PRESERVING VIRTUAL WORLDS 
 FINAL REPORT 
 
 
Jerome McDonough & Robert Olendorf 
Graduate School of Library & Information Science 
University of Illinois at Urbana-Champaign 
501 E. Daniel Street 
Champaign, IL 61820 
 
Matthew Kirschenbaum, Kari Kraus, Doug Reside & Rachel Donahue 
College of Information Studies 
Maryland Institute for Technology in the Humanities 
University of Maryland 
College Park, MD 20742-7011 
 
Andrew Phelps & Christopher Egert 
Interactive Games & Media Department 
B. Thomas Golisano College of Computing & Information Sciences 
Rochester Institute of Technology 
Rochester, NY 14623-5603 
 
Henry Lowood & Susan Rojo 
Humanities Research Group, Stanford University Libraries 
Green Library, 557 Escondido Mall 
Stanford University 
Stanford, CA 94305-6004 
 
 
Aug. 31, 2010 
 
 
Preserving Virtual Worlds was conducted with the support of the  
Library of Congress' National Digital Information Infrastructure for 
Preservation Program 
 
 
 
 
The report  is made available under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 license 
(See http://creativecommons.org/licenses/by-nc-sa/3.0/). 
? 2?
 Acknowledgements 
 
Numerous individuals assisted in the research described in this report, including many 
students and colleagues who have assisted in gathering data, proofing drafts, and helping us 
think through the difficult problems associated with preservation of digital games and 
interactive fiction. Several stand out for special recognition, however. The authors would like 
to thank Janet Eke, Patricia Hswe and Maeve Reilly for their tireless efforts in managing an 
extremely complex project and herding cats (and otters) as necessary. This project could not 
have happened without all of their extraordinarily capable assistance. We would also like to 
thank Beth Dulabahn at the Library of Congress for all of her help and for providing a 
valuable sounding board for ideas. The support and collaboration of Neil Fraistat and 
Dennis Jerz were invaluable, as was the assistance of Henrik Bennetsen and Matteo Bittanti.  
We are extremely grateful to J.P. Dyson, director of the International Center for the History 
of Electronic Games (ICHEG) and vice president for exhibit research at the Strong National 
Museum of Play, for being free with his time and his knowledge of preservation of gaming 
hardware, and to the Internet Archive for their support for our efforts. Finally we would like 
to thank the all of the game developers, artists and authors who have given us worlds to 
explore. 
? 3?
  
Table?of?Contents?
 1. Executive Summary ...............................................................................................................5?
 2. Introduction ............................................................................................................................9?
 What is a Virtual World? .....................................................................................................................9?
 Cultural and Economic Significance ............................................................................................ 12?
 Challenges to Preservation.............................................................................................................. 13?
 About the Project .............................................................................................................................. 15?
 About the Project Partners and Personnel ................................................................................... 16?
 About this Report.............................................................................................................................. 18?
 3. Games & Interactive Fiction: Collecting for Preservation ......................................... 19?
 Issues of Appraisal & Selection ..................................................................................................... 19?
 Working with Developers ................................................................................................................ 22?
 Context Information Reconsidered............................................................................................... 27?
 Literature & Ephemera ................................................................................................................... 29?
 4. Collections ............................................................................................................................ 33?
 Collections: Bibliographic & Archival Description ................................................................... 33?
 Collections: Data and Documentation ......................................................................................... 37?
 Problems/Opportunities in Access............................................................................................... 48?
 5. Software Preservation and the Law ................................................................................. 52?
 Copyright Issues: Complex Ownership and Orphan Works.................................................... 52?
 Emulators & the Law ....................................................................................................................... 53?
 6. Preservation Strategies....................................................................................................... 58?
 Analysis of Hardware Preservation ............................................................................................... 58?
 Analysis of Emulation and Virtualization .................................................................................... 61?
 Analysis of Migration ....................................................................................................................... 78?
 Analysis of Re-enactment................................................................................................................ 83?
 Digital Forensics and Disk Image Analysis ................................................................................ 85?
 Recommendations on Preservation Strategies ........................................................................... 87?
 7. When Strategies Fail: The Case of Second Life ............................................................ 89?
 Problems in Archiving Second Life............................................................................................... 89?
 Basics of Archiving a World............................................................................................................ 91?
 Creating a Manifest of an Island ................................................................................................... 91?
 Obtaining Permission ...................................................................................................................... 93?
 Copying a World ............................................................................................................................... 94?
 Failures in Archiving ........................................................................................................................ 96?
 8. Packaging Virtual Worlds ................................................................................................. 98?
 Virtual Worlds and the OAIS Reference Model.......................................................................... 98?
 Analysis of Packaging Requirements .........................................................................................101?
 Metadata & Packaging Recommendations...............................................................................102?
 9. Steps for the Future ..........................................................................................................105?
 Library of Congress National Collecting Plan ..........................................................................105?
? 4?
 Intellectual Property & Digital Preservation: Widening the Discussion.............................112?
 Representation Information and Format Registries/Tools ...................................................118?
 Digital Game Canon.......................................................................................................................120?
 Reimagining Videogame Asset Management & Preservation (ReVAMP) Symposium ..121?
 A Research Agenda for Preservation of Video Games & Interactive Fiction .....................123?
 10. Conclusion ........................................................................................................................126?
 Bibliography...........................................................................................................................128?
 Appendix A - Virtual Worlds that Died During the Grant ............................................134?
 Appendix B ? Media Coverage of PVW............................................................................140?
 Appendix C ? Preserving Virtual Worlds Ontology .......................................................143?
 Appendix D ? Multi-Institutional Collaboration: Lessons Learned ...........................162?
 Appendix E ? Publications .................................................................................................163?
 Appendix F ? Museogames Exhibit at the Mus?e des arts et m?tiers .......................166?
 Appendix G ? Second Life Deed of Gift ...........................................................................170?
 Appendix H ? Gaming Websites Identified in PVW Project .......................................172?
 Appendix I ? Sample Output from CopyBot ...................................................................187?
 ?
? 5?
 1. Executive Summary 
 
The Preserving Virtual Worlds project is a collaborative research venture of the Rochester 
Institute of Technology, Stanford University, the University of Maryland, the University of 
Illinois at Urbana-Champaign and Linden Lab, conducted as part of Preserving Creative 
America, an initiative of the National Digital Information Infrastructure and Preservation 
Program at the Library of Congress. The primary goals of our project have been to 
investigate issues surrounding the preservation of video games and interactive fiction 
through a series of case studies of games and literature from various periods in computing 
history, and to develop basic standards for metadata and content representation of these 
digital artifacts for long-term archival storage. The games included within our case studies 
were: 
 
? Spacewar! (1962) ? a space combat simulation for the PDP-1 computer; 
? Adventure (1977) ? one of the earliest of the text adventure games; 
? Star Raiders (1979) ? one of the more popular and complex games released for the 
Atari 2600 game console; 
? Mystery House (1980) ? the first work of interactive fiction to employ computer 
graphics as a significant part of the game, and not just text; 
? Mindwheel (1984) ? an interactive fiction work, notable for having been authored by 
U.S. Poet Laureate Robert Pinsky; 
? Doom (1993) ? the game which popularized the first-person, 3D shooter game; 
? Warcraft III: Reign of Chaos (2002) ? the real-time strategy game from Blizzard 
Entertainment; and  
? Second Life (2003) ? one of the most successful of the ?social? (i.e., non-gaming) 
virtual worlds. Given the large amounts of data involved in archiving all of Second 
Life, our project has focused on a small set of Second Life?s islands.  
 
Virtual worlds such as these are software artifacts, communities, and commodities. Their 
preservation is thus intertwined with issues of technology, social relationships and law, and 
our investigations have touched on issues in all of these realms. Significant problems we 
have identified for the preservation of these materials are summarized below. 
 
Obsolescence ? The most obvious problem affecting these materials is the obsolescence of 
the hardware and software infrastructures necessary to allow software to run. The earliest 
game in our case set, Spacewar!, currently exists in its original form stored on a punched paper 
tape intended to be read into the memory of a PDP-1 computer. There is, to the best of our 
knowledge, only one functioning PDP-1 computer left in the world, at the Computer 
History Museum in Mountain View, California, and paper tape readers are not exactly 
common equipment at this time. The fate of the paper tape of Spacewar! is the fate awaiting 
all games without the active intervention of preservationists. A book may pass 50 years on a 
shelf and still be readily accessible; rapid technological change and the resulting obsolescence 
of the technology necessary to access software mean that a computer game will not. 
 
Boundaries ? Identification of the exact boundaries of the object of preservation is difficult 
in the case of computer games. While we tend to think of the game as a relatively discrete 
package of software, the reality is that a functioning game involves a web of interconnections 
? 6?
 between the game?s executable, an operating system, the hardware platform used to execute 
both, and potentially network hardware and software and a multiplicity of other computer 
systems (witness the cases of Warcraft III and Second Life). Even a relatively simple, early game 
such as Adventure possessed a dependency on an operating system library in the source code 
that the original author, Will Crowther, sent to Don Woods. Games may also come in a 
multitude of official versions, as well as unofficial modifications (mods), rendering decisions 
about what to collect, even in the case of a single game, problematic. Determining what 
constitutes the game, and more importantly, what is necessary to preserve the game, can be an 
extraordinarily difficult exercise. 
 
Intellectual Property Law ? Copyright and related issues manifested themselves repeatedly 
in our studies. The Digital Millennium Copyright Act?s prohibition on defeating 
technological protection measures makes it impossible for a library to create a preservation 
copy of games employing DRM and anti-copying measures. While obtaining the permission 
of the rights owner to make a preservation copy offers a potential path around this obstacle, 
securing these permissions is complicated by the existence of a large number of ?orphan 
works? in the field of computer games, and the great difficulties encountered in trying to 
track intellectual property rights ownership in an industry as volatile as the game software 
industry. Intellectual property laws also may pose impediments to the development of 
emulation technology necessary for continuing access to some games. 
 
Collection Management ? As the Open Archival Information System Reference Model 
makes clear, preservation of any object requires the preservation of more than the object 
itself. Knowledge of how to decode and play back the information (representation 
information) must also be preserved, as must information which provides an intellectual 
context for the object to aid in understanding its full meaning and significance. Librarians, 
archivists and curators dealing with gaming materials must be extremely proactive in their 
collection management to insure that they are adding the requisite additional materials to 
their collections. Unfortunately, these types of materials can be exceptionally difficult to 
obtain in the case of computer games and electronic literature. Many of them may be in the 
hands of private companies that do not wish to see them released (or belonged to companies 
that have ceased to exist), and other information may be produced far outside of normal 
publication channels (e.g., machinima videos created in networked gaming systems and 
traded among fans), requiring significant time and energy to track down.  
 
Preservation Strategies & Significant Properties ? Standard preservation strategies for 
software entities such as computer games include migration (relatively easy if source code is 
available), emulation, maintenance of original hardware and operating systems (the 
?computer museum? approach), and in cases where these all fail, re-enactment/recreation 
(the game Mystery House is available on modern computing platforms in significant part 
because of the reimplementation undertaken by the Mystery House Taken Over project, 
which created a bug-compatible reimplementation of the original). None of these 
approaches, however, provides a perfect solution to the problem of preservation, and 
migration, emulation and re-enactment pose significant risks of altering the appearance and 
performance of games. Our research on emulation in particular shows that significant visual 
and aural aspects of the work can be strongly affected by running under emulation. Without 
a clear understanding of which aspects of a game are likely to be considered significant by 
scholars in the future, it is extremely difficult to choose an appropriate preservation strategy, 
? 7?
 and preserving games without any change in their appearance and play may simply not be 
achievable in many instances. 
 
Despite all of these significant issues, we have identified a number of immediate and 
intermediate steps that libraries, archives and museums can take to assist in the long-term 
preservation of games. These include: 
 
? Developing packaging standards for ingest of gaming materials into digital 
repositories that explicitly include support for application of the FRBR and OAIS 
data models to insure that the identification of materials and links between them are 
sufficiently precise and detailed to support preservation activity. We hope that the 
OWL ontology developed by our project, and demonstrations of its application with 
METS and OAI-ORE, might serve as the basis for further development in this area. 
? Have organizations such as the Library of Congress, the National Digital 
Stewardship Alliance and others concerned with large-scale collaboration between 
institutions in support of preservation take a lead role in trying to negotiate the 
distribution of responsibilities for collection development and management that will 
be essential to preserve computer games and interactive fiction in a distributed 
fashion. Given the common need for repositories of representation information and 
supporting documentation, the Library of Congress may wish to consider starting 
discussions with the National Institute of Standards and Technology and the Unified 
Digital Format Registry about the creation of such a repository, focusing initially on 
data format standards. 
? Help develop preservation systems that are accessible by and can accept 
contributions from the gaming community. The How They Got Game project, as 
part of the Preserving Virtual Worlds efforts, has established a sub-collection within 
the Internet Archive?s moving image collection called ?Archiving Virtual Worlds.? It 
contains video documentation of a large number of virtual worlds, much of it created 
by those worlds? users. Efforts such as this, that provide a stable environment in 
which to preserve the contributions of the gaming community and assist them in 
documenting their own activities and culture, are essential to the preservation of 
computer games and interactive literature. Moreover, such efforts help promote 
dialog among librarians, archivists and curators and the larger gaming community, 
and help build the partnerships essential to preservation activity in this field. The 
Library of Congress and members of the National Digital Stewardship Alliance 
should seek out further opportunities for such collaborative efforts with the gaming 
community. Development of emulators and emulation technology, an area where the 
gaming community has already invested significant effort, might be a potential arena 
for collaboration, as might establishing repositories of documentation of hardware 
platforms, such as the recently launched Maryland Institute for Technology in the 
Humanities? Vintage Computers website.1 
? Intellectual property laws as they currently stand represent serious obstacles to 
preservation of computer games and interactive fiction. The inability of libraries and 
other cultural memory organizations to make preservation copies of materials 
employing technological protection measures (TPMs) will certainly doom these 
????????????????????????????????????????????????????????
 1?Available at http://mith.umd.edu/vintage-computers/ 
? 8?
 materials to a rapid demise. The ?1201 anti-circumvention rulemaking process, due 
to the time limitations established on rules allowing circumvention of TPMs, the 
time and difficulty in coordinating applications for exemptions to the anti-
 circumvention prohibitions in the law, and the uncertainty of the outcome of any 
petition, does not provide libraries with a stable legal environment in which to 
conduct preservation planning. In addition to seeking collaboration with the 
inhabitants of virtual worlds, libraries, archives and museums need to build 
relationships with gaming companies to work on legislative changes that will enable 
preservation of computer games to proceed in a manner that protects the rights of 
games? creators while insuring that their creations are available to future generations. 
The ReVAMP symposium discussed in Chapter 9 represents the type of discussion 
that is necessary if a legal infrastructure conducive to digital preservation is to be 
forged. 
 
Computer games and interactive fiction form an essential part of our cultural heritage. These 
virtual worlds are unique forms of art, places for education, socializing, business and 
entertainment, and seem certain to play an increasing role in people?s lives. We hope that the 
research we report on below will help contribute to librarians?, archivists? and curators? 
efforts to insure that these virtual worlds remain living worlds. 
 
 
 
Jerome McDonough 
Preserving Virtual Worlds, Principal Investigator 
? 9?
 2. Introduction 
 
?This is our history, and just a handful of people are saving it.?  
 
? PixelVixen707, screen name of ?Rachael Webster,? herself a fictional character in 
the alternate reality game Personal Effects: Dark Art 
 
 
What is a Virtual World? 
 
Virtual worlds are software artifacts, communities, and commodities. They are places and 
spaces whose geography and landmarks can be as familiar as your own neighborhood, 
teeming with personalities that are rich and genuine and multifaceted, but?simultaneously 
and paradoxically?they are also always finally layers of logical abstractions mediated by the 
conventions of digital computing. They are also (often) branded media properties capable of 
generating massive revenues, or else generating very little, a phenomenon not irrelevant to 
their frequent and sudden demise.  
 
As software artifacts exhibiting complex dependencies on platform, operating system, and 
network environment, virtual worlds are undoubtedly among the most imperiled forms of 
interactive digital content. As communities?shared spaces and places?they are defined by 
no less delicate and idiosyncratic skeins of people, relationships, memories, and folklore akin 
to those found within oral cultures. Virtual worlds are not virtually?that is, ?almost??real. 
They are instead, to borrow a phrase from Jesper Juul (2005), precisely half-real: they are 
human products, scripted and engineered out of millions of lines of code written by dozens 
or hundreds or even thousands of individuals, but they are also focalizers for powerful 
collective acts of the imagination that rely on the same willing suspension of disbelief that 
characterizes immersion in other media, like novels and films. Virtual worlds are governed 
by the non-virtual realities of limitations on processor and rendering speeds, bandwidth, 
memory, and server technology, but they have also functioned to spur innovation and 
technical breakthroughs in those very same areas. Most virtual worlds are also buffeted by 
the inexorable pressures of the marketplace. During the two-year span of our project, no 
fewer than seventeen virtual world properties winked out of existence, with varying degrees 
of bang and whimper (see Appendix A). Even when profit-motive is not a consideration and 
a virtual world is built and maintained out of something like love, it remains vulnerable to 
the patience, attention span, resources, disposable income, and of course ultimately the 
inescapable mortality of its founders. 
 
The online journal Virtual Worlds Review defines a virtual world as ?an interactive simulated 
environment accessed by multiple users through an online interface.? Six essential features 
are prescribed: shared space (multiple users), a graphical user interface, immediacy 
(?interaction takes place in real time?), interactivity (?the world allows users to alter, develop, 
build, or submit customized content?), persistence (?the world's existence continues 
regardless of whether individual users are logged in?), and socialization, or a sense of 
community. While presumably capturing what many immediately think of when they conjure 
the mental image of a virtual world, we find this definition and these criteria limiting. For 
example, the requirement for a graphical user interface would seem to exclude text-only 
? 10?
 Multi-User Dungeons or MOOs (MUDs Objected-Oriented) from the definition of a virtual 
world. Likewise, Maze War, a remarkable 1974 multiplayer game which anticipated the first-
 person shooter genre would not qualify, since there is no persistence to the game world and 
precious little in the way of social interaction. 
 
The Preserving Virtual Worlds project has instead taken a deliberately broad and catholic 
approach to defining its scope and purview. Not all of the virtual worlds in our case set 
exhibit each of the above properties. This is partly a result of our mandate to explore a 
variety of different cases and preservation scenarios. However, it is also out of the 
conviction that no single quality?three-dimensional graphics, multiple users, user-
 contributed content, persistence?defines a virtual world. For our project, a virtual world 
means something like a digital setting whose properties are stable and coherent enough to 
deliver a consistent ludic or interactive experience to two or more users or to the same user 
over time. A virtual world also, however, inevitably involves some form of imaginative 
projection based on the promise of inhabitable space. An Excel spreadsheet is thus not a 
virtual world because despite being capable of producing consistent interactive experiences it 
fails to create the sense of immersive possibility so crucial to the experience of space and 
place. We believe virtual worlds thrive on this human craving for possibility. 
 
Perhaps no single example illustrates this more effectively than Battlezone, a coin-op arcade 
tank combat simulator first introduced in 1980 and subsequently ported to the Atari 2600 
and numerous other consoles and home computer systems. Battlezone, which was also 
eventually adapted by the US Army for 
training purposes, presents its players 
with a sparse and largely empty world, 
populated by abstract geometric 
obstacles and an array of hostile tanks, 
missiles, and spacecraft. The kill or be 
killed ethos of the game matches this 
stark, vector-drawn landscape 
perfectly. There is, however, one 
unusual touch: on the horizon of this 
world is a mountain range; peaks and 
valleys rising and falling like an EKG. 
Amongst the mountain peaks is a 
solitary volcano, a distinctive landmark 
capped off with a few blinking pixels 
meant to suggest lava or debris. 
 
This volcano became the stuff of gamer legend, with the rumor spreading that if you drove 
your tank far enough you could eventually get to it. And there was more: upon reaching the 
volcano you could drive up its side, and peer over the rim, and inside was a castle, and inside 
the castle treasure . . . One commentator puts it this way: 
 
The attraction of Battlezone?s world was so strong that many players wanted to 
turn their back on the fighting and drive their tank up into the mountains to 
go exploring. The designers of the game had to put in a routine to send a 
The Volcano in Batt lezone  
?
? 11?
 missile after would-be explorers so that arcade owners wouldn?t lose money 
on the peaceful tourists who didn?t want to fight. Many great legends 
emerged from the arcades that centered on finding a way to leave the fighting 
behind. 
  
This behavior anticipates the appeal of later, much more complex and fully rendered virtual 
worlds, such as those found in the Grand Theft Auto franchise, in which players often engage 
in ?sandbox? mode, foregoing the missions assigned to advance the plot in favor of ... just 
driving around. (The kind of free-form exploration that Situationist Guy Debord (1956) 
called a d?rive.)  Such activity is possible because all virtual worlds exist on a continuum 
between (again using terminology from Jesper Juul, who also discusses the example of 
Battlezone in this context) coherence and incoherence. A virtual world is coherent or 
incoherent to the extent that the player is able to use the information that is presented to 
him or her to fill in the gaps and blanks about other aspects of the world that are not literally 
portrayed. Battlezone, for example, gives us nothing with which to conjecture as to why its 
antagonists are fighting for possession of this barren planet, or how the crews of the tanks 
(are they even human?) eat or sleep or what they do to pass the time when they?re not all 
busy fighting. But specific details, like the volcano, can become the catalyst for such 
imaginative projections because they tantalize us with the promise of a more replete world 
just over the vector limned horizon. Human-computer interaction designer Stan Ruecker 
(2007) refers to this in a different context as the ?aesthetic function,? where attention to 
seemingly arbitrary cosmetic details?in this case, the erupting volcano, with its pixelated 
lava?has the salutary benefit of capturing a user?s attention and engendering trust in the 
surrounding environment. Virtual worlds, regardless of where they fall on Juul?s continuum 
between coherence and incoherence, exploit and depend on the effects of the aesthetic 
function on their users. 
 
We draw no hard and fast distinctions between traditional computer games and massive, 
multiplayer online virtual environments like Second Life. Our case set encompasses examples 
of both, as well as several works of interactive fiction which are at least as much instances of 
belles-lettres as games or entertainment. Steve Russell?s Spacewar!, programmed in 1961, is thus 
a virtual world for purposes of our project: indeed, it is a very literal virtual world, with its 
primitive visualization of a solar system on a screen at MIT (or today at the Computer 
History Museum in Mountain View). Will Crowther and Don Woods?s text-only 
ADVENTURE is a virtual world, its faithful recreation of Kentucky?s Bedquilt Cave 
complex yielding an uncannily accurate world model replete with objects, places, and non-
 player characters. John Romero?s and John Carmack?s DOOM is a virtual world, an 
architectural tapeworm enticing users to wonder what lies beyond the vanishing points of its 
seemingly endless corridors and mazes?precisely the question taken up by those who 
learned to build their own game levels, as well as early creators of machinima who used the 
game engine as a platform for narrative. Finally, we might recall the well-known origin story 
behind Will Wright?s landmark game, The Sims. Wright was at work on his first game design, 
a by-the-numbers shoot-?em-up for the Commodore 64 called Raid on Bungeling Bay; as he 
tells it, he became captivated by the idea of using the level editor to design new landscapes 
and combat zones, eventually finding this activity more compelling than the actual game 
play. This is the kind of world-making?made up of imaginative projection, social 
interaction, and unabashed play?that forms the basis for the virtual worlds we examined. 
 
? 12?
 Cultural and Economic Significance 
 
The word ?game? comes to us from the Old English gamen, itself descended from the 
Germanic, meaning joy, fun, and amusement. This etymology is consistent with most 
people?s contemporary sense of the term, and so it is not surprising that games are 
sometimes regarded as trivial, disposable, or at best tangential to those objects of cultural 
heritage seemingly possessed of more inherent gravitas. Yet even a moment?s reflection 
should reveal the shortcomings of this thinking. Historically, card, tabletop, and parlor 
games of all sorts have been integral to the human experience across cultures, spanning all 
levels of social class and caste. Yet as often as games are relegated to the category of the 
frivolous, they are just as often associated with violence, addiction, and dangerous 
tendencies. Digital games in particular seem to suffer for their close association with youth 
culture, and of course debates around violence and militarism in contemporary digital 
gaming are constantly in the headlines.  
 
This report does not seek to intervene in such debates, other than to point out that the sheer 
topicality of computer games and virtual worlds at this particular moment in our collective 
history would seem to make preserving accurate and authoritative records of them an 
essential aspect of the mission of an institution such as the Library of Congress. Likewise, it 
is worth remembering that the history of technology reveals consistent patterns of anxiety 
around new media and new forms of immersive storytelling. Consider the following, for 
example: 
 
A whole family, brought to destitution, has lately had all its misfortunes 
clearly traced by the authorities to an ungovernable passion for novel-reading 
entertained by the wife and mother. The husband was sober and industrious, 
but his wife was indolent, and addicted to reading everything procurable in 
the shape of a romance. This led her to utterly neglect her husband, herself, 
and her eight children. One daughter, in despair, fled the parental home and 
threw herself into the haunts of vice?. The house exhibited the most 
offensive appearance of filth and indigence. In the midst of this pollution, 
privation, and poverty, the cause of it sat reading?and refused to allow 
herself to be disturbed in her entertainment. 
 
? ?T. C., ? The Christian?s Penny Magazine and Friend of the People (1859); 
Qtd. in The ?Dangerous? Potential of Reading: Readers and the Negotiation of 
Power in Nineteenth Century Narratives (Aliaga-Buchenau, 2003). 
 
This warning about the perils of too much novel reading seems instantly familiar, evoking 
many of the jeremiads about the addictive qualities of gaming one associates with public 
debates around Everquest, World of Warcraft, and Second Life. The popular pastimes of the 
present?outright dangerous or at best mere distractions from more wholesome pursuits?
 are the new cultural norms of tomorrow. Just as it would be very difficult to analyze a 
painting such as Manet?s Olympia without reference to Titian?s Venus of Urbino, it will prove 
very difficult in the future to discuss games such as Star Wars Galaxies without reference to 
Spacewar!. It will also prove difficult to provide a full accounting of our cultural heritage 
outside the realm of gaming without preserving games as well. In an age where games such 
as Tomb Raider, Resident Evil, Final Fantasy and DOOM spawn media franchises including 
? 13?
 films, novelizations, comic books and a variety of action figures and collectibles, and where 
games such as Microsoft?s Halo and Blizzard?s World of Warcraft provide media production 
platforms for the creation of machinima videos, it is impossible to provide a proper 
contextualization for much of our existing popular culture without preserving games. 
 
Virtual and gaming worlds also exhibit increasingly complex transactions with segments of 
society beyond popular entertainment, for example the placement of ads in XBox games by 
the Obama presidential campaign (GamePolitics.com, 2008). Bainbridge (2007) has noted 
the potential impact of virtual environments on the research community, both as virtual 
laboratory spaces and as settings for economic and social research. Increasingly, any 
complete understanding of modern society and culture requires an understanding of the 
world of gaming, and if cultural heritage institutions are to adequately serve researchers, we 
must develop means of preserving these new information resources. 
 
Games have also redefined the social spaces around them: whether it is the coin-op cabinet 
in a neighborhood tavern or pizza parlor, the video arcade as a new kind of adolescent 
hangout, the changing dynamics of the family living room?where the TV set takes on a 
new role as a game-playing device?or the LAN parties that characterize online gaming or 
cybercaf?s in Seoul, it is obvious that games have reconfigured physical as well as virtual 
space. 
 
Economically, computer games have become a significant part of the global economy. 
Global sales of video games were estimated at $46.5 billion in 2009 (Wu, 2010), and within 
the United States, direct and indirect employment by the gaming software industry accounts 
for over 80,000 jobs (Siwek, 2007). Socially, video games have become one of the most 
popular forms of entertainment within the United States, with 67% of American households 
playing computer or video games (Entertainment Software Association, 2010), and with 
online gaming sites in the United States reporting over 190 million visitors in a single month 
in 2008 (ComScore, 2009). 
 
Challenges to Preservation 
 
Unlike a book in a library, computer games have very poorly defined boundaries that make it 
difficult to determine exactly what the object of preservation should be. Is it the source code 
for the program? The binary executable version of the program? Is it the executable program 
along with the operating system under which the program runs? Should the hardware on 
which the operating system runs be included? Ultimately, a computer game cannot be played 
without a complex and interconnected set of programs and hardware. Is the preservationist?s 
job maintaining a particular, operating combination of elements, or is to preserve the 
capability to produce an operating combination using existing software and hardware? Is it 
both? Once these questions of the boundaries of the preservation object are addressed, there 
are a host of other difficulties presenting the would-be preservationist. What information, 
beyond the game itself, will we need to ensure continuing access to the game? How should 
librarians, archivists and preservationists go about organizing the body of information 
needed to preserve a game? What strategy should we adopt to preserve software in a 
technological environment in which computing hardware and operating systems are 
undergoing constant and rapid evolution? Given the costs of preservation of normal library 
? 14?
 and archival materials, how can we possibly sustain the additional costs of preserving these 
complex and fragile technological artifacts? 
 
Given that the entire report really revolves around the challenges to preserving and 
maintaining access to games and virtual worlds, here we simply run down, in abbreviated 
form, the main problem domains. 
 
? Hardware obsolescence ? The original console or computing platform used to run 
the game may cease to be supported or even available in the aftermarket. 
? Software obsolescence ? The original software needed to run the game?operating 
system, drivers, frameworks?may lose support, cease development, or become 
incapable of running on future hardware/software configurations. 
? Scarcity ? Some video games are produced in limited quantities and are subject to 
the dangers of media decay. This is especially likely to be the case for special editions 
and releases, recalled games, or art games. 
? Third party dependencies ? Currently most emulators are developed by the game 
community and are of questionable legality. They are also typically created without 
the benefit of the original specifications and are themselves at risk of becoming 
obsolete. 
? Complex, proprietary code ? And an associated lack of documentation. Digital 
games are generally released as compiled binaries with no documentation of the 
compiling process, or even the programming languages used. Not having access to 
the source code or language specifications makes migrating or emulating software far 
more difficult. 
? Authenticity ? The elephant in the digital preservation room, proving that a digital 
object is what it claims to be, free from tampering or corruption. Digital games enjoy 
many versions between the first prototype, the official release (on multiple 
platforms), and cracked or otherwise altered unauthorized editions. Especially for 
older games, the only extant copy may exist in a fan-run web repository, making the 
authenticity impossible to establish. 
? Intellectual Property Rights ? The game development industry is highly creative 
and competitive, leading developers to be conservative with their intellectual 
property. Most have instituted restrictive shrink-wrap licenses reflecting this. And 
yet, once a game is no longer actively marketable, they are less likely to respond to 
inquiries about licensing for it. 
? Significant properties ? What are the significant properties of a game that must be 
maintained with each transformation/preservation action? What makes Mario 
Mario? How important are font size and color palette? What about the speed of text 
scrolling or sprite movement? What about controllers? How faithful must we stay to 
the original code? Significant properties are essential to define, as they play a major 
role in determining authenticity.  
? Context ? Although not an immediate threat to the preservation of games, building 
contextuality is important to creating understanding for future users. This is truer for 
digital games than many other record types because, as technology advances, game 
players who have only been exposed to the latest and greatest may be apt to play an 
older game and say, ?so what?? even though the game might have been revolutionary 
for its time. For example, Sierra?s Mystery House was the first text adventure to 
? 15?
 incorporate graphics. An amazing breakthrough in its day, it seems crude in 
comparison to today?s virtual environments.  
 
About the Project 
 
Preserving Virtual Worlds (PVW) was a two-year collaborative research venture of the 
Rochester Institute of Technology, Stanford University, the University of Maryland, the 
University of Illinois at Urbana-Champaign and Linden Lab (2008-10). The project is one of 
several launched as part of Preserving Creative America, an initiative of the National Digital 
Information Infrastructure and Preservation Program at the Library of Congress intended to 
seek novel solutions for the preservation of commercial digital content. Preserving Virtual 
Worlds has focused on developing methods for the preservation of digital games and 
interactive fiction through the creation of standards for metadata and content representation. 
 
The PVW project has employed a case set approach in its investigations, choosing a set of 
games and using those to identify representative types of preservation problems posed by 
each member of the case set. We have selected a number of games from different periods in 
gaming history, from different platforms and with different intellectual property status to try 
to maximize the opportunities for identification of problems. Games, interactive fictions, 
and virtual worlds within the case set include: 
 
? Spacewar! (1962) ? a space combat simulation for the PDP-1 computer; 
? Adventure (1977) ? one of the earliest of the text adventure games; 
? Star Raiders (1979) ? one of the more popular and complex games released for the 
Atari 2600 game console; 
? Mystery House (1980) ? the first work of interactive fiction to employ computer 
graphics as a significant part of the game, and not just text; 
? Mindwheel (1984) ? An interactive fiction work, notable for having been authored by 
U.S. Poet Laureate Robert Pinsky; 
? DOOM (1993) ? The game which popularized the first-person, 3D shooter game; 
? Warcraft III: Reign of Chaos (2002) ? The real-time strategy game from Blizzard 
Entertainment; and  
? Second Life (2003) ? one of the most successful of the ?social? (i.e., non-gaming) 
virtual worlds. Given the large amounts of data involved in archiving all of Second 
Life, our project has focused on a small set of Second Life?s islands.  
 
For each of the games in our case set we have tried to identify specific problems that might 
impede a library?s or archive?s ability to preserve the game in the long-term. Identification of 
these problems has involved content analysis of the games themselves (including 
documentation provided with the game) as well as research into the various games? 
intellectual property status. Common problems identified for games in our case set include 
the unavailability of the original game, content management problems resulting from the 
complex versioning issues surrounding software, storage media obsolescence and fragility, 
degradation of game play due to the use of emulation or migration, unavailability of 
representation information documenting the file formats used in game software, and a 
number of intellectual property issues. 
 
? 16?
 Based on our investigations into the games in our case set, we have developed a set of 
requirements for game preservation. We have also implemented a new ontology for game 
description that draws upon both the Functional Requirements for Bibliographic Records 
Final Report and Open Archival Information System Reference Model, and successfully 
used that ontology to create archival information packages for these games in conjunction 
with existing digital preservation standards such as METS, OAI-ORE and BagIt. We have 
loaded those packages into preservation repositories at Stanford University and the 
University of Illinois at Urbana-Champaign. 
 
Project infrastructure included a listserv with all participants subscribed; biweekly conference 
calls, both project-wide and for the technical team (primarily personnel at Illinois and RIT); 
several face-to-face all-hands meetings over the course of the project; a public website, 
including a blog (http://pvw.illinois.edu/pvw); and a wiki, with many (but not all) content 
areas also open to the public (https://apps.lis.illinois.edu/wiki/display/PVW/Home). 
Readers are encouraged to consult these online resources, as well as the project partners? 
publications (see Appendix E) for further information and discussion. 
 
About the Project Partners and Personnel 
 
The University of Illinois at Urbana-Champaign was the lead institution, with Jerome P. 
McDonough, assistant professor in the Graduate School of Library and Information Science, 
serving as principal investigator. Project personnel included a diverse array of staff, faculty, 
and graduate and professional students from Illinois and three other universities, thereby 
extending the project?s demographics to additional information schools, a program in game 
and interactive media design, a university research library, and a digital humanities center. 
The team included specialists in game design and game studies, library and information 
science, archives, digital humanities, intellectual property, and descriptive and analytical 
bibliography. The participating units within the four partner institutions are further 
described below. 
 
University of Illinois at Urbana ? Champaign (Lead Institution) 
Graduate School of Library and Information Science (GSLIS) 
 
The Graduate School of Library and Information Science began as the first library science 
program in the Midwest, founded in 1893 by Katharine Sharp. More than a hundred years 
later, it is consistently ranked as one of the very best in the field. The School?s faculty 
believes strongly that librarianship and newly emerging forms of information science must 
develop together, in order to ensure that libraries resist obsolescence and newer institutions 
learn the importance of access, privacy, and service. The mission of the Graduate School of 
Library and Information Science is to provide: graduate education for leaders in research and 
practice in the fields of library and information science; groundbreaking research to advance 
preservation of and access to information in both traditional and digital libraries and in the 
many settings outside of libraries where large amounts of critical information are collected; 
and useful service to librarians and other information service providers, as well as to the 
citizens of Illinois. 
 
? 17?
 Rochester Institute of Technology 
Department of Interactive Games and Media (IGM), and 
B. Thomas Golisano College of Information Sciences and Technology 
 
The Department of Interactive Games and Media is renowned for its innovative approaches 
to media-centric computing that merge the creative design of the interactive experience with 
the development of content, technologies, and systems that form the basis of such work. 
The department will support, wherever and whenever possible, multi-disciplinary work that 
fuses these elements in pursuit of its academic mission. IGM is comprised of talented and 
motivated individuals from a variety of academic backgrounds with a shared interest in 
computing as it relates to interactive and social media, new media, games, simulations, and 
media-centric systems of all varieties. The department?s mission is to provide a sustained 
educational environment that supports and encourages creative and collaborative academic 
inquiry by both faculty and students into these areas. The department?s programs, 
coursework, research, and development efforts will provide students with the knowledge and 
skills to pursue meaningful and rewarding careers in this arena, while simultaneously 
advancing the field and helping to provide a well-rounded educational experience. 
 
The Golisano College is one of the largest and most comprehensive computing colleges in 
the nation. The College has garnered accolades and recognition as a premier computing 
education and applied research facility. Housed in a 125,000 sq. ft. state-of-the-art building, 
the College showcases cutting-edge innovation and world-class faculty who are passionate 
about their work. The Golisano College includes the departments of Computer Science, 
Information Sciences & Technologies, Interactive Games & Media, Networking, Security, 
and Systems Administration, Software Engineering, as well as the Ph.D. in Computing and 
Information Sciences program, the research arm of the College. This mixture of applied 
computing disciplines is unique and allows the College to offer a strong, diverse series of 
programs centered on computing, from infrastructure to the end user. 
 
Stanford University 
Stanford University Libraries and Academic Information Resources (SULAIR) 
 
Stanford University Libraries and Academic Information Resources includes more than 30 
libraries and programs supporting research, teaching, and learning at Stanford University. 
SULAIR acquires and delivers library collections in all formats, establishes policies and 
standards to guide the use of academic information resources, develops training and support 
programs for academic uses of computers, and maintains a broad array of electronic 
information resources, including the online library catalog and several hundred article and 
indexing databases and electronic journal subscriptions. In each library unit, knowledgeable 
professional staff provides assistance in locating and using print and online information 
resources. 
 
University of Maryland 
College of Information Studies, and 
Maryland Institute for Technology in the Humanities (MITH) 
 
The College of Information Studies, Maryland?s iSchool, engages in collaborative, 
interdisciplinary, and innovative research, teaching, and service. We educate information 
? 18?
 professionals and scholars, and we create knowledge, systems, and processes. The iSchool 
offers Master?s degrees in Library Science (MLS), Information Management (MIM), and a 
doctorate degree in Information Studies. In Fall 2009, 344 students were enrolled in the 
MLS program, 141 enrolled in the MIM program, and 26 enrolled in the doctoral program. 
Approximately 70% of the total student body is female. The iSchool has 35 faculty and staff 
and 33 adjunct faculty representing diverse subject areas in information studies. The iSchool 
serves the mid-Atlantic region. 
 
The Maryland Institute for Technology in the Humanities is the University of Maryland?s 
primary intellectual hub for scholars and practitioners of digital humanities. On a day-to-day 
basis, MITH functions as an applied think tank for the digital humanities, supporting faculty 
fellows and engaging in sponsored research clustering around digital databases and tools, 
thematic research collections, text mining and visualization, and the creation and 
preservation of electronic literature, digital games, and virtual worlds. MITH has sponsored 
over two dozen faculty and graduate student fellows, and serves a community of several 
hundred researchers and interested members of the public who attend its events. 
 
About this Report 
 
This is the final report of the Preserving Virtual Worlds project, co-authored by the 
members of the PVW team. While presenting broad ranging and at times in-depth 
discussion, together with conclusions and recommendations and various supporting 
documentation in the appendices, it nonetheless cannot claim to be a comprehensive 
summation of the full range of PVW?s activities over the last two years. The principle 
audience for the report is library and information science professionals at the Library of 
Congress and other collecting institutions who have already or will soon be developing 
collections policies for computer games, interactive fiction, and virtual worlds. We also, 
however, hope that the report will be of interest to game developers and designers, many of 
whom as yet fail to take even rudimentary steps necessary to ensure the preservation of their 
own creative legacy and intellectual property; to the academic game studies community, 
especially those scholars with an interest in understanding the material underpinnings of the 
platforms and systems they study; and finally, but not least, to the fan and game player 
communities who have already done so much to safeguard future access to the content they 
cherish. 
 
The report is made available under a Creative Commons Attribution-NonCommercial-
 ShareAlike 3.0 license (see http://creativecommons.org/licenses/by-nc-sa/3.0/). 
? 19?
 3. Games & Interactive Fiction: Collecting for 
Preservation 
 
Issues of Appraisal & Selection 
 
In The Study of Games (1971), an essential treatment of the anthropology of games, Elliott 
Avedon and Brian Sutton-Smith asked: 
What are games? Are they things in the sense of artifacts? Are they 
behavioral models, or simulations of social situations? Are they vestiges of 
ancient rituals, or magical rites? It is difficult and even curious when one tries 
to answer the question ?what are games,? since it is assumed that games are 
many things and at the same time specific games are different from one 
another?but are they? (Avedon & Smith, 1971, p. 419) 
 
We are confronted with a number of questions here. One concerns the essential nature of 
games. Either games are fixed objects?perhaps authored texts or built artifacts?or 
alternatively, they are the experiences generated by a framework of rules, codes, or stories 
and expressed through interaction, competition, or play. Text or performance? Artifact or 
activity?  
 
Another question is whether there are general structural similarities among all sorts of games. 
Answering this question fully would be out of scope for our project, which was focused on 
practical issues of collection and preservation. However, it is worth noting that the authors 
of The Study of Games conclude with ?seven elements in games? distilled from studies by 
psychologists, mathematicians, and others. These structural elements include ?procedures 
for action,? ?roles of participant,? ?participant interaction patterns? and the like, taking 
games away from the notion that they are stable artifacts or texts. These elements 
underscore the importance of documenting interactivity as a historical phenomenon, 
something that predated computers (an obvious statement to any player of Diplomacy or 
Dungeons and Dragons [D&D]). As fashionable as it has become to discuss games as cinematic 
or as narratives, let us not forget that actions and responses are fundamental to the nature of 
interactive games. Games provide a structure within which players do something?whether 
the game is baseball, D&D, or Myst?and this structure is not compelled to be a linear 
narrative. Collections of digital games provide a tailor-made opportunity to document 
interactivity in ways that are not provided by other media. Any collecting activity thus must 
consider not only the fixed content of games as authored texts or software, but also 
examples of game development, game play, and player response and activity that provide a 
more complete picture of digital games as an interactive medium. We will have more to say 
about the implications of these considerations later in this report. 
 
Emphasis on both selection and appraisal underlines the diverse nature of the game-related 
collections, from collections of published games well-suited to selection policies traditionally 
in the domain of library collection development, to collections of complex software, 
artifacts, and archives better fitting notions of appraisal and ?special? collecting developed in 
the realms of archives, manuscript divisions, and museums. The common element uniting 
? 20?
 selection and appraisal is the intellectual effort of describing collecting and preservation 
priorities. 
  
Two of the PVW partners have active collecting programs: Stanford in games and virtual 
worlds, and Maryland in interactive fiction. The Library of Congress asked for our input with 
respect to the selection/appraisal problem, in order to aid its own eventual collecting 
activities in the areas of digital games and related forms of interactive software. In this spirit, 
we offer some thoughts that derive in part from our project work. 
 
With the growth of interactive software as a basis for entertainment and cultural production, 
it is safe to predict that many libraries will establish collecting programs for digital games and 
related artifacts, such as hardware consoles and interface devices. The foundation for such a 
collection in most cases is a library of published console- and personal computer-based 
games. A collection development policy is therefore a statement of the scope and extent of 
this library. Based on institution-specific program and patron need, the scope will encompass 
chronological, technical, and access considerations. A minimal or basic information-level 
collection, for example, would likely focus on current and recently published games that run 
on readily available and currently supported platforms (XBox 360, Wii, PS3, Windows PC, 
Apple Macintosh). With the progression through deeper levels of collecting (instructional 
support, research, comprehensive collections), decisions need to be made about the 
acquisition of content and hardware for obsolete platforms, some of which are still 
supported through third-party vendors (e.g., Atari VCS, Nintendo NES) and others of which 
are largely unsupported, historical operating systems (e.g., DOS), or titles that involve 
particular difficulties with respect to security, account management, or patron access (server-
 based games, handheld games, games distributed by web services such as Steam or XBox 
Live). Institutions supporting research or comprehensive-level collections will likely need to 
support multiple access solutions, such as a managed space for use of recent games on 
currently supported hardware platforms, a ?special collections? solution for archival and 
historical materials, and a digital repository for media migration and long-term preservation, 
as well as possible future access solutions. 
 
Once the overall scope of the collection is decided upon, the next level of selection policy 
should cover a mix of traditional library decision points (format, chronological categories, 
language of publication, country of origin) and categories more specifically tuned to digital 
games and virtual worlds. With respect to the former, it should be noted that ?chronology? 
in the digital game industry is very closely linked to decisions about platforms of interest; the 
entire history of published digital games currently spans fewer than four decades, yet this 
history is closely tied to a very strong notion of platform and operating system 
?generations.? Also, the country of origin is a particularly important matter, especially given 
the historical importance of Asian game publishers, notably from Japan, Korea, and China.  
 
Several criteria for selection decisions are oriented towards specific uses or capacities of 
digital games. As forms of software, for example, many digital games can be modified or re-
 used as platforms for activities that are not restricted to the original game content. Some 
software platforms, such as Flash, console environments (such as XBox Live Arcade) or 
even ?ancient? hardware platforms, such as the Atari VCS, provide a foundation for 
independent games produced by teams as small as a single designer/coder. Thus a game 
might be modified from an existing title or created as a one-off project for an art installation, 
? 21?
 used to create animated movies, or provide a basis for commentary on current events 
(?newsgames?). Such uses have come to define entire areas of activity that require special 
tactics of collection and preservation, such as ?art games,? machinima, serious games, and so 
on. These areas, in turn, provide specific foci for collecting activities, depending on local 
priorities. Thus, a library might well decide, in addition to a broad collection of current titles, 
to collect a specific type of game, such as art games or ?auteur? games as a focused 
specialization. Defining such areas in a collection policy statement is a more useful approach 
for collection development than typically applied categories for discussion of digital games, 
such as genre, because such categories are often loosely applied and may in fact produce 
artificial divisions of content based entirely on fine distinctions of game-play characteristics 
rather than inclusive research priorities. 
 
Other important aspects that may figure in a collection policy statement include 
technological aspects, social or ?community? aspects (particularly with respect to network-
 based games), competitive aspects (digital games as e-sports), surface qualities (graphics, 
audio, music) and inter-textual or cross-media qualities of games (e.g., games in relationship 
to literature or cinema). Thus one collection might be differentiated from another by an 
emphasis on competitive, multi-player games, including documentation of events, 
tournaments, and ?virtual communities? of fans through the collection of websites and 
replays, or perhaps by a focus on games based on films, which might include a companion 
collection of films based on games. Finally, with respect to collections based in the United 
States, it may be important to differentiate regional collecting activities. Designers, 
developers, publishers, game, and technology companies have historically been concentrated 
in a few regions, such as Northern California, Washington state, Maryland, and Texas. Thus, 
Stanford University?s game-related collections are closely tied to the older Silicon Valley 
Archives, and the University of Texas? efforts have thus far emphasized the collection of 
papers and artifacts from Texas-based designers, technologists, and musicians. 
 
The linkage of game collections and archival collecting efforts raises the related, but separate 
question of appraisal of game-related collections that consist primarily of documentation, 
personal papers, corporate records, or other categories we might call archival or special 
collections material. Appraisal of games as a part of archival practice was not a core issue for 
investigation by the PVW project. It refers to the process of evaluating collections of records 
for acquisition, retention or disposition, an assessment made in the context of a particular 
institution?s mission and/or research needs. Richard Pearce-Moses (2005), in the Society of 
American Archivists? A Glossary of Archival and Records Terminology, defines appraisal as ?the 
process of determining whether records and other materials have permanent (archival) 
value? and notes that it ?may be done at the collection, creator, series, file, or item level.? In 
this sense, appraisal is an issue for organizations such as corporate archives in game 
development companies who must make retention decisions on a regular basis; some of the 
questions around this issue of corporate stewardship of archival records were addressed in 
Rachel Donahue?s work included in the IGDA?s Before It?s Too Late: A Digital Game 
Preservation White Paper (Monnens et al., 2009). Academic institutions with active programs in 
game studies or game development will eventually face similar issues when making decisions 
about the retention of faculty and project papers in these areas. While PVW was not focused 
on the appraisal process, we do note that there is a need for careful attention in archival 
studies to this issue as it relates to game-related archives, particularly with respect to a better 
understanding of production work-flows, intellectual property considerations, and other 
? 22?
 matters as they are specifically shaped in the game industry. We also note that some work 
has begun in this area, notably by Prof. Megan Winget at the University of Texas. 
 
No single institution can do the work of comprehensively collecting and preserving digital 
games and their history. Competent institutions must join together to build archives of 
computer game history. Lay historians of digital games, mostly players and fans, have created 
websites and entire communities of game players dedicated to the preservation of game 
content and technology, and have made emulators and collections of game movies available 
to their communities. Institutions committed to game preservation and history could enlist 
these pioneers in the effort to create more permanent historical resources, and intelligent 
collection policies might provide a basis for initial discussion among potential participants. 
The Digital Game Canon, discussed later in this report, is one example of how such an 
effort might proceed. Indeed, participation of the game industry, museums, and academic 
institutions in this project can help to defuse the adversarial relationship between, say, the 
emulation community and publishers by developing mutually acceptable practices with 
respect to intellectual property and access. Better communication of collection priorities 
from libraries and museums working in this area might even provide a solid basis for 
industry support and participation. 
 
Working with Developers 
 
Preservation of video games and interactive fiction requires preserving not only the games 
themselves but also contextualizing material necessary to understand games? origin and use. 
Significant amounts of this type of material can be found in the possession of the software 
developers and designers who create games and the companies that distribute and support 
them. Much of this material, unfortunately, is not made public, although in some instances it 
may end up in archival collections. Our project has attempted to provide an initial 
classification of materials that those engaged in game preservation might wish to try to 
collect from developers: 
 
? Source Code & Other Game Assets (including text, still image, audio, video, and 3D 
files) ? As noted in the discussion of FRBR and games as bibliographic objects, a 
game can have many expressions, and while most users will wish to interact with a 
binary executable version of the game, obtaining and preserving source code for a 
video game can serve at least three significant purposes: 1.) it provides information 
on game production that is not available from the executable version to users 
interested in the underlying technology of games; 2.) it enables migration as a 
potential preservation strategy for a game when emulation is not viable; and 3.) it can 
improve the results of using emulation as a preservation strategy in those cases 
where producing a faithful rendition of game play requires recompilation of the 
source. In the case of Spacewar!, for example, in order for the online version of the 
game to run correctly on an emulated version of the PDP-1 computer, minor 
changes to the original code had to be made (clock speeds on modern computers 
being somewhat different than a PDP-1). In addition to source code, having copies 
of draft and final versions of assets used in a game (including text, still image, audio 
files, video files and 3D models) can ease scholarly analysis and use of these 
materials. 
? 23?
 ? Technical Documentation ? software and hardware development teams for games 
will typically accrue a variety of technical materials in the course of game 
development regarding the underlying technologies necessary to enable a game?s use, 
ranging from documentation of APIs for particular hardware devices, programming 
language documentation, their own notes on issues like enabling cross-platform 
compatibility of code, etc. This information may be valuable to scholars wishing to 
understand game production processes, as well as to preservationists trying to 
achieve a better understanding of the platforms on which a game was designed to 
run. 
? Production Materials ? understanding of the life cycle of a game can also be 
enhanced by access to materials the game?s designers used in its development. This 
can include design notebooks, storyboards, mock-ups created of game objects, 
scripts, character profiles, maps of the game?s terrain, and similar materials. 
? Designer Stories ? some of the more interesting material that can be collected 
regarding games? origins and use does not reside in documents, but in the designers? 
minds. Interviews with game designers and recordings of designers presenting their 
work at conferences can be a remarkably rich source of data for those wanting to get 
a more complete understanding of the history of games? development. 
? Records of Interaction with the User Community ? Game companies inevitably 
generate a fair amount of information in the process of interacting with their users, 
including bug reports on software, records of support calls, discussion forums and 
wikis. Some of this information may not be available due to privacy issues or other 
considerations, but it can provide valuable information regarding the game designers? 
relationship with users as well as the culture of game use. 
 
The game source code, production assets and the technical documentation acquired and 
generated in the course of game production are likely to have the greatest direct impact on 
the preservability of a game. The source code and game assets widen the set of potential 
preservation strategies, and technical documentation regarding the platform on which the 
game was intended to run is second only to representation information about the game files 
themselves in terms of its significance for supporting technical preservation activities. 
Production materials, oral documentation from game designers, and records of interaction 
with a game?s user community are of great value to anyone seeking to understand the life 
cycle of a game over time. 
 
Unfortunately, game companies may be unwilling or unable to part with all of these 
categories of materials for varying reasons. While there are exceptions, such as id Software?s 
treatment of DOOM, game companies are typically not particularly enthused about publicly 
releasing their source code, technical documentation or production materials, which might 
alert competitors to the design processes. Any social scientist who has studied software 
companies can attest to the fact that getting a programmer or designer to take the time to sit 
down for an interview is a difficult process, as companies do not wish to sacrifice the 
valuable time of their employees in an industry which is extremely driven by release dates. 
They are also disinclined to have their internal operating procedures made public. Records of 
interactions with users can raise obvious issues of confidentiality and privacy. 
 
This is not to say that game companies will necessarily be unwilling to share these materials, 
but there are obviously a number of factors working against making them available that 
? 24?
 libraries, archives or museums seeking to collect and preserve games must address. The most 
fundamental issue for any cultural memory organization attempting to collect this material is 
trust. Many software companies would view sharing something like source code as 
equivalent to handing over the crown jewels. Assurances that material will be ?dark? 
archived and made available only at some later date or under certain conditions are not of 
any real significance to a software company unless they already have trust in the individuals 
and the institution making the promises. 
 
Familiarity is key to building trust, and so cultural memory organizations seeking to collect 
materials from game designers need to continually engage with those from whom they are 
seeking materials, and not just as part of the process of negotiating donation of a collection. 
Librarians and archivists need to be in regular attendance at events that draw the video game 
design industry, including industry-oriented conferences such as the Game Developers 
Conference and E3 Expo, gamer-oriented conferences such as PAX Prime and VGXPO 
and academic conferences such as the International Conference on Computer Games, the 
IEEE Conference on Computational Intelligence and Games, and the Conference & 
Festival of the Electronic Literature Organization. Archivists are well aware that personal 
relationships can be the key to successful acquisition of materials, and maintaining a visible 
presence in the venues inhabited by game companies is critical to that effort. 
 
Game companies are also more likely to be willing to discuss donations with organizations 
that they see make active contributions to the world of gaming. Cultural memory institutions 
that wish to acquire materials from game developers need to have public (and permanent) 
activities around video games and electronic literature. Points of presence for gaming 
materials in the library, both physical and virtual (see, for example, the University of Illinois 
Gaming Collection (http://www.library.illinois.edu/gaming/index.html), Stanford 
University Library?s Stephen M. Cabrinety Collection in the History of Microcomputing 
(http://sulair.stanford.edu/depts/hasrg/histsci/index.htm), the University of Maryland?s 
Maryland Institute for Technology in the Humanities? Deena Larsen collection 
(http://mith.umd.edu/larsen/dlcs/), and the Video Game Archive at the University of 
Texas at Austin (http://www.cah.utexas.edu/projects/videogamearchive/index.html), as 
well as sponsoring events such as workshops on gaming research and game nights, help the 
institution promote an image of itself as actively contributing to gaming culture. To the 
extent that an institution is seen as having goals that align with the gaming industry regarding 
the importance of promoting games, it will be more likely to be seen as trustworthy by the 
industry. 
 
Choosing Versions: What?s the Difference? 
 
Digital games and virtual worlds are not static or homogeneous entities. Like books with 
different printings, editions, and translations, like art works with derivatives, prints, copies, 
and fakes, like films with director?s cuts and abridgements, games often exist in numerous 
different versions. These versions include ports for different platforms, patches and 
upgrades, user-contributed modifications (?mods?), pirated copies and hacks, and sequels 
that may prove more or less contiguous to the original. And here we are only talking about 
the game as a finished product. During development, the game will iterate through alpha and 
beta versions with greater or lesser availability to the public or to smaller communities of 
play testers. The constituent digital components of the game, meanwhile, from source code 
? 25?
 to graphics, movies and soundtrack, will manifest their own complex hierarchies of variants 
and version forks on the development tree. Little wonder then that programmers rely on 
version control repositories to manage these relations as they check code or master files in 
and out of the object library in the course of their work. Versioning is, in short, a fact of life 
in the digital world?indeed, one can argue that it is essential to the ontology of the digital at 
its most fundamental level since (strictly speaking) every time a file is accessed a new copy of 
it is created. 
 
Libraries and collecting institutions will face constant decisions over which ?version? of a 
game or virtual world to acquire, often also including the question of what source media it 
comes stored on. Limited resources will inevitably force trade-offs and decision-making. 
One cannot collect?let alone preserve and maintain access to?everything. Sometimes a 
particular version of a game will be important to collect since it is the first of its kind, even if 
it was never widely disseminated. Adventure from our case set is just such an example. The 
version of the game that propagated over the nascent ARPANET, and then to nearly every 
mainframe and personal computer system in the industry thereafter, was not the ?original? 
programmed by Will Crowther, but rather the expanded version of the game released a year 
or so later by Don Woods. Anyone who remembers playing Adventure on their first computer 
almost certainly is recollecting the expanded Woods version, not the Crowther original. Yet 
Crowther?s source code has recently been recovered, is now available on the Internet, and 
has been implemented in a variety of different virtual machines. Which version of Adventure 
does a collecting institution want then, the Crowther original or the version modified and 
expanded by Woods that proved so influential? The answer, of course, is probably both, but 
recognizing the difference may not be within the capability of a cataloger or archivist who 
does not specialize in games research. 
 
Much like an upside down stamp, some versions of games are significant because they 
contain bugs, so-called Easter eggs, or other anomalies. Sometimes the question of version 
will be dictated by access: an institution may reach a decision to collect only PC versions of 
games, because that corresponds with the vintage hardware they have on hand. Sometimes 
version may be dictated merely by happenstance: what comes into an archivist?s hands as the 
result of an acquisition or what is available that day on eBay. This was precisely the method 
by which we obtained our Macintosh-compatible copy of Mindwheel, a format not even listed 
in the Moby Games entry for the game (http://www.mobygames.com/game/mindwheel). 
(It was left on the doorstep of one of the project members, who narrowly avoided stepping 
on it on his way out.) 
 
We have done considerable work with FRBR as a framework for mapping relationships 
between different versions, ports, and storage media for games, with Adventure serving as our 
focal point (McDonough et al., 2010). While librarians have long recognized the distinction 
between a work as an intellectual creation and its embodiment within a particular physical 
form (and the need to adequately describe both), the publication of the Functional 
Requirements for Bibliographic Records Final Report by the IFLA Study Group on the 
Functional Requirements for Bibliographic Records (FRBR) marked a pronounced increase 
in the level of attention that the library community has devoted to these issues. FRBR 
proposed a formal model for bibliographic description that recognizes four classes of entities 
as implicated in descriptive practice: Works (unique intellectual or artistic 
creations), Expressions (the realization of Works), Manifestations (the physical 
? 26?
 embodiment of particular Expressions), and Items (single exemplars of a Manifestation). 
Attributes commonly found in bibliographic description, such as publisher or title, are 
bound in the FRBR model to one of these four entities. 
 
At first it might seem that all versions of Adventure should be grouped under a single 
?Work,? a particular instance of the game (the last version modified by Don Woods, for 
instance) should be the ?Expression,? a particular file with a unique MD5 hash should be the 
?Manifestation,? and an individual copy of that file (perhaps on a Commodore 64 664 Block 
disk) would be the ?Item.? But what if the text read by the reader is exactly the same, but the 
underlying code is different? These variants might be simple (a comment added to the 
FORTRAN source code), peripheral (such as the ability to recognize ?x? as a synonym for 
the command ?examine?), or very large (a port of the code from FORTRAN to BASIC). 
Should these code level variants be considered different expressions? To further complicate 
matters, what if the FORTRAN code was exactly the same but compiled to two different 
chips? For example, an IBM mainframe and a Commodore 64 might both have a 
FORTRAN compiler, but the two compilers will interpret the FORTRAN to a different set 
of set machine instructions. It might also be the case that two FORTRAN compilers 
designed by different programmers will generate slightly different machine language. Even 
the same compiler might generate slightly different machine code from a single source code 
file depending on the options with which it is invoked. Should these compiled executables, 
different in their binary structure but based on the same FORTRAN code, represent 
different ?Manifestations? or different ?Expressions?? 
 
Finally, even two files with exactly the same MD5 signature participate in a larger software 
environment at runtime. The drivers that run the display interface, the keyboard, the 
memory, and the disk drives arguably become part of Adventure when the user is playing the 
game. For instance, the experience of playing the game using the 6507 chip in a Commodore 
64 hooked up to a black and white television may be different than the experience of playing 
the game on the same chip in a Commodore SX64 (the all-in-one machine some felt fit to 
call ?portable?). Should the software environment on which the binary is executed be a part 
of the classification scheme at all?  Would playing the game on a video monitor (which 
displays only a fixed number of lines at a time) provide a substantially different experience 
from a session with the same game played on a Teletype (which saves the output indefinitely 
on paper)?  
 
Such questions are reminiscent of issues and debates that have long been debated in more 
traditional arenas for cultural heritage, such as literature and textual scholarship. ?The root 
cause of any version is revision,? the Melville scholar and editor John Bryant (2002, p. 70) 
has contended. But what is the threshold by which mere variants in a text promote the text 
to a new version of the work? Bryant delineates the following characteristics: that versions 
may be either physical (literally distinct documents) or inferred from the evidence available 
on one or more extant documents; that a version of a work can always be linked to another 
version; that versions are revisions of the work, and that they may be initiated, either 
deliberately or inadvertently, by the author or some other agency like an editor or the public 
at large; that versions do not depend on the authority of the author to be authorized as 
versions; that versions are manifest by their degree of difference from other versions; that 
versions entail some reconceptualization or reimagining of the work in question; that 
versions are partly defined by their impact on their audiences; and finally, that versions are 
? 27?
 constructed critically by virtue of other artifacts in the documentary field, meaning that the 
existence of a version is always finally arguable. If we step back from this delineation, what 
we have then is an account of versioning wherein a version itself is always critical, 
contingent, relative or relational, functional, rhetorical, and perhaps above all, consequential. 
This seems no less true for digital games and virtual worlds than for the literary and 
historical documents that concern Bryant. A version is not an inferior derivative of some 
legitimized and sanctioned work, nor is it simply a collection of observable variants. Versions 
are, instead (to paraphrase Gregory Bateson) the differences that make a difference. 
 
Our work with Adventure and other digital games has highlighted two deficiencies in the 
FRBR entity-relationship model. The first problem arises from the complex tangle of 
derivative works associated with any particular game. Neither catalogs as they exist today nor 
FRBR provide sufficient facilities to ease collocation of these works for users. Computer 
games provide one of the stronger arguments for the concept of a superwork and adding 
support for superworks to our bibliographic systems, and to the FRBR model. The second 
problem is the omission of any mention of intellectual property rights within the FRBR 
model. While the IFLA Study Group (1997) made it clear that they were not enumerating 
every attribute of or relationship existing between bibliographic entities, the failure to 
account for intellectual property relationships between Group 1 and Group 2 entities is 
extremely problematic for those attempting to describe computer games, and we suspect 
much other digital material. Alignment of legal theory and cataloging theory regarding the 
separation between artistic/intellectual creations and their expression in particular forms is, 
we suspect, a difficult task that will require the input of the both communities. 
 
In sum, there are no easy formulas or checklists to aid in decision-making regarding which 
versions of a digital game or virtual world to collect. There is no one rule of thumb or 
absolute heuristic. Bryant?s criteria, while compelling, are likely to remain resistant to formal 
codification. Unraveling the relationships between even the simplest digital works and their 
constituent elements is orders of magnitude more challenging than for most artifacts in the 
analog world. That said, these same challenges lend themselves well to visualization and 
mapping techniques, and one can imagine catalogers and archivists working with increasingly 
sophisticated tools to articulate and display the relationships between digital objects in their 
collections. All of this also requires considerable subject knowledge, and here the input of 
both game scholars and fan communities is likely to be invaluable. 
 
Context Information Reconsidered 
 
Unlike many creative works, video games and virtual worlds are rarely able to stand on their 
own as time progresses and technological advancements render them obsolete. That which 
was startlingly innovative in the early 1980s appears simplistic to the modern eye. Take, for 
example, Mindwheel (1984) and Fallout 3 (2008), two apocalyptic adventures played from a 
first-person perspective. 
 
? 28?
  
(Pinksy & Synapse Software Corporation, 1984) 
 
 
(Bethesda  Softworks, 2008) 
 
Fallout clearly has superior graphics?Mindwheel has none and, even if it did, the 16 color 
graphics of its day would not be competitive. The difference in sound quality is similarly 
? 29?
 dramatic. Traversing the world of Fallout 3 is an instantaneous experience, interrupted only 
by loading screens for significant changes in setting. In the screenshot above, nearly 20 
seconds elapsed between entering the command, ?DOCTOR, BEGIN GAME? and the text 
fully loading. In its simplicity, Mindwheel?s data barely hit megabytes, while Fallout 3 occupies 
several gigabytes.  
 
Games and interactive fiction are not the only works that suffer when viewed after their 
time. Virtual worlds?from Second Life to Neopets to World of Warcraft?are not experienced in 
a vacuum. The value and meaning of a virtual world is primarily derived from the actions 
and interactions of its players. Imagine stepping into Second Life, which doesn?t even have the 
benefit of plotlines or non-player characters (NPCs), years after the last user signed off. The 
world would be empty; interactivity limited to the virtual equivalent of archaeology: 
examining buildings and prims in an attempt to build a picture of how Second Life was lived. 
 
So how do we prove the historical significance of Mindwheel or other early video games to a 
modern gamer? By spending as much effort preserving the context of gameplay as the 
software that enables it. 
 
Literature & Ephemera 
 
As with any type of history, contemporary publications play a large role in providing an 
adequate context in which to understand video games. The most relevant of these are 
documents such as game manuals and press releases, and third-party publications including 
strategy guides, reviews, and interviews.  
 
A video game?s retail package is rarely limited to a copy of the software, especially if it was 
released before the advent of downloadable content and sophisticated digital rights 
management (DRM) technology. The box?a valuable artifact itself?typically contains an 
instruction manual, registration card, and advertisement or catalog at a minimum. Depending 
on the game, maps, control cheat sheets, posters, and comics or other game related fiction 
might also be included. Mindwheel was bundled with a 93-page novella. Many older computer 
games required information from these physical materials as a way to prevent illegal copies 
from being played. 
 
The materials bundled with a game become 
particularly important if the game itself is no 
longer playable. Manuals include introductions 
to plot and characters, screenshots, credits, 
and explanations of controls. These controller 
instructions can provide some limited insight 
into gameplay. For instance, in Mortal Kombat 
II, the character Shang Tsung can transform 
into any of the other player characters with 
specific sequences of one to four button 
presses (or holds). This type of complex 
button combination for special moves is as 
intrinsic a part of the fighting genre as the 
? 30?
 hardboiled detective is in film noir. If nothing else, this may aid with understanding ?button 
mashing? in a time when games are increasingly controlled by motions instead of joysticks. In 
a similar fashion, the keyboard card packaged with Eidos? Thief: The Dark Project illustrates 
the level of thought and detail put into the game?s design as well as the difficult controls 
required by such nuance. When text adventures were the games of the day, players often 
hand drew maps to keep track of their location; in the golden age of the 16-bit role playing 
game (RPG), world maps became common pack-ins. 
 
Strategy guides, generally published by third parties, provide a much richer picture of game 
play than can be gleaned from the materials a game is packaged with. By their nature, these 
books offer detailed descriptions of game play from beginning to end, and often pair 
screenshots with the guide text. Strategy guide content may include: 
?
 ? Screenshots 
? Maps 
? Concept art 
? Tips for mastering difficult game controls 
? Puzzle solutions 
? Lists of in game items, their purpose, and where to find them 
? Cheats, secrets, and Easter eggs 
?
 The last item on the list might be viewed as a type of game developer marginalia (or graffiti). 
Easter eggs are messages, events, or effects hidden by game developers that are typically not 
encountered during regular game play, or require inside knowledge to catch, if they are.  
They often contain programmer signatures, homages to earlier titles, advertisements for 
upcoming games, or genre in-jokes. Sierra On-Line was notorious for using Easter eggs as 
marketing tools. In particular, the various Quest series all contained references to each other; 
King?s Quest II hides a preview of Space Quest I in a snake home and a guard in Space Quest II 
asks if you've played King?s Quest (?Easter Eggs and cheats,? n.d.). DOOM has a well-known 
cheat code (?idkfa?) which, when typed during game play, gives the player every weapon, full 
ammo, every key, and full armor. In Heretic, another id Software title, typing this code 
removes all acquired weapons and calls the player a cheater (?Easter egg,? 2010). 
 
Promotional materials released by game developers with the launch of new titles can provide 
important insight into the contemporary state of video game technology. While marketing 
claims are likely to be exaggerated, they are unlikely to be outrageously overblown. Thus 
given the fact that both Blizzard?s 2010 press release for StarCraft II and Activision?s 1980 
promo reel emphasize graphics, it can be assumed that each of the games is fairly 
representative of the state of graphics technology at the time of its creation. This is the type 
of detail that helps with the Mindwheel vs. Fallout scenario presented above; with it, a modern 
researcher unfamiliar with the timeline of video game advancement gains a better 
understanding of the quality of a game at release. 
 
As with literature, game ephemera may include items from both the game developer and 
licensed third parties. These tie-in items range widely in form and material, from ordinary 
posters and patches to pocket lint and potions. The table below illustrates some of this 
variety.??
? 31?
 ?
 Item Manufacturer Year 
Small?Centipede?Poster? Atari?? 1982?
 Dragon's?Lair?Lunchbox?and?Thermos? Aladdin?Industries?Inc?? 1983?
 Q*Bert?Miniature?Figures?? Kenner? 1983?
 Mindwheel?novel? Synapse?and?Pinsky?? 1984?
 Packet?of?pocket?fluff?(Packaged?with?
 the?Hitchhiker's?Guide?to?the?Galaxy)?
 Infocom?? 1984?
 Nintendo?Cereal?System?? Ralston?Cereals?? 1988?
 Breath?of?Fire?Map? Capcom? ?1993?
 Chrono?Trigger?Key?Chains? Bandai? ?1995?
 The?Book?of?Atrus?(First?in?a?series?of?
 novels?inspired?by?Myst)?
 Miller?? 1997?
 Trance?Vibrator?(Packaged?with?Rez)? Sony?Computer?Entertainment?? 2001?
 DOOM:?The?Boardgame? Fantasy?Flight?Games?? 2004?
 Yoshi?Tag?and?Run?Meter?? McDonald?s?Happy?Meal?toy? ?2006?
 Master?Chief?Costume? Rubies?Costume?Co? ??2008?
 Monster?Hunter?Health?Drink? Bandai?? 2008?
 Sonic?the?Hedgehog?Plush?Backpack?? Hot?Topic?? 2009??
 ?????Key:? Developer?Third?Party?[Plain]?
 ? Retail?Package?[Bold]?
 ? Limited?Edition?[Italic]?
 ? Nostalgic?[Bold?Italic]?
  
The meaning constructed by these items is as diverse as their physical forms: 
 
? The lunch kit, Happy Meal toy, and cereal indicate popularity with school-age 
children. 
? The map implies elaborate world building (for the time). 
? The pocket fluff appealed to the humors of Douglas Adams fans. 
? The Myst novels represent a reversal of the usual path of adaptation; book to game 
(as was the case with The Hitchhiker?s Guide to the Galaxy) is much more common. 
? The board game could imply either outreach to the wider (non-video) gaming 
community, or greater than average popularity among tabletop gamers. 
? The costume speaks to the devotion of the games fans, especially when accompanied 
by photographs of those fans dressed up for video game or genre themed 
conventions. 
? 32?
 ? Nostalgic items like the Sonic backpack, faux vintage t-shirts, and rereleased classic 
games speak to the lasting influence upon American culture. 
 
In addition to these textual and artifactual materials, soundtracks, movies, television series, 
audio recordings, and orchestral performances have all emerged from popular video game 
franchises. While the specific details of ephemera are important for contextualizing games, 
that some varieties exist at all is a valuable appraisal aid. While any game might be sold with a 
poster or figurine, only the leading titles are likely to inspire entire lines of action figures or 
musical arrangements. Certainly, there are many factors other than simple popularity that 
contribute to a game?s enduring value, and not all titles of import come with a large market 
impact. But, as a history of video games would be incomplete without the blockbusters, 
some decisions for a large repository are simple. 
? 33?
 4. Collections 
 
Collections: Bibliographic & Archival Description 
 
Bibliographic records are one of the primary tools that any library has for management of 
collections. As has been noted with regards to moving image materials (McDonough & 
Jimenez, 2007), however, catalog records that may serve perfectly adequately as mechanisms 
to aid access often fail as mechanisms to aid in preservation. In the case of computer games 
and interactive fiction, traditional bibliographic description may not even prove adequate for 
purposes of access. 
    
Our project focused on research collections of games, such as those at Stanford University, 
when considering issues of description for access, and our examination of current practices 
of bibliographic description was performed with an assumption that our main interest was in 
determining whether they adequately supported the research activities of scholars concerned 
with games and related materials. We note that scholars? interests in these materials should 
not be considered uniform. The needs of scholars working within a program such as the 
Department of Interactive Games & Media at RIT, which is focused on game design and 
development, are not identical to those of scholars working within the digital humanities 
who might be interested in analyzing the rhetorical nature of games (Bogost, 2007). 
 
Another point which is often neglected in library and information science literature is that 
bibliographic information does not only support the work of patrons; it must also support 
the work of librarians and archivists. Coyle (2004) has examined the ways in which the 
current discussions of the FRBR final report and its implications for bibliographic record 
design have had an unfortunately narrow view of the functions which should be included in 
?Functional Requirements,? particularly with respect to libraries? internal operating 
procedures. When examining forms of bibliographic description from the point of view of 
preservationists, we have tried to keep in mind that description must support the work of 
information professionals as well as that of the research community. 
 
A significant problem with existing practices of bibliographic and archival description from 
the point of view of scholars interested in computer games and interactive fiction is the lack 
of detail provided by existing formats and descriptive practices, particularly with respect to 
the issue of versioning and editions of particular games. Consider this record from the 
Library of Congress? online catalog: 
 
? 34?
  
 
This record does provide some information regarding the version of the game, inasmuch as 
it indicates in the title that it contains ?Episode 1? of DOOM and that the CD-ROM in 
question apparently dates from November 1994. However, by November 1994, DOOM was 
already in version 1.7a; in fact versions 1.2 through 1.7a were all released during the first 
eleven months of 1994, and there are significant differences in game play, technical support 
and capabilities between the different versions. There is also nothing in this record to 
indicate whether this CD-ROM contains a copy of the shareware version of DOOM, or one 
of the commercial releases. If the shareware version, it would only contain Episode 1, but 
the mail order releases of the original DOOM contained further episodes. For a scholar 
researching games, this record is singularly unhelpful with respect to the version of the game 
contained with the Library?s collection. 
 
This level of detail, unfortunately, counts as the epitome of service when compared against 
traditional descriptive practice found in archival finding aids, where item-level description is 
minimal at best. The finding aid of the Stephen M. Cabrinety Collection in the History of 
Microcomputing at Stanford University Libraries provides the following information for 
anyone looking for the game Star Raiders: 
 
? 35?
  Box 136 Atari, Inc. ST Star Raiders, 1986 
   Physical Description: 1 computer disk ; 5 ? in. Atari 
 
 Box 72 Atari, Inc. Star Raiders, 1982 
   Physical Description: computer cartridge Atari 
 
 Box 134 Atari, Inc. Star Raiders, 1982 
   Physical Description: computer cartridge Atari 
 
 Box 142 Atari, Inc. Star Raiders, 1980 
   Physical Description: computer cartridge Atari 
 
 Box 152 Atari, Inc. Star Raiders, 1982 
   Physical Description: computer cartridge Atari 
 
Presumably we can depend on the fact that Box 136 contains the version of Star Raiders for 
the Atari ST, but as for the rest, whether the cartridge in Box 142 is intended for an Atari 
2600 system or one of the Atari 400/800 systems is unclear, and for the remaining three 
boxes, the cartridges could be in theory be for the Atari 400/800, 2600 or 5200 systems. And 
only a scholar with knowledge of the exact release dates for the various editions and 
platforms would be able to deduce that information from this finding aid. 
 
As discussed in Chapter 3, our project did extensive exploration of the application of the 
FRBR data model to game materials, and found that it provides reasonable if not perfect 
support for identification of versions and editions of games. However, FRBR in and of itself 
is not a solution to the problems of adequate description of game materials. As analysis of 
the current MARC and Anglo-American Cataloging Rules (Delsey, 2002) shows that the 
bibliographic format and rules of description employed by the library community already 
provide support for much of what?s contained within the FRBR Final Report. What is 
needed is a commitment to providing an enhanced level of description necessary to support 
scholarly work. While clearly there are painful financial ramifications for any library (and 
even more painful for any archive) trying to create more detailed bibliographic descriptions 
of these materials, existing practices are not adequate to meet the needs of the scholarly 
community. 
 
Nor do these descriptions provide anything like the information that would needed by 
preservationists, for whom the details of version and edition information are equally as vital. 
The cartridges employed by varying Atari systems (e.g., 400/800, 2600, 5200) differed in the 
pin-out arrangement used by the cartridge slots to physically interface the cartridge with the 
rest of the system. While Atari systems are still available on the secondary market today, 
eventually these systems will cease to function, and any ability to access the information on 
such a cartridge will be dependent on our ability to read the data off the cartridge and move 
it into some form of emulator. Without sufficiently detailed knowledge of what types of 
cartridges are in hand, a library or archive would not even know what additional technical 
information they might need to collect in order to try to insure long-term access. While the 
bibliographic record for DOOM above indicates that DOS is required to run the software it 
fails to provide any indication of which version of DOS is necessary.2  Long-term access to 
software requires detailed knowledge of the technical environment in which it was designed 
????????????????????????????????????????????????????????
 2?The original DOOM required MS-DOS version 3.3 or higher. 
? 36?
 to execute. Current descriptive practices do not provide anything resembling sufficient 
information on the technical aspects of software to support preservation activity. 
 
One of our critical findings regarding the preservation of gaming materials is that 
preservation of a game itself is insufficient; we need to also preserve the information that 
contextualizes the game and helps researchers achieve a more complete understanding of the 
game?s significance and use. Much of the contextualizing material that we might wish to 
preserve, however, is archival in nature, existing in relatively few copies. The UT Videogame 
Archive at the Dolph Briscoe Center for American History at UT Austin, for example, has 
papers and files donated by Warren Spector, a game designer famous for his contributions to 
the Ultima series of games from Origin Systems, Inc. and Wing Commander. These materials 
are unique, one-of-a-kind contributions, which is to say, the type of material archives 
specializing in personal papers and manuscripts acquire on a regular basis. A scholar at 
Stanford University looking at the Ultima games contained within the Stephen M. Cabrinety 
collection would probably be extremely interested in the material held at UT Austin, but 
bibliographic links between collections held by different institutions are so rare as to count 
as non-existent. Preservation demands that libraries, archives and museums begin to rethink 
their practices of bibliographic description and consider the possibility that catalogs and 
finding aids should help users obtain the information they need regardless of whether it 
happens to be housed at the current institution. The fact that a catalog provides a valuable 
service as an inventory of an institution?s holdings does not mean that that is the only 
purpose it should serve. 
 
Problems with linking bibliographic information with databases maintained by other 
organizations are not limited to the realm of descriptive metadata. One of the basic tenets of 
the Open Archival Information System Reference Model is that an archival information 
package for any digital object should contain representation information that allows a user to 
interpret the bits comprising an object as syntactically valid and semantically meaningful 
information. As much of this information will be identical for different objects of the same 
data format, many institutions are hoping that institutions such as the developing Unified 
Digital Format Registry will become hosts for representation information needed by other 
organizations. But if this is to occur for games, we need to be able to link the individual files 
comprising a game stored within an institution?s repository systems with representation 
information stored elsewhere. This will require the addition of linking information that 
libraries, archives and museums have never had to store previously, as well as the use of 
standards for linking within a distributed system of repositories, standards that are at best 
inchoate at the moment. 
 
As the above discussion should make clear, problems with descriptive metadata created 
within libraries, archives and museums today are not a matter of existing metadata being 
erroneous; rather it is simply insufficient. Given the increasing costs of generating metadata 
manually, this is clearly problematic. While some of the forms of metadata described above 
(e.g., links to representation information) will require decisions by professionals with 
significant expertise within the preservation field, other forms (e.g., contextualizing 
information) could be provided by anyone with a strong interest and knowledge of games. 
Given the number of individuals answering to that description, for games at least, cultural 
memory organizations should explore how they might enlist the aid of the communities they 
serve in the creation of metadata necessary to preserve these objects. 
? 37?
  
Collections: Data and Documentation 
 
Collections of digital games and virtual worlds are, to a large extent, collections of software. 
Many of the preservation issues faced by repositories holding these collections are therefore 
problems of software preservation. However, software preservation only addresses one 
aspect of digital games and virtual worlds; focusing on software preservation was an 
important aspect of the PVW project, but other equally challenging problems emerged and 
were addressed by the project investigators. 
 
It is important to emphasize here that virtual worlds are historical in two senses of the 
phrase: They are worlds of historical interest, and they are going to go away. In PVW, we 
explored some of the implications of the life-cycle of virtual worlds, especially of their 
extinction, for thinking about how the history of computer-based ?worlds,? as well as their 
use by communities of players or ?residents,? could be documented. The moment when a 
virtual world ?is history??when it shuts down?reminds us that every virtual world has a 
history. These histories of individual virtual worlds are inextricably bound up with the 
intellectual and cultural history of virtual world technologies and communities. They are also 
venues for historically specific events and activities. An important part of the historical 
context for virtual world history is the fact that human beings (through their avatars) fill 
these digital environments with meaning that emerges from their activities in social spaces, 
regardless of whether the spaces are synthetic (digital) or physical. So, in addition to 
preserving the software that provides the technical underpinnings for virtual worlds, our 
project faced the problem of how to identify, collect and preserve documents that convey 
events and activities that take place in virtual spaces. 
 
Perfect Capture 
 
Thinking of a game world or a virtual world as historical brings us directly to issues of 
historical documentation, digital preservation and curation of virtual worlds. What will 
remain of virtual worlds after they close down, either individually or perhaps even 
collectively, i.e., when the technology has become pass?? 
 
A notion that plays into the preservation discussion is particularly relevant here: namely, that 
of the potentially perfect reproduction of digital data. Recall that our digital personae, our 
avatars, and our player characters are ultimately all bits of data on a machine. Death switches 
count on that. If we can only get access to these data, shouldn?t it be possible to copy them 
forever? 
 
Consider an actual historical case, that of Chris Crosby, a.k.a. NoSkill. Crosby was the first 
of the highly skilled players of the on-line multiplayer game DOOM to be recognized as a 
?Doomgod.? An active player from about 1994 to 1996, the young father was killed in a car 
crash in 2001. His memorial site on the web, like many others, depicts him in the prime of 
life, holding his young son, but it also offers a number of files for downloading (NoSkill 
Memorial Site, 2004). These files are demos recorded from games he played between May 
1995 and April 1996. A demo, also called an lmp (from its .lmp or ?lump? file extension), is 
a replay file. It is a recording of a game session in the form of a sequence of commands that 
correspond to input control states during each frame of the game, or ?tic.? DOOM players 
? 38?
 could generate a demo file by simply entering the command ?-record? in the console, a 
command-line interface that could be called up while playing the game. In other words, they 
could create a script?a sequence of instructions?generated from game data and save it as a 
demo recording. A recording in this format is much more compact than video captured 
from the screen. The catch is that the demo data must be run and executed inside a copy of 
the same game from which it was generated, and even from the exact same version of that 
game, if the game engine is to render the action correctly. 
 
It is easy enough to download Chris Crosby?s demo files from his memorial site and play 
these files inside the correct version of this old game, originally published towards the end of 
1993. Indeed, DOOM is in the collection of digital games that have been preserved as part of 
PVW, so that several versions of the game, including the version Crosby played, will be 
available in the Stanford Digital Repository, for example. Playing Crosby?s replay files is a 
profound experience. It means nothing less than experiencing a dead (historical) game 
through the eyes of a dead player, that is, seeing the game as he saw it. NoSkill in this sense 
comes back to life, as the replay file activates the game engine to carry out the exact 
sequence of actions enacted by the now dead player. Again, in this first-person shooter, we 
can see the game action through NoSkill?s eyes. The player is dead, but it is now possible for 
his avatar in some sense to live on through an act of perfect reproduction.  
 
Historians will be unable to help but contrast the potentially infinite repetition and perfect 
reproduction of NoSkill?s game-play to the fading memories of his life ? and death. At the 
same time that we are powerfully affected by revisiting a past experience, we know that what 
we are seeing may be an historical event, but without documentation it is not history. It is a 
remarkable act of software and data preservation, but it would be a mistake as we begin to 
stage early work on preservation of games and virtual worlds to frame these projects like 
PVW primarily, or even exclusively, in terms of software preservation and the perfect 
capture mode of game replays. This would be a barren exercise with respect to the 
documentation of the events and activities?the history?that has occurred in these worlds. 
 
Perfect Loss 
 
This is because future historians and others interested in the history of virtual worlds will not 
just want to experience what it was like to play a historical game or visit a world like Second 
Life, they will want to know much more about the things people were doing in virtual 
worlds, why they were doing them, and what their activities meant to them.  
 
So perfect event capture with respect to digital data is possible, and replay offers a paradigm 
for perfectly reproducing the past, even seeing through the eyes of players who are no longer 
with us. From a historian?s point-of-view, perfect capture is half of a paradox, for it must be 
placed alongside the very real possibility of ?perfect loss? in digital spaces. If we save every 
bit of a virtual world, its software and the data associated with it and stored on its servers, it 
may still be the case that we have completely lost the history. To date, the virtual world has 
not yet been produced that offers vestiges or traces of the past after transaction logs or 
content have been removed from it. When the data is gone, it?s gone. That is not the entire 
problem, however. 
? 39?
  
An example might help here. 
A few years ago, a series of 
nasty protests in the virtual 
world Second Life led to an 
attack on in-world buildings 
owned by the National Front. 
Reports about this clash 
appeared in blogs and forums 
visited by members of the 
Second Life community and 
others interested in virtual 
worlds not long after the 
events had occurred. After 
reading a witness?s account of 
the events, many readers 
jumped into Second Life to see 
what was going on, only to 
find there was absolutely 
nothing to see. The National 
Front had already abandoned its Island and deleted all of the content there, essentially 
stripping the turf of every trace and artifact. The ?world? revealed nothing of what had 
recently been a hotbed of activity and conflict. The only sources left were documents created 
and stored outside the virtual world itself ? blog entries, forum posts, screenshots, etc., on 
the 2D web. 
 
The historian Timothy Burke has described the difference between game-generated data and 
historical documentation in terms of what he calls the ?proprietary? data of virtual worlds, 
meaning the data that is owned, or present on the servers that support that world: ?... I think 
the one thing that *isn?t* in the proprietary data is the history of unusual or defining 
episodes or events in the life of particular virtual worlds ?  The narrative history, the event 
history, of any given virtual world, may in fact be obscured by the kinds of god?s-eye view 
data that developers have. After all, they often don?t know what is happening at the 
subjective level of experience within communities, or have to react to it after it?s happened. 
(Say, when players stage a protest.)? (Burke, 2006)  Thus, focusing on preservation of what 
Burke calls proprietary data matches up poorly to the likely needs of future scholars of 
virtual worlds. 
 
Consider another example that illustrates this point. In the first hours after the WTC and 
Pentagon attacks on 11 Sept 2001, online communities used systems such as massively 
multiplayer role-playing games as a medium for responding to the attacks. In games such as 
Everquest and Asheron?s Call, players read news alerts either via in-game text or system 
announcement, while outside the world but still on-line, other players caught up via player 
community websites. Of course, others watched television, heard from friends, or even 
experienced the events up-close and personal. Within hours, players organized candlelight 
vigils for the victims of the attacks, using glowing weapons or other objects, taking 
screenshots and posting online to document their in-world activities and discuss what they 
meant in the context of the dramatic historical events unfolding around them. For a vigil 
? 40?
 held on Everquest?s Luclin server on 12 September in response to ?yesterdays [sic] 
disheartening display of events,? players were invited to ?mourn and discuss? on the 
Everlore website (Lowood, 2008). Players commented on the meaning of this action to 
them; one of them, with the player name Keeter, argued that, ?Just because you are in a 
game doesn?t mean the world outside doesn?t effect [sic] you. Many people would like to 
mourn and share peace along side [sic] people they have battled long and hard side by side 
with. Yes, I can go to a church to mourn, but I would like to do it with my comrades around 
the country/world, which is impossible everywhere else. If you don?t want to be a part of it, 
then dont [sic]. You can choose not to do it. But respect the people who would like to. We 
don?t bother people that want to run naked gnomes through the country, so don?t bother 
people that want to gather and discuss something important to us all.? Documentation such 
as this quotation is necessary both for a full description of an event and for a rich 
interpretation of what the activities associated with that event meant to participants, no 
matter what kind of world we are talking about. 
 
There are three important points here with regard to a preservation project focused on 
software and game-produced data. First, inside Everquest today there is no trace of these 
events. Assuming that the game world has not been deleted, erased or remade (untrue), that 
we are on the right server or shard of the game, and that we are standing on precisely the 
spot where such a vigil occurred, it is generally not possible to dig beneath the surface, 
scratch underneath a poster, or find a file cabinet of documents or an old newspaper in a 
nearby building. There are exceptions, such as a monument on a specific Asheron?s Call server 
that commemorated a unique achievement by its players, but such exceptions are rare.  
 
Second, this lack of in-world artifacts and documentation clearly has implications for long-
 term preservation that focuses on game software and server-side data. Assume that we are 
able to capture every bit from a virtual world server, everything from 3D models to account 
information, that we are able to reverse engineer or disable authentication and log-in 
controls after the original server is no longer live, and that we have received permission from 
every rights holder ranging from game developers to third-party developers and players to 
copy, store, and use what they created, show their avatar, or reveal their identity and 
activities. The chances of all this actually happening are near zero, of course, but assume that 
it all could be done. Then assume that we can sync up every state or version of the software 
to the matching states of databases. It might then be possible to run a simulation of the 
virtual world as an archival time-machine, flying around on a magic carpet in spectator mode 
but never interacting with events run by the game engine and player data, much like a game 
replay. Turn the dial to 12 September 2001, and you might find a group of players standing 
around with brightly colored weapons and wands in their hands. But what are they doing, 
and what does it mean to them?  
 
The third point then is that the documentation that is a prerequisite for future historical 
studies of virtual worlds may not be located on game servers at all. The most important 
qualitative documentation may be somewhere else, on a blog or a wiki, in a player-created 
database or Flickr screenshots, or a YouTube video. The same may be true for some of the 
contextual information desired under whatever set of transfer protocols or preservation 
specifications a project is using, even sometimes for technical aspects such as software 
dependencies or relationships among objects. Game researchers such as Dmitri Williams 
have extracted and analyzed a wealth of quantitative data from virtual worlds; they have used 
? 41?
 these empirical data to explore social and economic aspects of these worlds. Such research 
rarely has access to server-side data, but instead relies generally on surveys, participant 
observation or data harvested on the client side using bots or automated characters. 
(Williams et al., 2006). The point is that writing the history of virtual worlds on the basis of 
software and of associated data alone would be a barren exercise. Installing Everquest in 2050 
will not reveal much about the virtual world that emerged from the software, even if future 
writers and historians have access to everything needed to run a fully functioning version of 
the game. Certainly, there are still important reasons for preserving this software, whether as 
artistic or cultural content, for technology studies, or for forms of scholarship that treat 
aspects of digital games and virtual worlds as authored texts or artistic objects. Still, we need 
to think more about virtual world history in terms of events and activities, much as an 
archivist or historian would in the real world, and attend more carefully to preservation of 
forms of documentation in digital form that are external to virtual worlds as software 
environments. 
 
Implications for Collections 
 
One way of characterizing the mix of software preservation and documentation activities 
necessary to preserve virtual world content and history is to think of the primary collections 
as a mix of ?library? (published digital games and virtual world content) and ?archives? 
(documentation about game/virtual worlds). This is, of course, a familiar model for cultural 
repositories such as libraries. However, the nature of digital games and virtual worlds as 
software objects and the particular ways in which events and experiences in these virtual 
spaces are mediated and reported carry significant implications for collecting strategies as 
well as preservation. 
 
It is high time to review a few of the issues that have surfaced in the Preserving Virtual 
Worlds project with respect to collection and preservation of data and metadata. The first 
issue has already been set up in the discussion of virtual worlds as history. What exactly are 
we trying to preserve?   Specifically, is virtual world preservation focused on the software 
and server-side data that in some sense defines or encompasses the ?world? as a created 
artifact, or are we looking for materials in digital form that document the activities of players 
or ?residents? of these spaces?  There are at least two other ways to shape this distinction. 
The first is to separate developer-created or -managed materials from those created or 
managed by players. The second, as already noted, is to distinguish repositories of virtual 
world data as essentially libraries or museums of created artifacts or texts from archives of 
documentation about events. However, depicting virtual world preservation as a binary 
proposition such as developer vs. player or artifact vs. document is ultimately unnecessary 
and counter-productive. Certainly, there is value in preserving both software artifact and 
event histories. The most productive approaches to virtual world preservation will be those 
that integrate artifact and documentation in terms of collecting focus, evaluation of digital 
content objects, organization of content transfer packages, metadata creation and access 
strategies. 
? 42?
  
 
 
Consider a problem that might seem to be entirely a matter of treating the essential task of 
game preservation as software preservation, yet turns out to have crucial implications for 
documenting player behavior and history: software versions. In the Preserving Virtual 
Worlds project, we decided early on to limit our attention to a dozen or so representative 
case studies, rather than a comprehensive collection of software or data. One of our key 
cases has been the computer game DOOM, created by id Software. DOOM is a multiplayer, 
first-person shooter, and while now is not the time to delve into game history, suffice it to 
say that this game immediately transformed competitive, multiplayer gaming into the 
leading-edge genre for computer games through the 1990s. There are two other aspects of 
DOOM that are important with respect to preservation. First, it was distributed first and 
throughout its history in shareware versions that featured a limited number of episodes of 
the game; the idea was that the shareware version would hook players on the game, so they 
would then purchase the full version. Second, the developers of DOOM openly embraced 
modification of their software by the player community, and by implication revision of the 
notion of game authorship, which immediately de-stabilized the notion of a canonical 
version of the game. Defining a version of the game DOOM therefore involves considerable 
attention not only to a sequence of patches and versions, but also to combinations of 
developer-produced software, third-party add-ons and player-developed modifications, 
better known as mods. 
 
De-stabilizing the notion of a fixed version of software is not just a problem when we 
attempt to determine what versions to preserve and how to account for revisions and 
? 43?
 changes. It also raises significant issues that affect the documentation of player activities in 
the game. A crucially important category of objects that provide this documentation are 
produced by players? efforts to capture their experiences through replays, screen captures 
and screenshots. In the case of DOOM, we have already seen how NoSkill?s demo files make 
it possible to view the actions of one of this game?s earliest and best competitive players. 
Again, DOOM demos were essentially replay files, saved sequences of instructions from a 
previously played game that, when executed by the game software, would show the same 
game from the same (first-person) perspective of the original player. In other words, demos 
are recorded game play sessions that are played by the game engine. The same is true of 
replay files in later games, such as Blizzard?s Warcraft III, up to the present day. Unlike video 
files captured from the screen or video-card output, demos or replays allow different views 
and settings as permitted by the game software and the best visual quality that the software 
will produce. However, this all is possible only when a running version of the game engine is 
available in order to view these replays. Not only that, the version used to view the demo or 
replay nearly always must correspond exactly to the version that was played when it was 
created. Therefore, any decision about which version of the game will be preserved 
determines which replay or demo files will be viewable in the future. Likewise, any decision 
about which demos or replays are historically significant in terms of game culture or history 
will presuppose preservation of the appropriate version of the game software. Treatment of 
the software artifact affects documentation, and selection of documentation affects 
treatment of the software artifact. At least in the realm of virtual world or digital game 
history, separation of these treatment decisions into specialized areas or departments may 
lead to disastrous consequences for future archivists and historians. 
 
 
A second example with respect to virtual world data and metadata also speaks to the 
necessity of maintaining contact between collections and their contexts, as well as between 
projects of software preservation and historical documentation. This example suggests that 
documentation can also serve as a category of metadata for virtual world data. 
 
An interesting quality of virtual and game worlds is that many of them can be navigated by 
in-world coordinate systems, much like real-world cartography. Two well-known examples 
are the ?Second Life URL? (SLURL) in Second Life and the UI coordinate system in World of 
Warcraft. Just as we can mash up data by attaching GPS coordinates to real-world maps, 
photographs, and other media, these virtual world coordinate systems might make it possible 
to match documentation we have assembled in our virtual world collections not only to 
locations in virtual worlds, but also to each other. In the case of the video collection, 
however, the metadata scheme based on Dublin Core already provides the ?coverage? 
element for individual objects. As the Dublin Core specifications tell us, this element can be 
applied ?for the use of multiple classification schemes to further qualify the incoming 
information? such as latitude and longitude or other ?native coordinate representations.? 
(Becker, et al. 1997)   The Internet Archive?s Heritrix-based crawler allows collection 
curators to input metadata tags at the document level for web pages harvested through 
crawls; by adding a tag for a specific SLURL, say, to a document describing an event that 
occurred in Second Life, we can thus link documentation to in-world locations. 
? 44?
  
 
Using virtual world coordinates might help us bridge the gap between documentation and 
mute server-side software and data with respect to ?event history,? perhaps offering a 
solution to the problem of perfect loss in virtual world history. Here is a specific use 
scenario. A cultural historian is interested in the use of game worlds for scholarly 
communication and learns about the first science conference held in World of Warcraft, in 
May 2008 (Bohannon, 2008). She finds videos documenting this event in the Archiving 
Virtual Worlds collection, but they are a bit grainy and she is curious about the locations 
chosen for the event. So as part of her ?fieldwork,? she installs and fires up the game world. 
Then, using the coordinates conveniently provided by the collection metadata, she ports to 
the location where the conference was held and, using her avatar, walks the terrain depicted 
in the video. This scenario will work better in game worlds, where developers maintain 
relatively stable environments with respect to content, than in virtual worlds such as Second 
Life, where residents such as the National Front are free to delete everything they created. 
However, if we are able to maintain backups of content as part of a package of data 
associated with a virtual world, this problem will be alleviated. For now, it is sufficient to 
observe that, as in the case of DOOM demos, useful connections between documentation 
and data will only be available to historians and other researchers, if curators and archivists 
work closely along these lines with software preservation specialists. 
? 45?
  
Examples of Collections and Impact of Preserving Virtual Worlds? Collection-
 Building Efforts 
 
Currently, there are only a few significant library collections of historical game and virtual 
world software. These collections also include archival documentation, hardware artifacts 
(such as game consoles and realia), and print collections related to digital games. They 
include the Stephen M. Cabrinety Collection in the History of Microcomputing in the 
Stanford University Libraries (http://www.oac.cdlib.org/findaid/ark:/13030/kt529018f2), 
the International Center for the History of Electronic Games at the Strong National 
Museum of Play (http://www.icheg.org/) and the UT Videogame Archive at the Dolph 
Briscoe Center for American History, University of Texas at Austin 
(http://www.cah.utexas.edu/projects/videogamearchive/).  
 
The Cabrinety Collection was a resource for the Preserving Virtual Worlds project, as it is 
held by one of the project partners, the Stanford University Libraries. Stanford's How They 
Got Game was founded in 2000 to begin work on the history and preservation of digital 
games and interactive simulations. The founding of the project was stimulated by Stanford?s 
acquisition of the Cabrinety Collection three years earlier. Today, the collection remains 
perhaps the largest collection of microcomputer history held by a major cultural repository, 
with approximately 20,000 software titles, roughly 85% of which are digital games, some 75 
hardware platforms, publications, ephemera, and archival materials. How They Got Game 
continued earlier work in software history and archives carried out under the auspices of the 
Silicon Valley Archives at Stanford, and this activity is closely tied to related archival 
collections, such as the papers of Steve Meretzky and Hal Barwood, and records of the 73 
Easting Simulation and HPS Simulations. The linkage of library and archival collections is 
similar at both UT Austin and at the Strong Museum. 
 
As part of the Preserving Virtual Worlds project, the How They Got Game group at 
Stanford has created three collections to document virtual world events using largely player-
 generated content. The first is the Archiving Virtual Worlds collection 
(http://www.archive.org/details/virtual_worlds) hosted by the Internet Archive as part of 
their Moving Image Collections; the second is the Machinima Archive 
(http://www.archive.org/details/machinima), a curated collection of the emerging medium 
of game-based moviemaking known as machinima; and the third is a curated collection of 
websites related to digital games and virtual worlds hosted by the Internet Archives Archive-
 It service (http://archive-it.org/). The Archiving Virtual Worlds collection is particularly 
relevant to the discussion of documentation. It consists in large part of video footage made 
with real-time screen capture tools such as Beepa?s Fraps. The ?Final Countdown? video 
that documented the last minutes of EA-Land is an example of the content preserved in this 
collection. The Machinima Archive is dedicated to the academic investigation and historical 
preservation of the emerging art form known as machinima. Machinima is filmmaking 
within real-time, 3D virtual environments such as games and virtual worlds, usually made 
from existing video game engines, whether through editing of replay files (?demo?), video 
captured from the screen or graphics card, or composited game assets. Machinima ?capture? 
of in-game performance, sometimes called ?virtual puppeteering,? provides a model for the 
creation of high-quality video that documents player activities inside a game or virtual world. 
Last but not least, the How They Got Game project established its subscription with the 
? 46?
 Internet Archive?s Archive-It service in early 2008 for the purpose of crawling, saving, and 
making searchable numerous game- and virtual world-related websites. In two Archive-It 
collections (?digital games? and ?virtual worlds?), we have seen to the preservation of 
videos, weblogs, wikis, player-created websites, maps and many other forms of 
documentation that provide information about player activities. These activities might 
include modifying game software, demonstrating skills through superior game-play, 
commenting on events such as protests or artistic performances, or anything else that 
someone who has spent time in a virtual world might consider important. A particular 
emphasis of the documentation activity undertaken through Archive-It has been the 
preservation of web-based collections of player-created maps, walkthroughs, exploit videos, 
glitch hunting, and other activities that deeply explore in-game and in-world spaces. 
 
Compared to institutional collections of digital games and design archives, collections of 
player-created documentation relating to activities and events in game and virtual worlds 
have only begun to emerge in similar institutional contexts backed by preservation solutions. 
On the other hand, vast troves of such documentation maintained by individual collectors 
and players, fans and fan groups, and projects (such as MobyGames) have been assembled 
and maintained. The Archiving Virtual Worlds collection, the Machinima Archive, and the 
How They Got Game Project?s Archive-It collections will all be preserved as part of the 
Internet Archive?s collections, with backup in the Stanford Digital Repository. These 
collections complement and supplement the project?s efforts to work out schemes for long-
 term preservation of digital game software and assets by documenting in-game and in-world 
activities and events, player culture, and the efforts made by the player and, to some extent, 
the developer community to save, preserve and provide access to historical documentation 
about games and their cultural impact as well as game content. 
 
The Contributions of Game Communities 
 
Prior to recent efforts such as those of the institutional repositories mentioned above and 
the Preserving Virtual Worlds project, it is fair to say that nearly all of the significant efforts 
towards collection, preservation, documentation, enunciation and emulation of game 
content, technology and culture were due to the work of players and fans, or individuals with 
experience in the game industry. Examples of such efforts include the MAME community 
on the emulator front, the Software Preservation Society (founded as the Commodore 
Amiga Preservation Society) on the data migration front, or the work of collectors such as 
Frank Cifaldi?s ?Lost Levels? project and website (http://lostlevels.org/). 
 
One of the activities specified by Library of Congress in the award of the NDIIPP grant to 
the Preserving Virtual Worlds project was outreach. In the project, we emphasized two kinds 
of outreach with the specific goals of (1) encouraging better contact and communication 
between communities of players, developers and fans and cultural repositories working in 
the area of game and virtual world preservation; and (2) opening up channels for the 
exchange of information among these various groups and institutions. 
 
We believe it will be necessary to encourage more contact between the various communities 
and individuals with an interest in preserving the history of digital games and virtual worlds 
and cultural repositories such as museums and libraries building programs for game 
preservation. The disconnect between active collectors and programmers building software 
? 47?
 such as emulators is not unique to game preservation, but is a problem for computer history 
more generally. An important moment in awareness of the value in forging such contacts 
was provided, for example, by the ?The Attic & the Parlor: A Workshop on Software 
Collection, Preservation & Access,? which took place at the Computer History Museum in 
Mountain View, California, in May 2006. One of our project members (Curator Henry 
Lowood at Stanford) was on the organizing committee of this meeting and facilitated the 
proceedings. The meeting included library and museum curators, corporate archivists, and 
individuals active in collecting and preserving software, and highlighted the benefits to 
repositories of working with broader communities, both with respect to collecting and 
technology.  
 
Preserving Virtual Worlds contacted a number of game companies, collectors, and players 
throughout the course of our various activities. On the basis of these contacts, we were able 
to open up channels for exchanging information among these various parties. While we 
believe that much more work can be done in this area, we can point to two specific examples 
of our activity as possible models for future projects and programs. The first is our work 
with the International Game Developers Association (IGDA), specifically, the Game 
Preservation Special Interest Group (SIG). Quoting from its own website, ?The 
International Game Developers Association is the largest non-profit membership 
organization serving individuals who create video games. We bring together developers at 
conferences, in local chapters and in special interest groups to improve their lives and craft? 
(http://www.igda.org/). While the IGDA serves its mission in many ways, the special 
interest groups are the focus of member-based activities in a variety of areas, ranging from 
specific game design areas to quality of life concerns. The Preservation SIG serves as a 
?meta-resource, hub and community for those interested in digital game preservation and 
history.?  The SIG, with roughly 75 members, maintains an e-mail discussion that 
emphasizes game preservation related news and projects. It has sponsored events at the 
mammoth annual Game Developers Conference (such as the Digital Game Canon), where it 
also hosts an annual round-table for members and other interested parties, and aims 
?towards setting standards and guidelines for assisting industry/academia in establishing 
preservation efforts.? 
 
One of our project members (Lowood) served through the course of PVW as chair of this 
SIG. Reports on the status and results of the project were delivered to the group and 
discussed at the annual round-table meetings in 2008, 2009 and 2010. Moreover, several 
project members (e.g., McDonough, Donahue) joined the SIG and thus participated in on-
 line discussions of a variety of technical, game, and legal issues via the e-mail forum. In 2008, 
the SIG launched an effort to raise awareness among IGDA members of the importance of 
digital game preservation, both as an issue for the industry and as a matter of concern for 
individual members who collect games or for whom games are an important part of their 
lives and shared culture. A committee of the SIG began work on a two-part effort to 
produce easily distributed publications about game preservation. The first is a white paper 
describing the problem of the possible loss of game content and culture if no action is taken 
and the challenges facing preservation activities. The second would document best practices 
after the conclusion of the Preserving Virtual Worlds project. In March 2009, the first white 
paper was published under the title, Before It?s Too Late: A Digital Game Preservation White Paper, 
with publication support provided by the IGDA in the form of a grant to the SIG. It was 
announced at GDC 2009, where a limited number of copies were distributed; since that 
? 48?
 time, it has been available either as a free download or a modestly-priced print publication 
via Lulu Press (http://www.lulu.com). Before It?s Too Late includes contributions from PVW. 
Lowood edited the volume and Donahue distributed two surveys, ?Records Management in 
the Gaming Industry? and ?Preservation Activities in the Video Game User Community? as 
appendices to the white paper. It is anticipated that work will begin on the best practices 
document in 2011.  
 
A second major outreach activity of the project was the Play - Machinima - Law conference 
hosted by the Stanford University Law School and the How They Got Game group at 
Stanford. This two-day conference covered key legal issues associated with player-generated 
content. The specific focus was machinima, computer-animated cinema that is based on 3D 
game and virtual world environments, technology, and content. Issues of player-generated 
content play into many areas of concern to the Preserving Virtual Worlds project as a whole. 
The specific context of machinima production provided a lens for looking at these issues in 
detail, while also narrowing the scope of the conference to manageable proportions. 
Speakers included machinima artists/players, legal experts, commercial game developers, and 
game researchers; participants from Preserving Virtual Worlds included Lowood, Bittanti 
and Rojo (as organizers, moderators, and speakers), McDonough and Kraus (as speakers). 
Topics included game art, game hacking, open source and ?modding,? player/consumer-
 driven innovation, cultural/technology studies, fan culture, legal and business issues, 
transgressive play, game preservation, and notions of collaborative co-creation drawn from 
virtual worlds and online games. A particular focal point and the main topic of the second 
day was a document drafted by a team from the Stanford Law School as guidance 
concerning copyright law for machinima artists. With input from the conference and the 
IGDA Preservation SIG, that document will be produced as part of the Preserving Virtual 
Worlds effort and is discussed in more detail elsewhere in this report.  
 
The outreach efforts of the Preserving Virtual Worlds project, both with respect to the 
IGDA Preservation SIG and the Play - Machinima - Law conference, demonstrated the 
importance and potential gains from encouraging contact, discussion and collaboration 
among cultural repositories, the game industry, and the various communities with an interest 
in game cultures. 
 
Problems/Opportunities in Access  
 
Assume again that we are able to save all the software and server data for a virtual world, 
game world or simulation and put it into a digital repository in a library. When a historian 
fires up this virtual world fifty years from now, it will be empty. Moreover, the 
documentation that will be a pre-requisite for future historical studies of virtual worlds won?t 
be there. It never was. The most important qualitative documentation was somewhere else, 
on a blog or a wiki, in a player-created database such as Thottbot, in technical 
documentation, or Flickr screenshots, or a YouTube video. The same may be true when a 
future repository attempts to fire up some ancient software. If contextual information, such 
as software dependencies or descriptions of relationships among objects, is not provided 
under whatever set of transfer protocols or specifications a project is using?and all the 
retired software engineers are gone?it may be impossible to get old software to run. 
 
We suggested earlier in this report that nobody writing the history of virtual worlds will be 
? 49?
 able to do much with software and associated server-side data alone. Installing Everquest in 
2050 will not reveal much about the virtual world that emerged from the software, how it 
was built or used, even if future writers and historians have access to everything needed to 
run a fully functioning version of the game. Certainly, there are important reasons for 
preserving this software, whether as artistic or cultural content, for technology studies, or for 
forms of scholarship that treat digital games and virtual worlds as authored texts or artistic 
objects. Still, we also need to think about virtual world history in terms of events and 
activities, much as an archivist or historian would in the real world, and attend more carefully 
to preservation of forms of documentation in digital form that are external to virtual worlds 
as software environments. Software preservation without historical documentation of in-
 world events will be a barren project. 
 
These considerations bring us to specific future use cases and scenarios. Who might want to 
have a look inside a historical virtual world, and why?  And how might future librarians 
provide access to them?  Consider the following hypothetical use scenario: 
 
? A historian is interested in the cultural and social history of virtual worlds. His 
attention is drawn to vigils that were held in game worlds such as Everquest and 
Asheron?s Call after the WTC and Pentagon attacks on 11 September 2001. He would 
like to know more about virtual communities and the meaning to them of such 
events. However, it turns out that no data other than backed-up chat logs?which 
soon would be deleted?was ever kept on the game servers. Surviving 
documentation can only be found via the discussion boards, websites, and 
screenshots saved by players. Although he is interested in the virtual world?s history, 
his search for documentation depends on web archiving efforts and personal digital 
collections. Will full-text searching of this vast amount of material yield the 
documents he needs?  Or will finding needles in the web haystack require document-
 level indexing with specific date, location, or event tags? 
 
The moral of this example?  Documenting virtual worlds through software and data 
preservation alone will not do the job.  As we have already suggested, there will be no trace of 
past events inside a virtual world, at least not the present generation of these worlds. We 
noted one exception earlier (in Asheron?s Call), but they are rare. Here is what this means. 
Assume again that such a world has not been deleted, erased or remade, that we are on the 
right server or shard of a game or virtual world in which some documented event happened, 
and indeed that we are standing on precisely the spot where a historical event such as a 
conference or vigil occurred. We cannot dig, find old documents or inspect traces of the past 
on the site. Installing Everquest in 2050 will not reveal much about the virtual world that 
emerged from its game software. Nothing of the events that shaped player culture, game 
politics and relationships between players and developers will remain. Like Ozymandias? 
giant leg in the desert, the surviving software artifact will tell us nothing about what ?once 
dwelt in that annihilated place.?  Future researchers will be sorely disappointed if we do 
nothing to ensure that historical documentation about virtual worlds, but created outside 
them, is preserved along with software and proprietary data. 
? 50?
  
Now that we have introduced potential users for our virtual world collections, we will say 
more about access. Access is perhaps not a core concern for digital preservation per se, and 
there were no specific project deliverables that involved investigation of an access model or 
solution. Still, the Preserving Virtual Worlds project, due to its participants and its problem-
 sets, highlights the importance of finding solutions for problems such as identifying 
significant digital artifacts or developing standards for metadata. A key element in finding 
solutions to problems such as these is the collaboration of individuals and teams with 
different perspectives on software preservation, archives, and history. As a consequence of 
the Stanford group?s connection with a separate project team (led by the former project 
manager of Stanford PVW effort) concerned with the development of 3D environments for 
scholarly applications, we roughed out a concept and began development work on an 
access/delivery environment that might address some future access issues. 
 
Specifically, we began to re-think access to documentation and artifacts from virtual worlds 
collected in cultural repositories and, with that foundation, begin working on an approach to 
access that could lead to an alternative model for preservation of 3D artifacts from games 
and virtual worlds. A collection of such assets would include models, maps, geometries, 
textures, transactions, and so on. In working on this problem, we have been inspired by Ivan 
Sutherland?s earliest essay on virtual reality, ?The Ultimate Display.? In this case, however, 
we are working on an approach to access that may lead to an alternative model for 
preservation of 3D artifacts built for games and virtual worlds. This work depends once 
again on collaboration, in this case, on the allied project just mentioned above. Based in the 
Stanford Humanities Laboratory and the Computer Science Department, this project has 
been developing next-generation, virtual world platform called Sirikata. Sirikata is a BSD 
licensed, open source platform. The development team aims to provide a set of libraries 
and protocols that can be used to deploy a virtual world, as well as fully featured sample 
implementations of services for hosting and deploying these worlds. An alpha version has 
been used for a mixed reality performance at the MiTo International Festival of Music in 
Milan, 12-13 September 2009, and is currently live for an installation at the Bornholm Art 
Museum in Denmark.  
 
Think about the assets and content that go into the creation of a virtual world: models, 
maps, geometries, textures, and so on. We are not sure yet how future scholars will 
visualize, analyze, and understand these artifacts in a digital repository consisting of data files 
and metadata. In administrator mode, which is what we have today, the Stanford Digital 
Repository is essentially a file directory. Now think about another model of access to 
artifacts from an historical world, also largely models and suitable spaces for these models: a 
natural history museum showing dinosaur skeletons in a set that takes the visitor to a 
prehistoric savannah. Access to the information preserved there is visual and is reinforced 
by immersion in the world of the artifacts. It should be possible to do something similar 
with 3D artifacts from virtual and game worlds. 
 
We are investigating the use of the Sirikata platform for the creation of a new kind of 
repository, one in which 3D objects are stored, retrieved and investigated as 3D objects. This 
means that we would like to be able to move original geometry and texture data?archival 
assets?from their original environments into such a repository. The two cases we are 
investigating are (1) digital artifacts such as maps or levels from 3D games, beginning with 
? 51?
 early titles such as id Software?s DOOM and Quake, and (2) exhibitions created in virtual 
worlds such as Second Life by cultural institutions, including libraries and museums. Can we 
move these objects into an instance of an open virtual world platform such as Sirikata? If 
so, might we think of these instances as virtual wings of a library, rather than file 
repositories, places where the historical artifacts are deposited, preserved, found and 
investigated in an environment that puts documentation and narrative alongside the 
artifacts? 
 
Maps are among the most important artifacts in game development and player cultures. 
Players analyze them, re-create them as mods in games other than the ones in which they 
were originally created, and build viewers and projections so as to better visualize how to 
optimize their game-play in these spaces. As artifacts in a digital repository built with virtual 
world technology, historical maps would not just be artifacts, they might also provide spaces 
in which to site other objects and documentation?such as models, screenshots, videos, or 
documentation?that provide information about what took place in these settings. This 
might be where our future historians goes to check out the locations used for the science 
conference held in World of Warcraft, for example, without having to assemble, install and 
figure out how to use the original movement and navigation systems of the game's user 
interface. While further development work is necessary to realize this vision, we have already 
successfully exported levels (maps) from id Software?s Quake (1996) to the open VRML 
format, from which we are able to move unaltered geometries and textures from this 
historical digital artifact to other environments, whether Maya, 3DS Max, or Sirikata. When 
the pipeline to the Sirikata-based virtual repository is completed, it will be possible to drop in 
to Sirikata and see 3D objects with the same geometries and textures they were given in the 
original game. In fact, these artifacts will be created from certified copies of original game 
data used to produce them in the first place, utilizing forensics techniques for data extraction 
and the preservation methodologies worked out in the Preserving Virtual Worlds project. 
Future access environments for digital repositories will need to consider how to provide 
scholars with access both to digital artifacts from environments such as virtual worlds and to 
digital documentation about these artifacts and worlds. Bootstrapping virtual world 
technology as a means for re-conceptualizing the digital repository as a 3D environment 
rather than a file system provides a potential solution to this problem and deserves further 
investigation with the specific goal of developing access solutions for virtual world and game 
assets. 
? 52?
 5. Software Preservation and the Law 
 
Copyright Issues: Complex Ownership and Orphan Works 
 
The owner of IP has the right to transfer all rights by assignment, or a portion of the rights 
by license. The rights may be placed in the public domain, either intentionally, through IP 
misuse or neglect, or as the result of expiration of a registration. Or the owner may elect to 
make no use of the IP rights and to prohibit others from use. It has been charged that some 
patents are procured not to protect the owner?s use but to prevent use of the invention by 
competitors. 
? Introduction to Game Development, Ch. 7.5, ?Transfer of IP Rights? 
 
Copyright 
 
The most basic preservation action for any digital object is to convert it into a media-neutral 
format. This is typically done by creating an image of the original media; that is, making an 
exact, bit-for-bit replica of the disk that can then be mounted from the hard drive or burned 
to a fresh disk (e.g. an ISO or IMG file). In other words, the simplest and most common 
preservation process violates reproduction and (usually) anti-circumvention provisions in 
U.S. intellectual property law. 
Copyrights held by corporations endure for up to 120 years, and under the DMCA cultural 
heritage institutions enjoy no special privileges. A video game console generation typically 
lasts less than a decade. We are currently in the 7th generation of consoles, and personal 
computers have evolved in equally dramatic ways since Apple II and Commodore 64 began 
saturating the home PC market. Given the difficulty of identifying and then obtaining 
permission from the current rights holders of older video games, this translates into libraries 
risking fines of $200-150,000 per game were they to migrate their collection of classic 
software from 3.5" floppy disks to images stored on hard drives, an act comparable to 
rebinding a book or creating an access copy of a manuscript. 
 
Patents 
 
According to Stephen Rubin, video game patents (or at least litigation surrounding them) 
most often center on game controllers, consoles, and peripherals (Rubin, 2010). As patents 
have the shortest lifespan of all intellectual property protections, and outside of museums 
hardware is likely to be discarded in favor of accessing the games on modern computers, 
they have far less of an effect on cultural heritage institutions. Even so, at least two 
generations of game consoles have passed by the time a patent expires (20 years from the 
time of filing) and there is a danger of hardware becoming irreparably damaged after support 
has stopped, with no legal option to build a replacement. 
 
The Vectrex console, released in 1982, was unique for two reasons: it used vector rather than 
raster graphics, and it was the first console to offer a 3D peripheral. The Vectrex had a short 
market life, and the 3D Imager an even shorter one?making it a very rare, and expensive 
device. The rarity makes the games even more appealing to collectors and, in a display of 
? 53?
 typical resourcefulness, someone in the video game community began building 3D imagers 
from breadboards and portable CD players (Woolff). Because the Madtronix Imagers were 
produced in 2005, the original patent had expired and Woolff was able to sell them for about 
$75 each, a far cry from the $700 or more you might pay for one on the grey market, if a 
functional 3D Imager could even be found to purchase. Without this peripheral, a number 
of games are rendered unplayable, and preservation is meaningless without access. Although 
in many cases hardware upkeep is not sustainable, there are instances when the device is 
unique enough to be worthwhile. 
 
The majority of video game patents may be hardware related, but patents on software, a 
fairly controversial topic, do exist. Rabin cites Sega, Nintendo, Sony, and Microsoft as all 
having patents on game play. Patents of this nature serve to strengthen the protections 
already offered to individual games by copyright. 
 
Trademark 
 
Unlike copyrights and patents, as long as a trademark is used and protected (that is, the 
owner doesn?t look the other way when made aware of infringement), it never expires. 
Historically, trademarks have simply been logos or other commercial identifiers. With video 
games, flagship characters become brand identifiers and thus eligible to be registered as 
trademarks. This could potentially mean that works that would otherwise enter the public 
domain in a fixed time period could be monopolized by a progression of rights holders for 
centuries.  
 
Trade Secrets and Licensing 
?
 Trade secrets and licensing add further complexity to the web of video game IPR. Trade 
secret regulations are not as well known or clear-cut as the three major categories of 
protections. Without even getting into the mire of contract law created by end user license 
agreements (EULAs), licensed content (for instance, games released with a major motion 
picture) makes the task of identifying rights holders and obtaining permissions far more 
arduous. In the book Challenges for Game Designers, a great deal of time is spent discussing how 
to work with licensed content and how to pitch proposals for sequels and other licensed 
works. Combined with the proliferation of titles on the shelves using characters from 
Marvel, Disney, and other media giants, this seems to indicate that video games utilizing 
licensed IP are a sizable chunk of the existing catalog, and so are also likely to be included in 
any video game preservation projects. 
 
Emulators & the Law 
?
 We have argued for a greater degree of collaboration between the gaming community and 
the preservation community in insuring the longevity of games. The gaming community has 
been exceptionally active in the creation of emulators to enable access to older games. If 
emulation is to be used as a preservation strategy for some types of games, then it would 
behoove the preservation community to take advantage of the large amount of prior work 
done by the gaming community in creating and disseminating emulators for various 
platforms and, if possible, contribute to their further development as necessary. If 
? 54?
 preservationists are to become actively involved in the creation and dissemination of 
emulators, however, they need to do so with a full awareness of the legal issues surrounding 
emulators? creation and use. 
 
In the eyes of the law, the creation of an emulator constitutes a form of reverse engineering, 
in which an existing computing platform is disassembled and examined for the purposes of 
replicating its behaviors and operations in a new platform. While generally speaking the 
courts have found reverse engineering to be a legal activity, the ability to create an emulator 
can be limited by various laws, including copyright law, trade secret law, patent law, the 
Digital Millennium Copyright Act (DMCA), contract law, and the Electronic 
Communications Privacy Act (Electronic Frontier Foundation, 2010). To date, most of the 
court cases regarding emulators? creation have focused on copyright law, in particular on the 
extent to which an emulator?s creator may examine and use software from the original 
computing platform when creating the emulator. 
 
Two of the most significant pieces of case law regarding emulators have been the cases of 
Sega Enterprises Ltd. v. Accolade Inc. (977 F.2d 1510 [9th Cir. 1992]) and Sony Computer 
Entertainment v. Connectix Corp. (203 F.3d 596 [9th Cir. 2000]). In Sega, Accolade reverse 
engineered the Sega Genesis console platform by wiring a decompiler to a Sega Genesis 
system and loading different games on it in order to determine the interface specifications 
for the console. Accolade used the information obtained by decompiling the Sega Genesis 
software to create a software specification that they in turn used for creating Sega Genesis 
compatible games. The Ninth Circuit Court found that Accolade?s actions constituted fair 
use, as they had engaged in copying of the software embedded in the Sega Genesis system 
only to be able to access the ?unprotected ideas and functional concepts embodied in the 
code,? and disassembly of the code was the only means of gaining access to those ideas and 
functional concepts. 
 
The Sony case, however, is the more definitive ruling on the legality of producing and 
marketing an emulator. In this case, Connectix extracted the contents of a Sony Playstation?s 
BIOS and disassembled it in order to determine its behavior as part of creating an emulator, 
the Virtual Game Station, that would run Playstation games on Windows or Macintosh 
computers. As in the Sega case, none of the actual code from the Playstation system was used 
in the Virtual Game Station; Connectix merely studied the original code in order to replicate 
its behavior. The Ninth Circuit Court found that the extraction and examination of the Sony 
Playstation BIOS constituted fair use, as in the Sega case. The Sony decision, however, goes 
into far more detail regarding the court?s analysis of the four factor test involved in a 
determination of fair use. 
 
The court found that the purpose and character of the use made of the Sony BIOS by 
Connectix was ?modestly transformative,? in that it established a new platform on which 
consumers could play Playstation games, and that therefore this factor favored Connectix. 
On the second statutory factor, the nature of the copyrighted work, while the court 
recognized that software deserves copyright protection, it had previously ruled that such 
protection is less than that offered to ?traditional literary works,? and found that in this case, 
the standard set in the Sega case, that copying be necessary to access the functional aspects of 
the software had been met, and this factor also favored a finding a fair use. While the third 
factor, amount and substantiality of the portion used, weighed against Connectix (inasmuch 
? 55?
 as they had copied the entire work), the court also recognized that the copying had occurred 
as an intermediate step in the production of the Virtual Game Station, and that none of 
Sony?s software had actually appeared in the released product. On the fourth factor, the 
effect of the use upon the potential market, the court somewhat interestingly found that 
while the Virtual Game Station might negatively impact Sony?s sales of the Playstation, the 
transformative aspect of the Virtual Game Station meant that the VGS did not simply 
supplant the Playstation, and that therefore, this factor favored Connectix as well. 
 
The courts to date, then, have found that as long as an emulator does not include software 
taken from the original system, it does not directly infringe upon copyright. The gaming 
community has discussed the possibility that the creators of an emulator might be taken to 
court for contributory infringement, on the theory that by enabling others to play games on 
a platform other than the original, the creators of an emulator are inducing others to illegally 
copy game software to enable it to work with the emulator. As far as we know, the courts 
have not directly addressed this issue. In other cases that have focused on technological 
devices? role in contributory infringement (see, for example, Sony Corp. v. Universal City 
Studios, Inc., et al. 464 U.S. 417 [1984] and MGM Studios, Inc. v. Grokster, Ltd. 545 U.S. 913 
[2005]), the courts have focused on the issues of whether the technology has significant non-
 infringing uses, whether those responsible for producing the technology were in a position 
to control the behavior of those infringing on others? copyrights, whether those producing 
the technology promoted the use of the tool for infringing activities, and whether those 
producing the technology materially benefited from the infringing activities of others. Any 
library or archive wishing to become involved in the creation of emulators needs to be aware 
of these issues surrounding contributory infringement, and avoid any activity that might be 
seen as encouraging others to infringe copyright or profiting from others infringing 
activities. 
 
There are a number of other legal issues that may arise with respect to the creation of 
emulators other than copyright issues. In the Sony v. Connectix case, one of Sony?s accusations 
was that Connectix had tarnished Sony?s Playstation trademark by the negative associations 
that would arise in users? minds regarding Playstation games if they played them under the 
Connectix Virtual Game System and had an inferior experience of the game. While the court 
rejected this argument, stating that consumers would be able to distinguish between the two 
platforms and avoid misattributing quality issues to the Playstation brand, the existence of 
this complaint does indicate that those creating emulators may need to be careful to avoid 
any appearance of intruding on companies? trademarks. 
 
Contract law and the Digital Millennium Copyright Act may also impede programmers? 
ability to create an emulation of an existing platform. In Davidson & Associates DBA Blizzard 
Entertainment, Inc.; Vivendi Universal Inc. v. Jung et al., 422 F.3d 630 (8th Cir. 2005), Blizzard 
Entertainment sued the developers of BNETD, a reverse-engineered implementation of the 
Battle.net online gaming servers that support multiplayer online games from Blizzard such as 
Diablo, Starcraft and World of Warcraft. Blizzard claimed BNETD?s programmers had agreed to 
the Battle.net end user license agreement and terms of use when they signed up for the 
service, and the EULA and TOU for Battle.net specifically prohibit reverse engineering. The 
court upheld the license agreements as enforceable contracts and found BNETD?s 
developers in violation of the EULA. The court also found that BNETD?s developers 
? 56?
 violated the anti-circumvention provisions of the DMCA by emulating the authentication 
sequence used by Battle.net servers when establishing communication with a client game. 
 
As a sidebar to this court ruling with respect to emulation, it should be noted that BNETD 
source code continues to be widely disseminated and that hundreds of servers around the 
world based on this code are openly operated. This observation raises the possibility that 
content created on such a private server (replays, machinima, modified assets) might become 
part of a future archive or collection; indeed, it is likely that some of this content is already 
evident in a collection such as the Machinima Archive. As part of our collaboration with the 
Center for Internet & Society at the Stanford Law School, we consulted with legal staff and 
students at the law school to discuss the specific issue of the liability of a repository (such as 
the Internet Archive) in such an instance. While in the absence of a specific example or legal 
action, PVW was not in a position to obtain legal advice and there has been to date no clear 
legal precedent in this area, it appears that the exposure of an archival repository to legal 
action due to the inclusion of such content in a video, for example, is likely to be quite small. 
In the case of an objection to making such a video available for viewing, it is likely that 
removing public access to the archived content would be an acceptable response. This has 
been the practice of the Internet Archive, for example; however, the application of such a 
policy to the case of content produced on a ?private server? (as a separate issue from issues 
of reverse engineering and emulation) has not yet been tested in the courts. 
 
Finally, developers of emulators may find that even if they do not fall afoul of copyright law, 
they may encounter problems with patents. There has been substantial litigation in recent 
years regarding patents and emulation technology, most of which has settled out of court. 
With copyright law providing only a limited ability to restrict the actions of those developing 
emulators, companies that do not wish to see emulators for their products are more 
frequently turning to patents as a means of restraining the development of competing 
technology. 
 
The above issues all relate to the development of emulation technology. There are also legal 
issues involved in the use of game emulation technology, most notably copyright and the 
DMCA?s anti-circumvention provisions. Most of the copyright issues center on the creation 
and use of ?ROM? files from game cartridges for game console platforms. In Atari, Inc. v. JS 
& A Group, Inc., 597 F. Supp. 5, Atari sued the JS & A Group for copyright infringement and 
patent infringement, stating that JS & A, in marketing its product PROM BLASTER (which 
allowed users to make copies of Atari 2600 game cartridges), was guilty of contributory 
copyright infringement. JS & A, citing ? 117 of the Copyright Act, stated that the PROM 
BLASTER was intended to allow users to make archival copies of their legally owned 
computer programs and hence its primary use was non-infringing. The District Court 
decided that, as  ? 117 had been rewritten based on the input from the final report of the 
National Commission on New Technological Uses of Copyrighted Works (CONTU) that 
states that archival copies of digital media should be allowed ?to guard against destruction or 
damage by mechanical or electrical failure,? and since game cartridges as media are not 
particularly subject to mechanical or electrical failure, making an archival copy of a game 
cartridge constituted infringement. While this ruling was mitigated somewhat by the case of 
Sega v. Accolade, where the Ninth Circuit found that making a copy of a game cartridge could 
be fair use if done for the purposes of making an intermediate copy needed to reverse 
engineer a game system, the case law at the moment is clear that copying a game cartridge 
? 57?
 for the sole purpose of being able to play the game using the copy, rather than the original, is 
copyright infringement. 
 
The DMCA may also inhibit a library?s ability to use certain emulation technologies. Some 
modern computer games employ digital rights management technology, such as SecuROM, 
which checks for the original installation CD-ROM or DVD-ROM when it runs an installed 
program. While there are disk emulation technologies, such as Daemon Tools and Alcohol 
120%, that claim to be able to create a copy of a protected CD or DVD and mount it in a 
virtual drive that will enable the game to run even if the original disk is not present, such an 
action constitutes bypassing a technological protection measure. While the use of such disk 
image emulation technologies is fairly common among the gaming community, the legality 
of such an action has not, to the best of our knowledge, been tested in court. 
? 58?
 6. Preservation Strategies 
                                                             
Analysis of Hardware Preservation 
 
As part of the preservation of video games and virtual worlds process, it is important to 
understand how current collectors are faring as part of their acquisition and preservation 
strategy. To this end, the research team conducted an interview with J.P. Dyson, director of 
the International Center for the History of Electronic Games (ICHEG), and vice president 
for exhibit research at the Strong National Museum of Play. The mission of the ICHEG is 
to collect, study, and interpret electronic and computer games from a number of 
perspectives, including the cultural impact of games upon society, and how people play and 
learn from such experiences. Strong and the center have the additional mission of catering to 
two distinct audiences: the general public and patrons who visit the Strong National 
Museum of Play, as well as academics and researchers engaged in the study of video games 
and play experiences.  
 
ICHEG has processed close to 20,000 game artifacts and 100,000 game-related artifacts in 
the last 18 months. Dyson states that the preliminary focus of ICHEG has been three-fold: 
acquiring primary artifacts related to electronic games hardware and software, restoration of 
the collection for both display and interaction, and gathering of ancillary artifacts related to 
the play experience and documentation of the production process. Acquisition of content is 
procured through private donation as well as active investigation and outreach by the 
museum staff.  
 
Restoration processes vary depending upon the type of artifact. ICHEG and Strong rely 
upon a number of sources for the repair and restoration of electronic game equipment, 
including internal personnel, experts from the community (from both academia and 
industry), external paid experts, and others. For arcade cabinets, the process involves either 
internal specialists or external vendors. Special attention is paid to the type of restoration. In 
some situations, it is only acceptable for the restoration to occur with original parts (the 
preferred method). In other situations, original part restoration may be secondary, such as in 
situations where motherboard components are no longer available. For consoles and 
computers, ICHEG often acquires multiple copies of the same artifact. In such situations, 
working components can be selected or reassembled from the duplicates within the 
collection. When a component does not work, it can be relegated as a display-case sample. 
Museum personnel have also acquired parts collections that can, in turn, be used to repair 
broken devices. 
 
Along with parts, the restoration process for an individual component can take a number of 
forms. Reconditioning may include, depending upon the condition, resurfacing of a CD-
 ROM and the restoration of magnetic media components. Archival process includes the 
capture of video and image related to the play experience. For components that find their 
way into Strong?s interactive installations, reinforcement or reconditioning may include 
processes by which the artifact can be made ?bulletproof? for the duration of the exhibit 
showing. For example, joysticks and other input devices may have to be replaced by versions 
that can survive the repeated use by patrons of all ages. The museum has used this technique 
for the Atari 2600 version of Pac-Man, due to its overwhelming popularity. 
? 59?
  
According to Dyson the museum currently focuses on the acquisition and assessment of 
artifacts. To create the proper experience for the general public and researchers, a number of 
steps must be undertaken. For consoles, reproducing an experience means coupling 
hardware with televisions and display devices of the proper era. Dyson has noted that the 
scarce nature of older televisions coupled with technological problems of newer televisions 
can cause impediments in re-creating how the game can be experienced in its original form.  
 
Dyson said that ICHEG and the museum are starting to explore the use of emulation and 
virtualization technologies as a method of preserving game experiences. The Strong National 
Museum of Play has used emulation technologies in at least two situations where the original 
artifact was damaged or was too valuable to exhibit for public interaction.  
 
Dyson states that there are a number of challenges looming on the horizon for museums. 
Currently, ICHEG is concerned about games delivered through technologies such as Steam 
or the iPhone Apps store. In such cases, the distribution platform is a part of the game 
experience, and if the distribution channel were to cease operations, the process of capturing 
the game experience would be exceedingly difficult. Dyson is also concerned about the 
preservation of experiences related to massively multi-player games and games that exist and 
persist over a series of distributed servers. Such preservation leads to larger questions 
regarding the overall preservation of software, firmware, operating systems, and application 
infrastructures so that an experience can be properly recreated.   
 
Dyson described current efforts of the museum and ICHEG to create a digital collections 
policy. The policy is intended to determine what it means to collect artifacts that are digital 
in nature, and is a starting point for the process and for communication related to the 
preservation of digital artifacts.  
 
Dyson also considers establishing communication between collectors and the companies 
that produce electronic game hardware and software as one upcoming challenge. With better 
connections, the potential exists for greater collaboration related to preservation. 
 
ICHEG?s long-range plans are to move their processes beyond acquisition, restoration, and 
presentation. ICHEG and the Strong National Museum of Play are committed to creating 
plans for large-scale preservation and presentation policies that will require emulation, 
virtualization, and migration strategies. 
 
In the Preserving Virtual Worlds project, our research team gained insight into the problems 
that are too often encountered by museums in their quest to provide preservation and access 
to historical video games.  
 
One such problem dealt with the process of video conversion for both experiencing game 
play, as well as capturing video of game play for preservation, which first occurred in a case 
study related to the Atari 2600 Star Raiders game. The Atari 2600 utilized a circuit called the 
?Television Interface Adapter,? which output a radio frequency signal. The signal could be 
received by a standard television set and was treated in a similar manner as an ?over the air? 
transmission. The signal was routed into the television set through a switch block, which 
allowed the player to select either the game system or a standard television antenna. As part 
of the Preserving Virtual Worlds research, the team acquired three such Atari 2600 systems 
? 60?
 from different eras, including an original Sears Telegames Atari model, an Atari 2600 
?Heavy-Sixer? (manufactured around 1977), and an Atari 2600 ?Light-Sixer? (manufactured 
around 1980). In order to create an authentic game experience, the team was able to procure 
black and white and color televisions from the early 1970s (Sony 1974 Color TV 19", 
Motorola B&W 13" portable with components dated 1971). Both of these sets were 
characterized by split VHF/UHF inputs, rotary channel selection and tuning, and a wide 
bezel to guard against overscan issues related to that era cathode ray tube technology.  
 
Although the team was able to gain a satisfactory play experience from the 1970s television 
technologies, several problems manifested with progressively newer sets. For example, with 
televisions from the 1980s through 2009, the team experienced problems such as pixelization 
of image, fluctuation of brightness (photodiode and photoresistor compensation), problems 
with signal lock (digital tuner problems), problems with sound reproduction (volume, 
clarity), impedance conversion (conversion from 300 ohm to 75 ohm), and overscan 
problems (ghosting, artifacts around image edge). The problems are only exacerbated by 
newer television technologies based upon the conversion to digital TV signals in the United 
States. The team was already aware of television technologies that have eliminated the 
possibility of over-the-air input based upon the lack of analog signal as well as the decrease 
in popularity of VCRs, game consoles, and DVD players that use a direct RF broadcast 
signal. In such cases, the television is not even capable of receiving the signal, and as such, 
nothing is displayed.  
 
The problem also gains new dimensions if the video signal is to be recorded on a digital 
capture device. In such scenarios, additional signal transformations may have to occur. For 
example, in order for the team to record screen capture from the Atari 2600 using a digital 
process, the signal was transformed from an analog RF signal to a composite TV signal 
(using a VCR to do the transformation) to a digital signal (using a digital camcorder).  
 
The experience gained from the video conversion project was shared with a team of curators 
and game restoration specialists at the Strong National Museum of Play as part of an 
extended dialogue that has resulted in a partnership between the museum and the Rochester 
Institute of Technology, in which students gain cooperative education credit by assisting the 
museum in video capture and play experience restoration tasks. 
 
A second problem is the conversion of game-related media so that the media can be read 
and processed by newer technology. Many of the earlier game technologies utilized 
proprietary formats for their ROM cartridges, magnetic tapes, and floppy disks. As the 
systems age, it becomes harder to find modern technology equivalents that are capable of 
processing the information. For example, if a person no longer possesses an original Atari 
2600, the only way to retrieve the information from a cartridge is to either purchase an Atari 
2600 ROM reader from a third-party company (it cannot be purchased from the average 
computer or electronics store) or to extract the relevant ROMs from the cartridge?s circuit 
board and process them through a ROM reader. With magnetic media, combinations of 
media packaging, sizes, capacities, and protection techniques compound the problem. For 
example, when the research team worked on a case study with DOOM, not only did it take 
some time to hunt down a computer that was still capable of reading a floppy disk, but it was 
also important to identify a floppy drive that could still process both single as well as double 
density disks. 
? 61?
  
In summary, game museums are at an important crossroads in the preservation process. 
Curators at such institutions must balance the need to acquire rapidly disappearing artifacts 
related to the early history of video games while ensuring that their acquisitions will be able 
to be studied and enjoyed for decades to come. The promise of preservation technologies 
and their strategies may be able to span the bridge between these two desires. 
 
Analysis of Emulation and Virtualization 
 
As part of the preservation strategy for virtual worlds and electronic games, one must 
consider options that allow players the ability to experience a game well beyond the 
availability of a given hardware platform. Ideally, museums and collectors wish to preserve 
the original experience through maintenance and upkeep, but as hardware platforms age, it 
becomes more difficult to find repair parts. Furthermore, finding individuals with sufficient 
knowledge to repair such systems also becomes increasingly difficult over time. In order to 
augment the preservation of the original game system, techniques of emulation and 
virtualization can be used to provide game play experiences on newer hardware. This section 
examines the use of emulation and virtualization techniques as well as their implications for 
our case study games. 
 
Emulation Overview 
 
An emulator is a software application that is capable of simulating a particular hardware 
platform, including the hardware and appropriate firmware and support software. Playing 
through an emulator varies from the original experience because the emulator is designed to 
execute on hardware that is radically different from the original platform and corresponding 
system architecture. Emulators are designed to utilize the original executable code from the 
game. For most emulation systems, this involves extracting and providing a copy of the 
native cartridge ROM or game CD-ROM.  
 
 
? 62?
  
Over the last several years, the game community has seen a surge in emulation technologies 
designed to allow users to play many of their favorite, classic console games. In order for an 
emulator to successfully re-create a play experience, it must address several key areas of the 
original game environment platform, including the central processing unit, the memory 
manager, the environment code, and peripherals. In addition, the emulator must have a 
means of loading and processing the original game code. 
  
? Central Processing Unit ? An emulator has to be able to translate CPU instructions 
and must be able to emulate the functionality of a specific processor. This includes 
all arithmetic and logical operations, registers, storage modes, addressing modes, 
interrupt handling, and input/output operations.  
? Memory Manager ? An emulator must be able to provide the structure and access to 
main memory that is analogous to the original platform. This includes memory 
layout and management for system internal ROMs, scratch RAM, system RAM, 
device-based RAM and ROM (video, audio), and memory controllers (MMUs, DMA 
circuits). 
? Environment Code (ROM/Firmware/BIOS/OS) ? An emulator requires the 
appropriate software framework to operate. For some emulated systems, the 
firmware is separate from the instructions that are part of the game media. For 
example, DOS emulators will provide access to BIOS memory and some consoles 
will provide access to ROMs for common system functions (peripheral calls, math 
calculations, IO routines). Due to both maintenance and legal reasons, some 
emulator developers separate the environment code from the rest of the emulation 
system. Despite potential separation, this code is absolutely required for the 
emulation system to operate. 
? Peripheral Mapping ? An emulator must provide mappings from the original system 
to the target system. For example, in the case of most console systems, components 
external to code execution include video output, player input (joysticks, paddles), 
audio, and the game cartridge ports. In many cases, these components were not 
designed for general use, but were specific to the particular platform. For example, 
the Atari 2600 provided video and audio through a custom-designed Television 
Interface Adapter (TIA), which provided a means for the consumer to attach the 
console to a standard TV. The emulator must provide peripheral translation such 
that TV signaling can be displayed upon modern video cards and monitors, input 
devices can be mapped to modern joysticks, mice and keyboards, and ROM 
cartridges can be mapped to a modern file system. 
? ROM/Emulated Software ? Target software that can be run on an emulator. 
Technically, the target software is separate from the emulation 
environment. However, for most game console emulators, the ROM or Emulated 
Software must be processed so that it can be properly loaded and executed in the 
emulator. For example, to run an Atari 2600 cartridge in an emulator, a copy of the 
ROM code must be extracted with a ROM or cartridge reader and must be 
formatted in a particular manner such that the emulator can properly parse the 
resulting file.  
 
The community for emulation development, especially for video games, can be best 
characterized as a grass-roots movement led by a few dedicated programmers. For most 
? 63?
 emulator developers, the process involves reverse engineering the hardware and firmware 
present in the original systems. Currently, the emulation movement?s focus has been 
predominately on classical console systems, including the Atari 2600, Sega, Playstation 1, 
NES, Game Boy, and other early platforms. For general computing platforms, classic game-
 related computers such as the Apple II, TI99/4A, Commodore 64, Atari 400/800, and 
Amiga have all been the target of emulation development. At one level, the focus on earlier 
platforms may be based on their simplicity. Earlier platforms consisted of simpler CPUs 
(6502 series being a popular model), minimal memory management schemes, and limited 
peripheral detail. At another level, as these platforms age, hobbyists and fans are frantically 
looking at means to preserve these experiences. As the community often drives emulation 
efforts, the efforts are in various states of completion. A quick survey of popular emulators 
for the Atari, Sega, Sony, Commodore, Apple, and Nintendo brands showed that over 70% 
of the efforts never resulted in a viable emulation environment. In addition, for those efforts 
that were completed, portions were abandoned for a number of reasons, including developer 
abandonment and legal confrontation. To make the situation even more difficult, a survey of 
abandoned projects revealed that their websites were taken over by malicious software (e.g. 
many of the Apple II emulation sites have been replaced by malware websites as well as 
virus- and spyware-laden applications). 
 
Emulation Criteria 
 
In exploring emulation environments, the research team compiled a list of factors deemed 
critical for an emulator that was to be part of a preservation experience. The list of criteria is 
as follows: 
 
? Source Availability ? Is the source code for the emulator freely available? 
? Source Maintainability ? Is the programming style of the emulator source code 
readable?  How easy is it to maintain the code?  How well is the source code 
documented?  Does the source code?s architecture make sense?  How easy is it to 
obtain the compilation tools and libraries necessary to rebuild the emulator from 
scratch? 
? Licensing ? Does the emulator support a license model that is compatible with 
preservation efforts? 
? Range of Host Platforms and Operating Systems ? Does the emulator work on a 
variety of PC and Apple platforms?  Can the emulator run under common operating 
systems such as DOS, Windows, Apple OSX, and Linux? 
? External Documentation ? What is the extent of external documentation related to 
the emulator and the target system, including what is necessary to execute the 
emulator and related game ROMs? 
? Sensory Fidelity ? Does the emulator appropriately reproduce the visual, aural, and 
tactile interactions of the original game? 
? Peripheral Emulation ? Does the emulator properly reproduce the range of 
functionality related to the peripheral interfaces?  Are the peripherals properly 
reproduced from a code and experience standpoint? 
? Peripheral Expansion ? Does the emulator allow for the connection of original 
peripheral devices (such as gamepad or joystick)? 
? 64?
 ? Performance ? Does the emulator run within acceptable parameters on the host 
operating system?  Are there options available to adjust the performance of the 
emulation software? 
? Ease of Use ? Is the emulator easy to install and use by someone with a low level of 
technical proficiency? 
? Community Maintenance ? Is the emulator actively maintained?  How large is the 
community supporting the emulator?  Are there any active or looming impediments 
to the continuation of the emulator? 
? Robustness ? To what degree can the emulator execute the original games?  If there 
are still bugs, to what extent do they inhibit playability? 
? System Requirements ? What does it take to run the emulator? (CPU, memory, 
RAM, peripherals) 
? Emulation Variation ? Does the emulator only emulate the standard run-of-the-mill 
platform or does it emulate the various different manufacturing of firmware 
differences of a particular series (e.g., does an Atari 2600 emulator have any 
indicators that there were actually four different production versions of that 
particular model and does the emulator embody the nuances between these 
variations?) 
 
For purposes of this research, the team considered an emulator to be ?good? if it had the 
following characteristics: 
 
? The emulator is based upon freely available source code and appropriate licensing. 
? The emulator is actively maintained. 
? There is reasonable internal and external documentation for the project. 
? The emulator interface is easy to use by non-technical players. 
? The emulator supports a wide range of performance and tuning options. 
? The emulator is robust and provides a believable level of fidelity when compared to 
the original game experience. 
 
Virtualization Overview 
 
Virtualization is another technique that can be used to help in the preservation process. 
Similar to emulation software, virtualization software is an application that allows a user to 
execute code related to a different operating environment from the host system. Unlike 
emulation, virtualization is performed on CPU and system architectures in which the host 
and target share similar attributes. In addition, virtualization creates a level of abstraction 
that does not include the target operating system. Instead, virtualization provides a 
bootstrapping mechanism in which a legacy or different operating system can be loaded into 
virtual machine. From there, additional software can be loaded into the virtualization 
environment.  
 
? 65?
  
 
Virtualization?s focus is on providing the appropriate level of abstraction for the hardware 
environment, including the CPU, memory manager, and peripheral operation. Unlike 
emulation, virtualization employs a variety of techniques to leverage functionality of the host 
hardware architecture. For example, virtualization solutions developed for modern PC 
architectures may employ a technique known as hardware-assisted virtualization. This 
technique relies upon special processor instructions (Intel?s VT-x and AMD?s AMD-V 
extensions) that help accelerate the base assembly instructions. Some experimental systems 
use a technique called paravirtualization, in which the guest operating system is modified for 
use with the virtualization process. In either case, the virtualization system must provide 
access to the following hardware components or provide an appropriate level of abstraction: 
 
? Central Processing Unit (CPU) ? Ideally, in virtualization, the native CPU handles 
every one of the instructions handled by the virtualization system. Unlike emulation, 
which requires mapping and translation, this means that either the CPU must have 
the appropriate extensions to handle the virtualization instruction calls or that the 
target operating system has been modified to allow for code to be context switched 
from the virtualization environment to the actual CPU. 
? Memory Management ? The memory management system is a hybrid between the 
actual hardware and an abstraction process. At one level, the virtualization process 
must provide the equivalent architectural access to main memory, including multi-
 user and demand paging techniques used within the system. At another level, the 
virtual machine?s memory space should be separated from any other virtual machines 
or any processes running on the host system. 
? BIOS/Firmware ? The virtual machine should have access to the BIOS of the host 
system. In cases where a legacy operating system requires access to an older BIOS, 
? 66?
 the virtualization technology may provide a way of loading such ROM images into a 
section of RAM memory. When such techniques are utilized, the virtualization 
system also must account for any potential situations where devices are mapped into 
the ROM/BIOS memory space.  
? Bootstrapper ? For native systems, there is usually a process by which an operating 
system is loaded onto a hard drive. In such cases, BIOS or firmware code is 
necessary in order to search for peripherals that contain bootable environments. 
Virtualization systems must contain code to provide an operating system bootstrap 
for initial configuration. 
? Peripheral Mapping ? Virtualization technology must provide resource management 
for peripherals. In some cases, the peripheral is remapped, such as the use of a USB 
memory stick for a floppy drive as well as a large file for a primary hard drive. In 
other cases, the host peripheral is utilized in either a shared or exclusive mode, such 
as CD-ROMS, video, and sound cards. 
 
Unlike the emulation community, the target of the virtualization community is generalized 
and is often supported by larger corporations. For example, Oracle/Sun Microsystems, 
VMWare, the RedHat foundation, and other entities that have invested substantial resources 
into the process are conducting leading efforts in this area. 
 
Virtualization Criteria 
 
The research team compiled a list of criteria for virtualization technologies: 
 
? Source Availability ? Is the source code for the machine freely available?  If not, is it 
supported by a commercial entity with significant investment in virtualization 
technology success? 
? Source Maintainability ? Is the programming style of the virtual machine source code 
readable?  How easy is it to maintain the code?  How well is the source code 
documented?  Does the source code?s architecture make sense?  How easy is it to 
obtain the compilation tools and libraries necessary to rebuild the emulator from 
scratch? 
? Licensing ? Does the virtualization technology support a license model that is 
compatible with preservation efforts? 
? Range of Host Platforms and Operating Systems ? Does the virtual machine 
software work on a variety of PC and Apple platforms?  Can the emulator run under 
common operating systems such as Windows, Apple OSX, and Linux? 
? External Documentation ? What is the extent of external documentation related to 
the virtualization platform? 
? Peripheral Emulation ? Does the virtualization technology properly reproduce the 
range of functionality related to the peripheral interfaces?  Are the peripherals 
properly reproduced from a code and experience standpoint? 
? Peripheral Expansion ? Does the virtualization technology allow for appropriate 
device mappings?  Are there clear interfaces for the addition of shared, exclusive, or 
simulated device code?  Can the process of device inclusion be handled by a 
programmer with moderate systems programming skills? 
? Performance ? Does the virtualization technology run within acceptable parameters 
on the host operating system?  
? 67?
 ? Ease of Use ? Is the technology is easy to install and use by someone with a low 
level of technical proficiency? 
? Community Maintenance ? Is the virtual machine actively maintained? 
? Robustness ? To what degree can the virtualization system run applications?  If there 
are still bugs, to what extent do they inhibit a software experience? 
? System Requirements ? What does it take to run the virtualization system? (CPU, 
memory, RAM, peripherals) 
 
For purposes of this research, the team considered a virtualization technology to be ?good? 
if it had the following characteristics: 
 
? The virtualization technology is based upon freely available source code and 
appropriate licensing. In cases where the source was not available, the virtualization 
technology was being developed by a company dedicated to virtualization. 
? The virtualization technology is actively maintained. 
? There is reasonable documentation for the project. 
? The virtualization interface is easy to use. 
? The virtualization technology is robust and provides a believable level of fidelity 
when executing legacy software. 
 
Case Studies 
 
As part of this study, the research team selected four game titles from our case set for testing 
with emulation technology. The game selection included Star Raiders for the Atari 2600, 
Mindwheel for the Apple II computer, Mystery House for the Apple II computer, and DOOM 
for an IBM PC/MS-DOS environment. For the purposes of the study, the selected software 
titles were executed on modern hardware with a selection of commonly used operating 
system software. In all, the team selected ten Intel-based PCs, with Pentium IV processors, 
at least 4 GB RAM, Nvidia 8800, 9600, and GTX260 video options, CD/DVD ROM drives, 
at least 320 GB hard drives, integrated Realtek sound solutions, as well as standard mouse 
and keyboard devices. The systems were configured with a range of host operating systems, 
including Windows XP 32 bit edition, Vista Ultimate 32 bit edition, Windows 7 Ultimate 32 
bit edition (Release Candidate), Windows 7 Ultimate 64 bit edition (Release Candidate), and 
a variety of Linux variations including Ubuntu, Debian, Arch, Gentoo and Fedora. For 
Linux installations, additional packages were installed to create an appropriate graphical and 
audio environment for emulation and virtualization, including X11, Gnome, ALSA, 
PulseAudio, OpenAL, OpenGL, and other library and application sets. In addition to the PC 
environments, the team had access to both Intel- and PowerPC-based MacBook Pro 
computers along with Intel-based Mac G5 systems. These systems were configured with 
both Tiger OSX (10.4) and Leopard OSX (10.5). 
 
Case Study 1 ? Star Raiders 
 
Star Raiders, released by Atari for the 2600 platform in 1982, was a cartridge-based computer 
game for the Atari 2600, in which the player pilots a fighter spacecraft and engages in battle 
with opponents known as ?Zylons.? Star Raiders is an instrumental game in video console 
history for a number of reasons. First, it immersed the player in a first-person, ?3D? 
dogfight environment, which allowed the player to engage enemy craft via a combination of 
? 68?
 thrust and rotation interaction. As such, the game is considered to be the forerunner of an 
entire series of dogfight combat simulators for arcade and home console play, including such 
classics as Wing Commander, Star Wars, and Elite. Second, the game utilized a non-standard 
pad controller that allowed for keyboard-style commands on the Atari 2600. The player 
worked the keyboard controller in tandem with the joystick depending upon the particular 
phase of the game. The Star Raiders cartridge consisted of an 8K ROM, which stored the 
entirety of the game code.  
 
The target platform for Star Raiders included the Atari 2600 game console. Created in 1977, 
the Atari Game Video Computer System consisted of a MOS 6507 (6502 variant) 
microprocessor, a MOS 6532 RAM-I/O Chip (128 bytes static memory, timer, and 8 digital 
I/O ports), and a Television Interface Adapter Chip (TIA). The MOS 6532 was used to 
buffer the status of console chips as well as buffer the states of the joystick devices. The TIA 
provided the ability to generate an RF signal that could be received on channels 2 or 3 of a 
standard television set. The TIA generated both the video and audio signals used by the 
game system. The console also provided an adapter port for receiving game cartridge ROMs. 
 
In order to ensure emulation accuracy, the research team had access to several original Star 
Raiders cartridges as well as three Atari 2600 platforms, including a Sears Telegames Model, a 
?heavy sixer,? and a ?light sixer.?  As part of emulator testing, the research team selected 
four emulation systems to test on PC-based environments. The four systems selected 
included Stella (http://stella.sourceforge.net), z26 (http://www.whimsey.com/z26), PC 
Emulator (http://pcae.vg-network.com), and MESS (http://www.mess.org).  
 
All four emulation platforms were able to successfully run the Star Raiders ROM image. 
However, based upon the team?s criteria for emulation software, all but Stella were 
eliminated from contention. Specifically, MESS?s licensing policy was too restrictive for use 
as an archival system, and both PC Atari Emulator and Z26 are in an indeterminate state of 
upkeep, despite the latter?s popular following in some segments of the emulator community. 
In addition, PC Atari Emulator experienced difficulties with clock timings in the first version 
we evaluated (from a non-authoritative emulator site). The problem proved inconsistent, 
varying on target PC platform and operating system (problems were experienced on MS-
 DOS and FreeDOS).  
 
In side-by side comparison, there were several differences noted by the test team. First, there 
is a perceivable quality difference in the graphical output between the emulator and the 
original 2600 output for both black and white as well as color televisions (for authenticity, 
we did some of our testing on 1970s era televisions). The RF/television combination created 
an experience in which the graphics were not as sharp as the emulated version. Although this 
might seem as a detriment, the slight blurriness disguised several visual artifacts that 
distracted players on the emulated systems. Second, the sound quality was also different in 
the emulated versions. Although the sound approximation of the TIA was acceptable in the 
emulated version, there were some pitch and timbre variations noted in the audio 
reproduction when compared to the original experience. In part, this was due to differences 
in the fidelity when comparing the internal speakers of a 1970s era television versus a 
modern sound card output to commonly available desktop speakers. To verify the sound 
quality difference was not due to aging of the speakers, the team attached the Atari console 
to televisions from the 1980s and 1990s to compare the sound experience. The results 
? 69?
 correlated to the experiments performed with the 1970s era television sets. Finally, the 
mouse and keyboard experience did not properly substitute for the joystick and pad 
experience of the original. The team later found that an adaptor could be purchased 
(www.stellaadaptor.com), which would allow the original Atari 2600 joysticks and pads to be 
interfaced to a modern PC. The adaptor converted the original digital I/O signals from the 
9-bin connector to a USB format that could be utilized by z26 and Stella. Despite these 
issues, the Atari 2600 emulator community has created a robust product that provides an 
extraordinary level of support for reproducing the original Atari experience. The emulation 
technology is mature enough to support most Atari cartridges that were produced, and the 
involvement of the development community has ensured that most emulators support a rich 
set of tools for both debugging existing ROMs and allowing hobbyists a platform for 
software experimentation. However, this investigation has also revealed one of the critical 
weaknesses in the emulator community ? as emulators age, they too can fall into neglect and 
disrepair.   
 
Case Study 2 ? Mindwheel and Mystery House 
 
Mindwheel, published in 1984 by Broderbund, was available on a variety of platforms 
including the Apple II. U.S. Poet Laureate Robert Pinsky designed the game as an interactive 
novel that allowed the player to explore the thoughts and minds of four deceased characters 
within the fictional work, with the player trying to solve particular problems in search of a 
?Wheel of Wisdom.? The game was accompanied by a hard-covered instruction manual and 
a novella that served both as a way to introduce the story line and as an early form of copy 
protection, as the player was requested to enter words from particular pages and paragraphs 
in the novella.   
 
Mystery House, an Apple II title published in 1980 by On-line Systems, is a ?whodunnit? game 
in which the player must explore a mansion in search of clues that will lead to the identity of 
a murderer, who is also locked within the house. The game is based on the tradition of text-
 based exploration games, but also utilized simple line-based drawings to augment the text 
descriptions within the game. The game was designed to run on a standard Apple II 
computer and was distributed on 5.25" single-sided floppy disks for the Apple floppy drive 
reader. 
 
The target platform for these games was the Apple II/Apple II+ computer system. Based 
upon the MOS 6502 processor, the Apple II/II+ supported up to 48K RAM, cassette and 
floppy drives through expansion ports, and a NTSC composite output signal for both 
monochrome text and color graphics. The support ROM built into the Apple II/II+ 
provided direct support for Apple?s BASIC dialect. Sound was generated through pulsing a 
speaker. Crude wave generation could be computed on the computer, allowing for low 
fidelity playback of voice and multi-channel source sound (on a small internal speaker within 
the computer system).  
 
To ensure that the Apple II emulator experience was similar to the original, the team was 
able to secure time on an original Apple II+ system in order to determine the how well the 
emulator matched to an original system over a variety of Apple II applications and games. 
 
? 70?
 The team discovered a number of emulator systems for the Apple II. The following list 
includes general observations about each emulation application. 
 
Apple PC (http://www.zophar.net/apple2/applepc.html) emulates the Apple II+, Apple 
IIe, and Apple IIc computers. The emulator does not include the Apple II ROM, but the 
user community provides appropriate sources for downloading the ROM images. Apple PC 
is designed to run on a DOS system, and can run on DOS emulators such as DosBox. The 
software cannot run on limited DOS environments such as those available in Vista and 
Windows 7. Apple PC emulates monochrome green and white monitors, and provides 
sound from the PC speaker, Sound Blaster, or MockingBoard options. The emulation allows 
for the mouse to substitute as a joystick. Although the software does not require explicit 
installation, it can be run from a directory on the hard drive. Help is included, and the 
system includes an integrated debugger. Apple PC uses an inverted capslock scheme, which, 
when utilized, will persist in lowercase mode appropriate to later Apple models. Backspace 
and arrow key functions, along with cursor representation, are mapped such that backspace 
and left arrow move the cursor left. Apple PC is incapable of running Mindwheel. 
 
 
 
Appler (http://www.zophar.net/apple2/appler.html) is predominantly an Apple II+ 
emulator, although the README file enumerates several hardware options that span 
different Apple II class revisions. Appler is designed to run on a minimal 386 system, with at 
least 1MB RAM memory, 128K EGA graphics adapter capability, and MS-DOS 3.30 
operating system. Appler will not run on newer versions of Windows including Vista and 
Windows 7 without a DOS emulation environment such as DosBox. The emulator includes 
the Apple II ROM and supports a monochrome white monitor mode for text applications 
and a color (green, white, magenta) output for graphical applications. Appler supports sound 
and does not require any special installation. The emulator comes with help and includes an 
integrated debugger. Keyboard interaction varies with the application. For Mindwheel, the 
backspace and left arrow both perform the backspace, and delete does not perform any 
operation. In Mystery House, backspace and left arrow move the cursor back without erasing 
? 71?
 the previously typed text. Appler uses a different font than the standard Apple II character 
generator. In addition, there are issues with combined text and graphics renderings as well as 
issues with running Mindwheel. 
 
  
 
 
AppleWin (http://developer.berlios.de/projects/applewin/appler) by far was the best Apple 
emulator tested on the PC platform. Distributed under a GNU license, the emulator 
supports Apple II, II+, and IIe computers and includes the Apple II ROM. The emulator 
runs on Windows platforms (tested with Windows XP, Vista, and Windows 7 RC). The 
emulator supports a wide range of monitor formats, including four monochrome variants 
(white, green, amber, and custom), as well as color modes (standard, TV, text-enhanced, and 
half-shift). Sound is available with Mockingboard and Phasor options. AppleWin differs 
? 72?
 from other emulators in that it has a performance mode that allows for ?authentic? timings 
and run speeds as well as optimized modes that allow it to take advantage of the host 
hardware. AppleWin requires installation, and includes extensive help, information on the 
Apple platform, and a complete debugger. AppleWin supports a toggle capslock mode that 
starts in uppercase. In Mindwheel, backspace, left arrow, and delete all perform a backspace. 
In Mystery House, backspace and left arrow both move the cursor left. Delete replaces the 
highlighted character with a blank and the image of the cursor does not blink. 
 
 
 
 
 
 
? 73?
 Virtual Apple (http://www.virtualapple.org) is a browser-based emulator. The emulator is 
capable of executing on both Mac and Windows platforms, and is supported by popular 
browsers such as Internet Explorer, Firefox, Safari, and Chrome. The application emulates 
sound and provides two monitor modes ? a white on blue for text, and a green, white, 
magenta mode for images. There is also a Java implementation that can be run from the 
browser. There is minimal help and no debugging support. Capslock and keyboard reactions 
are fairly atypical, even when compared to other Apple II emulators. 
 
 
 
 
 
? 74?
 Virtual II (http://www.xs4all.nl/~gp/VirtualII/) is an emulator designed to run on the Mac 
platform. The emulator supports Apple II, Apple II+, and Apple IIe computers. It does not 
come with Apple II ROMs, requiring the user to search the Internet for an appropriate 
source or to extract the information from their own ROM set. The emulator requires at least 
the Mac OSX 10.4 operating system. The emulator allows for monochrome and color video, 
and supports a mode to toggle scan lines on and off. The emulator supports audio including 
Mockingboard emulations. Options allow for controlling refresh rate as well as CPU speed. 
Virtual II requires installation, includes help files, and a debugger. To emulate flexible 
keyboard behavior, Virtual II has the ?soft caps lock? option. 
 
 
 
 
 
? 75?
 By far, the greatest problem the team noted with emulation technology for the Apple II was 
the inconsistent use of a standard keyboard metaphor when dealing with the Apple II/II+ 
keyboard. The Apple II/II+ keyboard worked primarily with uppercase encodings, with 
such machines only allowing key font generation of the uppercase glyphs when typed for 
certain applications. The inconsistency among emulators leads to inconsistent behavior that 
is not representative of the original experience. Also, the difference in emulators means that 
emulator users should explore the keyboard behaviors, or else they could observe 
unexpected results in the play experience. 
 
Case Study 3 - DOOM 
 
DOOM, by id Software, is universally synonymous with the first-person shooter genre of 
video games. Developed in 1993 by John Carmack and John Romero, the game story 
involves a space marine (you, the player), who is dispatched to Mars? moon Phobos, to 
contain a dangerous situation created by the Union Aerospace Conglomerate, a mega-
 corporation responsible for processing radioactive waste and supplying the military. The 
game is known for its graphic nature, supporting a range of weapons as well as stylized 
violence. The goal for players was to navigate levels searching for keycards, which in turn 
allow access to additional areas of the game. Along the way, players have to both avoid and 
fight various monsters with a range of abilities and strengths. DOOM was groundbreaking in 
advancing several techniques for game engine design, including 2.5D game level design, 
better handling of BSP, complete texture mapping for room surfaces, and non-perpendicular 
space design. DOOM also provided a means by which the game could be experienced in a 
single-player mode as well as a multi-player mode over a network or a modem. 
 
DOOM was designed to run within an IBM PC environment. Its minimum specifications 
stated that it could be played on a system with an Intel 80386 or compatible processor, 4MB 
of RAM memory, a VGA graphics card, and a hard disk drive. The preferred specifications 
allowed for an Intel 80486 processor, a Sound Blaster Pro or Sound Blaster Pro compatible 
sound card, and a network card that could support IPX network protocols (with a terminate 
and stay resident application or DOS network driver stack). Although an operating system is 
not specified, the README file contained on the distribution media mentions setup issues 
with MS-DOS 5.0 and before, MS-DOS 6, and OS/2. 
 
With the advent of PC compatible clones in the 386/486 era, there is no standard reference 
platform for the target system. For this era system, motherboards could support 8MB to 
32MB of system RAM, support a PC compatible BIOS, and a number of expansion slots for 
ISA, VESA, EISA peripherals. Common peripherals included VGA cards that conformed to 
VESA specification, coax-based Ethernet cards, modems, drive controllers, COM ports, 
parallel ports, joystick ports, and multifunction IO adaptors for floppy drives and hard disks. 
Unlike modern PCs, most peripherals required that the user set through jumpers, 
appropriate IO addresses, memory addresses, interrupt lines, and DMA address channels. 
Jumper configuration problems could cause unintended consequences, including machine 
lockups, system crashes, and indeterminate behavior. When running many applications, the 
user would have to provide known settings for peripherals to the program in order for it to 
operate effectively. 
 
? 76?
 The team was able to secure an older 486/DX50 system for testing an original DOOM 
distribution (version 1.8 on both Floppy and CD-ROM [Shareware/Retail]). 
 
Before selecting virtualization and emulation technologies, the team examined two different 
ways of providing an operating system environment for the game. One choice was to use a 
Microsoft MS-DOS 6.22 distribution from floppy disk. The MS-DOS option is based upon 
Microsoft?s closed source operating system. Although Microsoft does not directly support 
the sales of MS-DOS, it is still available through the Microsoft Academic Alliance for 
educational use. The second choice was to use FreeDOS (http://www.freedos.org), an open 
source alternative to MS-DOS. FreeDOS has GPL licensing and was developed for both 
modern and legacy systems. FreeDOS can be used to provide DOS functionality for the 
Windows NT line of operating systems, Windows XP, Vista, and Windows 7. FreeDOS 
supports such features as FAT32 file systems, disk cache, memory management (EMS and 
XMS), mouse extensions, and multimedia extensions for CD ROM.  
 
The team worked with a number of virtualization and emulation applications for DOOM. 
The following is an overview of the various systems, detailing what worked and what did 
not.  
 
Virtual PC 2004 and Virtual Machine 2007 (http://www.microsoft.com/windows/virtualpc) 
are virtualization systems promoted by Microsoft that provide virtualization support for a 
Pentium II processor and/or the virtualized CPU of the host platform. Virtualization device 
mapping provided a uniform peripheral architecture including a S3 Trio 32 VESA card, an 
AMI BIOS, a Sound Blaster 16, and DEC 21XXX series Ethernet solution. Virtual PC 2004 
was able to run on Windows XP and Vista, but not Windows 7 RC. Virtual PC 2007 could 
run on all the Windows variants that were tested. Both versions of Virtual PC were able to 
run MS-DOS as well as FreeDOS, and supported the DOOM installation process. Video 
worked nicely in both versions. Sound effects were selectable for PC Speaker and Sound 
Blaster solutions. Music worked with SoundBlaster, Adlib, and Pro Audio Spectrum 
selections. For Virtual PC 2007, DOOM would tend to freeze on all tested operating systems 
except Windows 7 RC 64 bit. After several attempts, the team was unable to create a 
reproducible freeze point: the DOOM application would freeze the Virtual PC application at 
indeterminate points in time. 
 
DOSEmu (http://www.dosemu.org) is a combination hardware virtualization and emulation 
solution capable of running classic operating systems such as MS-DOS, FreeDOS, and DR-
 DOS. The program is available for Linux systems and provides a GPL license for its use. 
DOSEmu provides partial support for sound, with current workarounds for Sound Blaster 
music compatibility issues. DOSEmu was able to work with all five of the Linux test 
platforms. The distribution used by the team automatically installed FreeDOS as the target 
operating system. Although DOSEmu is supposedly capable of mapping to a MPU-401 
MIDI card, the testers were unable to reproduce this mapping successfully. Therefore, on all 
of the host platforms, music did not work. However, sound effects did work in speaker and 
Sound Blaster mode, except for the Gentoo build, which does not support standard audio 
device mapping in the base distribution set.  
 
QEMU (http://www.qemu.org) has aspects of both a virtualization platform and an 
emulation system, and is a contributing component for a variety of other VM solutions, 
? 77?
 including VirtualBox, Xen, and KVM. QEMU was deployed on all five of the host Linux 
systems. In all cases, it was able to install both MS-DOS and FreeDOS, and was able to load 
the DOOM game software. For sound effects, QEMU was able to provide PC Speaker 
access and Sound Blaster functionality. The sound support did not properly work on 
Gentoo and Fedora due to audio device mapping. Music was supported for Adlib, Sound 
Blaster, and Pro Audio Spectrum sound solutions. Fedora had some difficulties mapping the 
arrow keys in Fedora. Problems with Fedora also led to some situations where the game 
would manage to get into an unplayable state.  
 
VirtualBox (http://www.virtualbox.org) is a commercial-grade virtualization solution that 
was offered by Sun Microsystems and was recently acquired by Oracle along with Sun?s 
other operational assets. VirtualBox is distributed in both a commercial version as well as an 
open source edition. VirtualBox can virtualize a wide range of peripherals and provides 
native virtual execution, a just-in-time dynamic recompiler, and hardware-accelerated 
virtualization features. VirtualBox can be run on a wide range of host operating systems. In 
the testing phase for VirtualBox, the team was able to successfully run the environment 
under all Windows variants, as well as Ubuntu and Gentoo Linux. VirtualBox was unable to 
execute under a Fedora base installation due to missing files, and was unable to run under 
Debian and Arch due to VM initialization problems. Unfortunately, VirtualBox had many 
problems running DOOM: the virtualization process for Windows platforms made the game 
very slow on Windows variants. To verify this was not a problem with the installation: text-
 based applications that were run did not experience the same problems as in the running of 
DOOM. Although the Sound Blaster virtualization allowed for sound effects, music 
selections through the Sound Blaster and other devices were ignored, regardless of the sound 
card settings.  
 
The final virtualization test was with VMWare (http://www.vmware.com), a company 
dedicated to the creation of virtualization technology. VMWare was used as a baseline to 
compare the previously tested virtualization solutions. VMWare worked with the various 
versions of Windows, but worked only with Ubuntu and Debian Linux distributions. Arch, 
Gentoo, and Fedora were unable to provide re-compilation for the virtualization tools and 
device mappings. For all distributions, VMWare experienced problems with crashes and 
issues with sound virtualization. As such, VMWare did not provide a suitable execution 
environment for DOOM. 
 
The final application we tried was Peter Veenstra?s DOSBox. DOSBox is actually an 
emulator that provides an environment analogous to an IBM computer with an MS-DOS 
execution environment. The emulator provides a wide range of video support, from original 
Hercules monochrome, through CGA, EGA, VGA, VESA, and S3 Trio 64. For audio, the 
emulator supports emulation of the PC Speaker, Sound Blaster, MPU-401, Disney Sound 
Source, Gravis Ultrasound, AdLib, and Pro Audio Spectrum. The emulator has support for 
network emulation as well as advanced features for screen and movie captures. DOSBox 
does not require the user to own a Microsoft copy of MS-DOS. As such, FreeDOS is not 
applicable to this environment. In testing, DOSBox ran within every host operating system 
the team tried, including all Windows and Linux variations. DOSBox performed the best for 
the team, having only a freeze problem under Linux with the WaveBlaster sound driver set. 
By far, DOSBox provided the best experience for DOOM compared to the prior 
technologies. 
? 78?
  
There were virtualization problems with both MS-DOS and FreeDOS DOOM scenarios, 
including intermittent crashes, freezes, and sound mapping problems. The pure emulation 
solution created the best experience for the game.  
 
Analysis of Migration 
 
Another important part of the technology of preservation is the process of migration, the 
process of moving code from one target platform to a newer or better target platform that is 
capable of running the software application. There are two types of migration that one must 
consider for the preservation of virtual worlds and video games: source migration and binary 
migration. 
 
 
 
Source migration, or ?porting,? is possible if one has access to the original source code and 
assets for a game experience. The process of source migration is usually easier on equipment 
that possesses similar CPU series and related platform architectures, but can be performed in 
cases where the source and target architectures differ. To perform source migration, one 
must possess appropriate tools, such as interpreters, compilers, and assemblers that are 
capable of not only reading the source programming but also are capable of targeting the 
new environment. Along with the tools, the re-compilation process requires access to 
libraries and APIs related to graphics, user input, operating systems and firmware access, 
math functions, and data parsing. In some cases, source migration may require the original 
source code for the libraries, as well as the game. In other cases, the operating system of the 
API set may provide appropriate wrappers for legacy code (e.g. DirectX SDK support for 
previous versions). If such libraries are not available, one may have to rely upon conversions 
or library ports from the open source community. Along with the code, digital assets might 
need some special attention in the source migration process. At the simplest level, the 
? 79?
 original code and libraries may be able to use the content directly, and, as such, very little 
might have to be done. For example, in migrating image manipulation code, libraries will 
allow for older image formats such as BMP and JPEG to be loaded, but special attention 
must be paid to elements such as image bit depth, alpha channel designation, bit encoding, 
as well as file name format and case sensitivity. 
 
Binary migration is possible if one has access to binary and assets of the original game. 
Binary migration is usually only possible in systems where the CPU series and platform 
architecture is backward compatible with the legacy system, for example, the execution of a 
PC compatible Intel 386 binary application on a PC compatible Intel Pentium IV system. 
Although some binary game applications are self-contained, others rely upon dynamic 
libraries and modules provided by the target operating system or platform environment. For 
example, DirectX 9 Games are reliant upon a Windows operating system as well as the 
presence of the DirectX 9 run-time environment. Such games are often packaged with an 
integrated or stand-alone installer for the DirectX 9 components. Binary migration can be 
susceptible to small changes in computer architecture, changes in peripheral devices, CPU 
optimizations, and changes in BIOS and firmware code.  
 
Another difficulty encountered within the binary migration process is the conversion from 
original storage and execution media to newer media that can be accessed by modern 
devices. Since the advent of the personal computer, media storage formats have changed 
frequently. Tapes were replaced by magnetic disks, which gave way to optical disks and, 
most recently, to solid-state flash memory. Each format has numerous variants, and each 
variant requires its own drive connected to a computer through a series of different ports 
and cables often made obsolete within two years of introduction to the market. As a result, 
the cultural record of the last 30 years is scattered over a variety of media that cannot be read 
by humans (or modern computers) without the aid of obsolete and hard to find machinery. 
 
One of our test cases for this project was the preservation of Mindwheel, a piece of interactive 
fiction written in part by U.S. Poet Laureate Robert Pinsky (see also Case Study 2 in Section 
6). We obtained a Macintosh copy of this title, but it came to us on the original 400K 3.25" 
floppy disk. While, as of this writing, USB connectable drives capable of reading 1.44 MB 
disks are common, these devices are unable to read 800K or 400K. Fortunately, we had 
access to a Macintosh Powerbook G3 ?Wallstreet edition? that had both a floppy drive 
capable of reading the 800K disks as well as an Iomega Zip drive.  
 
Using this machine, along with an external USB Zip drive, we created an image of the floppy 
on a Zip disk using the Powerbook, transferred the image to the external Zip drive, and 
thence to a contemporary iMac. Unfortunately, the disk image was in the now obsolete MFS 
format and so accessing the individual files via a contemporary machine remained difficult. 
We solved the problem in two ways:  the first involved opening the disk image in the 
popular MacPlus emulator, vMac, transferring the files to a virtual disk image formatted in 
HFS (readable by both the emulator and a modern operating system). We later found the 
?MFSLives? software provided by Apple and were able to open the image within the native 
operating system. 
 
The chains of transfers from medium to medium and platform to platform described above 
are common in digital preservation projects, but the fragility of the links that compose them 
? 80?
 is palpable. The process described above depended, ultimately, on an aging laptop, available 
only through serendipity. As the chronological distance between the date of creation and the 
encounter in the archive grows, accessing old store media will likely become even more 
difficult. Already, formats even as common as the once ubiquitous 5.25" floppy disk present 
significant obstacles to preservation and access. 
 
Most scholarly work in born digital preservation to date has focused primarily on the 
problem of archiving, cataloging, and accessing material that has been transferred to modern 
hardware. The problem of copying data from older machinery has largely been solved 
informally on an ad hoc basis. Obsolete drives are pulled out of storage closets and 
connected to computers old enough to have the requisite ports but new enough to accept 
more modern connections. Hobbyists build custom-made cables online for connecting old 
drives to modern computers and sometimes post their schematics on the web, but their 
documentation is often understandably amateurish and rarely maintained. 
 
There are, however, a few well-supported 
solutions upon which this project will build. 
Our experience demonstrates that some 
computers built at transitional moments in 
the evolution of media technology can 
serve as ?Rosetta Stone? machines, 
translating the archaic and forgotten to a 
form recognizable by the modern. Of these, 
the Macintosh Wallstreet edition 
Powerbook G3 is among the best. The 
laptop, manufactured between May and 
September 1998, came with swappable CD, 
DVD, and floppy drives capable of reading 
800K and 400K disks. A swappable zip drive could also be purchased for the machine. An 
Ethernet port further allows data to be transferred from the computer using standard 
networking protocols, and PCMCIA slots permit the addition of USB ports through a third 
party card to which an external hard drive or even flash media can be attached. The 
hardware is capable of supporting older versions of Linux, and with it many contemporary 
open source software packages. The machine does not, of course, natively support 5.25" 
floppies or other more archaic formats, but it does serve as an example of the sorts of 
machines that may prove very valuable to 
digital preservation laboratories in the 
future. 
 
One of the best commercial tools produced 
in the present day for accessing old media is 
the Catweasel manufactured by the German 
company Individual Computers. The 
Catweasel is an internal PCI-card that allows 
more modern drives to access older formats 
(including those manufactured by 
Commodore). Unfortunately, it requires the 
end user to have an internal floppy drive of 
? 81?
 the appropriate size as well as an internal PCI slot, thereby all but limiting the tool to use in 
desktop computers. Digital archival and recovery work must often be performed using 
portable workstations that can easily be brought to the archival site. In these purposes, 
Device Side Data?s FC5025 USB 5.25" floppy controller, released in February 2010, provides 
most of the same functionality as the Catweasel but with an external USB connection. 
  
For 800K and 400K disks, much of the Catweasel?s functionality can be replicated with the 
robust OmniFlop software, developed and released as freeware by Jason Watton of the 
British company Sherlock Consulting Limited. OmniFlop replaces the default drivers in 
Windows XP with custom-made software that controls the mechanisms to permit standard 
floppy drives to read almost any format. Like the Catweasel, it requires a computer with an 
internal floppy drive (USB floppy drives are not usually compatible with the software). 
 
If one has access to the original machinery, 
it is sometimes possible to connect early 
personal computers to a modem (originally 
used for dialing up bulletin board systems 
or commercial services like CompuServe or, 
later, America Online). With the 
appropriate software, and a working old 
computer with the appropriate drive and a 
compatible modem, it is possible to send 
files to any other computer using what is 
called a null modem connection. Finding all 
of the necessary components is often 
difficult today, however, and one must assume they will become increasingly rare in the 
future. The procedure is a useful one, then, but not ultimately sustainable. 
 
In January 2009, Nate Lawson, a consultant at a California digital security company called 
Root Labs, announced a USB connection to a Commodore 1541 drive using what he calls 
the xum1541 cable. The device is an optimization of the xu1541, originally designed by Till 
Harbaum in 2007, but later adopted by Spiro R. Trikaliotis of the Institut f?r Automation 
und Kommunikation in Magdeburg, Germany. The xu1541 project includes schematics for 
building a USB connection to the 1541 using components that the author claims will cost 
less than 5 euros. Lawson?s updated xum1541 device is not yet available for purchase, nor 
are his schematics available online, but a YouTube video gives enough information about the 
component pieces to suggest an approach for building similar cables for the 1541 and other 
early 5.25" drives. 
While the amount of preservation work that has been accomplished even under such less 
than ideal circumstances is remarkable, a formalized and repeatable set of procedures are 
needed, and needed immediately. With every passing day, the problem becomes increasingly 
urgent as drives fail and 5.25" disks crumble like papyrus, and with them disappear large 
swathes of our creative history. 
 
? 82?
  
Single density Commodore 1541/Apple  II series ?flippy? 5.25? disks 
 
According to the 2003 Guinness World Records, the Commodore 64 is the best selling 
computer of all time. With over 30 million units sold, the machine represented for many an 
introduction to the personal computer. Although it did not originally ship with a disk drive 
(plug-in ROM cartridges were the only media that could be used with the machine out of the 
box), many users quickly purchased the 1541 drive. The drive, similar to the one Apple 
manufactured for its Apple II computer, could read only one side of the disk at a time, 
requiring users to physically turn the disk over and reinsert it into the drive to access the 
other side (hence the colloquial name, ?flippy? disks). The drive thus operated differently 
than later IBM-compatible drives (and for that matter later Commodore drives), which, 
thanks to the addition of another drive head, could read both sides of the disk at once. 
Given the staggering number of Commodores and Apples sold, developing an access 
strategy for the media format is at the top of our priority list. 
 
Double density, IBM-formatted 5.25" disks 
 
Although the Commodore and Apple II enjoyed nearly a decade of dominance in the home 
markets, business tended to use instead IBM and IBM compatibles. The most common 
5.25" drives on these machines could read both sides of a floppy disk simultaneously, and 
could access a higher number of tracks on a side (hence the designation, double density).  
 
Macintosh 800K/400K floppy drives 
 
The 3.5" drives on early Macintosh computers were formatted slightly differently than those 
used by DOS/Windows machines and later Macs. Unlike the 1.4MB disks which can be read 
with a (currently) commercially available USB drive, the disks used by these drives cannot be 
read using modern equipment. Many writers and artists who used Macintosh computers in 
the early to mid-1990s have saved their material on these floppies. 
 
? 83?
 3" Amstrad Drives 
 
Although never terribly popular in the United States, the U.K.-based electronics company 
Amstrad produced a series of personal computers, including the Spectrum series, which were 
very popular in Europe in the 1980s. These machines frequently used a proprietary 3" disk 
format that is now beginning to appear frequently in the collections of British authors. 
 
 
 
SyQuest Optical Disks 
 
In the mid-1990s, a company called SyQuest produced a series of ?removable hard disks? 
that vaguely resembled a compact disk permanently sealed in a plastic jewel case. These disks 
were read by a special drive that usually connected to a computer via a SCSI port. The drives 
were most popular with Mac users, and their disks can be found in several publicly held 
collections of artists? papers. Unfortunately, although connecting a SCSI device to a 
contemporary machine is not impossible (although still fairly difficult), there are no drivers 
available for modern computers to control the SyQuest devices. 
 
Analysis of Re-enactment 
 
Another method of preserving a game experience is through the use of technologies that 
support some level of re-enactment process. In this report, re-enactment covers a range of 
possible techniques, including the reconstruction of an experience from historical record and 
game-related artifacts, the reconstruction of the environment through technology, and the 
documentation of game-related experience through secondary sources, such as community 
content, advertisements, popular culture artifacts, and other such paraphernalia. In this 
section, such techniques are examined and the technological impact of the techniques is 
considered. 
 
? 84?
  
By examining such artifacts, two levels of implementation must be reconstructed. At one 
level, the logic and flow of the game experience must be re-created. This includes navigation 
related to title screens, credits, game play instructions, game settings, and game play states. It 
also includes the rules of the game for level presentation, character movement, win and lose 
states, as well as elements related to play difficulty. At another level, the content of the game 
must be reconstructed. This includes game art, animation, video, and sound. Fortunately, in 
some situations, partial information is available that may assist in the reconstruction process, 
especially in cases where a playable version is available (e.g. games that exist, but due to a 
loss of technological information, cannot be migrated to a new system). Such re-enactments 
vary in their accuracy and fidelity in re-creating the original game.  
 
Sometimes re-enactment can be used as an exercise with students who are learning the art of 
video game development. National programs in the field of game development often 
challenge students to re-create classic game experiences such as Asteroids, Space Invaders, and 
text adventure games so that students can gain greater understanding into the logic and fun 
of the game development process. Even as recently as this year with Google?s ?Pac-Man in a 
Browser? experiment, we have seen the impact of experience-based re-enactment. 
 
Another methodology is environment-based re-enactment. This is especially important for 
games that involve server-based technologies or distributed processing in the presentation of 
the play experience. For example, games such as World of Warcraft require a client-side 
component (the game world visualization component, input manager, and user interface 
system) as well as a server-side component (the realm that maintains the object catalog, 
object instances, and player instances). In such systems, if the server-side technology is no 
longer available, the game cannot be experienced. As we look back over the last several 
years, there have been a number of online games that have ceased operation or have 
changed so significantly that the original play experience has been lost (e.g. what was it like 
to play during the first three months of the World of Warcraft private beta experience?) When 
server-based games disappear, there are only a few options available to preserve the 
experience. 
? 85?
  
Environment-based preservation involves the game development company either 
distributing the server-side technology to their player base or involves providing the source 
code to the player community. An example of this approach can be seen with Cyan World?s 
Myst Online title, which was based upon the popularity of the Myst family of exploratory 
games. In April 2008, the Myst Online servers were shut down. However, in December 2008, 
Cyan Worlds announced plans to open source the project. Although this process is still 
ongoing, the possibility of an open source version of the game has re-sparked the player 
community and has already allowed Cyan Worlds to re-introduce some of the prior 
technology to the public. 
 
Another form of environment-based preservation involves the underground development 
community. There have been numerous examples of developers creating their own server-
 side technology so that a server could be hosted on a local server and a private version of the 
game (with a small group of friends) could be experienced. Such attempts include the UXO 
project for Ultima Online and the Mangos project for World of Warcraft. These projects have 
resulted in various levels of response from game companies, including active encouragement, 
non-committal acknowledgement, and legal cease-and-desist responses.  
 
Environment-based preservation can also be accomplished through software designed to 
process and re-create the environment in a stand-alone application. Such scenarios do not 
provide the ability for multi-user interactions, but instead allow for an interactive snapshot 
of what the game would look like at a precise moment in time. The technique involves the 
capture of game assets as well as information related to the position, orientation, and state of 
the game objects. 
 
The final methodology is artifact-based re-enactment. This approach involves the use of 
secondary sources to define the game experience. Key artifacts for this approach include 
video, images, audio, and first-hand testimonials from players and player communities. In 
addition, information from trade magazines, conventions, reviewers, commercial outlets, and 
other sources can supplement critical information. Popular cultural artifacts such as TV 
commercials, music, spin-off media (Saturday morning cartoons), merchandise, and other 
items can provide appropriate context for the time period and the experience. 
 
Digital Forensics and Disk Image Analysis 
 
Digital forensics is a rapidly emerging field that encompasses e-discovery, intrusion detection 
and incident response, data recovery, and the packaging and presentation of digital evidence 
to standards admissible in various legal settings. As libraries and archives acquire more and 
more data in born-digital formats, tools, techniques, and workflows from digital forensics 
offer solutions and technologies for processing this challenging new kind of material. 
Stanford University Library, along with the British Library and the Bodleian in the UK, are 
currently applying forensics hardware and software in exactly this role. See Kirschenbaum, 
Ovenden, and Redwine?s forthcoming report (2010) for a more comprehensive introduction 
to the field. 
 
The bedrock of all digital forensic techniques is the concept of a disk image, a bitstream 
representation of every bit of information originally stored on an instance of physical media. 
? 86?
 Bitstream imaging bypasses the file system, thereby allowing an archivist to capture 
information stored in slack space at the ?end? of a file or data clusters flagged for 
reallocation (the typical fate of supposedly ?deleted? material). A technique known as data 
carving allows an analyst to reconstruct the pieces of a file from the bitstream. File types can 
be identified by tell-tale headers (if they remain intact), allowing for some reconstruction of 
the file even it cannot be reconstituted in its entirety. More exotic techniques may allow for 
the recovery of information from data tracks that have already been overwritten at least 
once, but these are not within the realm of practicality for libraries and collecting 
institutions. 
 
Disk images have circulated freely in player communities and various Internet venues, 
including retro gaming forums and abandonware hubs. Like cartridge-based ROMs, the 
availability of disk images allows users to run older games via emulators. Sometimes, 
however, the disk image carries additional freight. A report on a web forum details how one 
player used an image of an old freeware demo disk to locate source files for the first 
Commander Keen game, an episodic series of side-scrollers published by id Software. Residual 
data captured as part of the image allowed the player to reconstruct a near-complete running 
copy of the original game: 
 
Most of the Keen 2 data had been overwritten by the sample games, but to my 
surprise most of the Keen 1 data was still intact and after scratching together the 
deleted files I managed to get almost a fully running game. Only EGAHEAD.CK1, 
EGALATCH.CK1, EGASPRIT.CK1, HELPTEXT.CK1 and LEVEL01.CK1 had 
been completely erased (well LEVEL01.CK1 had about 50% of it's [sic] data there 
but I didn?t bother combining it with the finalized one), so I took these missing files 
from version 1.0 of the game. 
 
Now I was a bit disappointed that I didn't get the graphics files together (usually 
graphics are what's most interesting in development versions of games), but it seems 
the other game files were actually quite different at this point as well. I don't know at 
exactly which stage the game is in this build but it at least looks pretty close to final 
with a bunch of minor level design differences in quite many levels. 
The story section is no doubt the most interesting difference, you actually get a scene 
where a Vorticon comes to your ship and takes the vital parts and then leaves 
(explaining why these tiles without and with the parts exist in the final game). 
Other points of interest is the lack of the GOD cheat and the fact that you can't 
complete the game. (http://www.pckf.com/viewtopic.php?t=1211) 
 
Large-scale, systematic analysis of disk images circulating within the player community may 
therefore provide an important tool for recapturing and reconstituting supposedly ?lost? 
works or versions of works. 
 
Nonetheless, it is important to recognize that even a disk image, while somewhat akin to a 
?facsimile? in the realm of photographic reproduction, is an abstraction, a projection (or, 
more precisely, an interpretation) of physical phenomena on the surface of the disk. Creating 
a disk image is not a homogeneous process, and not all imaging tools will acquire the same 
data. Some disk image formats, for example, will not indicate the presence of damaged data 
sectors. It is therefore easy to fetishize a disk image as a more authentic version of the data 
? 87?
 than just extracting the files and dumping them on a modern file system. A more useful 
understanding, however, would be that forensics gives you better documentation about how 
you reproduced a copy of the original data, not (necessarily) a superior copy. 
 
In one sense, digital games and virtual worlds present no distinct challenges or opportunities 
for applications of forensics, since the same techniques for data acquisition and analysis that 
are used with other forms of digital content can be applied just as seamlessly. Whether or 
not forensic methods are useful in a particular institutional setting will be determined by a 
cost-benefit analysis of time, resources, specialty training, and the kind of access and services 
it wishes to provide to its patrons. Techniques for authenticating digital content may, 
however, have a larger role to play in appraising material collected from the player 
community. Demos, speed runs, and other forms of player-generated content are 
occasionally subject to question or controversy regarding authenticity and provenance. In 
these cases, forensic tools and techniques can prove uniquely useful.  
 
Recommendations on Preservation Strategies 
 
The team has provided several recommendations that must be addressed in the preservation 
process. These recommendations are not only derived from the emulation, virtualization, 
and migration processes from the grant, they are also derived from interactions related to the 
metadata encoding process undertaken by the research team. The recommendations are as 
follows: 
 
? It is important for museums and other collectors to start the exploration process into 
emulation, virtualization, and migration techniques as one facet of the preservation 
process. 
? It is important to encode information related to game play based upon original 
platform, emulation, and virtualization information within preservation metadata. 
? There must be a mechanism to encode a testing rubric for virtualization and 
emulation into a metadata set, such that there is a template for the range of 
permissible options that is easily searchable and can be indexed. In addition, the 
testing rubric must contain mechanisms for freeform comments related to the 
emulation and virtualization process. 
? Information related to virtualization and emulation must be extensible and there 
must be easy mechanisms by which emulation and virtualization information can be 
updated across collections of metadata. Information update should be updatable 
based upon the growth and demise of emulation and virtualization community 
efforts. 
? In cases where re-enactment is necessary, there must be means of collecting and 
indexing metadata and content based upon re-enactment data. Since this information 
can be more expansive than the original game, there must be mechanisms to 
minimize repetition and support important artifact derivation. 
 
In looking towards the future, agencies such as the Library of Congress will want to fund 
exploration in a number of areas if the process of emulation, virtualization, and migration is 
to be an important part of the preservation strategy. Initiatives should be based upon certain 
guidelines: 
 
? 88?
 ? It is vital to encourage the dissemination of information to help museums and 
collectors understand the importance of emulation, virtualization, and migration. 
? Agencies must help support the development of emulation platforms for the express 
purpose of emulation. The goal would be to unite various grassroots development 
teams into a larger community dedicated to the preservation mission. 
? Agencies must help larger virtualization companies understand the importance of 
peripheral virtualization in the preservation stratagem. The effort should include the 
encouragement of developing adequate video and sound driver solutions, as well as 
creating avenues for developers to extend the virtualized device specification in an 
easy to understand format. 
? Agencies must help metadata format creators understand the need for a consistent 
way of encoding emulation, virtualization, and migration data into data set. Standard 
metadata encoding schemes need to be re-evaluated and extended to support such 
function. 
? Agencies must encourage game engine developers and producers to find 
technological and interaction solutions to re-enacting and re-scripting typical 
scenarios within games and engines. 
? 89?
 7. When Strategies Fail: The Case of  Second Life 
 
Second Life is a virtual world where users create nearly all of the content within the world, and 
also create most of the social, political and economic activity. While the social, political and 
economic activities may be the most interesting aspects of Second Life, any attempt to 
preserve those must in part rely on preserving the technological environment providing the 
context in which they occur. For our project, then, we wished to investigate the feasibility of 
preserving that environment, and, if possible, doing so in such a way as to avoid relying on 
the proprietary infrastructure of Linden Lab, if for no other reason than the fact that Linden 
simply does not possess the resources at this point to serve as a long-term preservation 
repository for archived states of their own virtual territories (known as ?islands?). As we also 
lacked the resources to try to tackle all of the various islands in Second Life, we choose to 
focus on a subset of five for our archiving efforts: 
 
? Life Squared ? A reincarnation of the archive of artist Lynn Hershman Leeson 
housed in Stanford University Libraries? Special Collections; 
? Stanford University Libraries ? a virtual library space established by Stanford 
University Libraries to support online collaboration, education and exhibits; 
? Democracy Island ? A project of New York Law School devoted to overcoming 
?some of the difficulties associated with civic participation ? in real space?; 
? International Spaceflight Museum ? a virtual museum of spacecraft and space travel 
managed by a non-profit corporation; and 
? Rumsey Historical Maps in Second Life ? a virtual exhibit of selections from the David 
Rumsey Map collection. 
 
Problems in Archiving Second Life 
 
Second Life is by far the most technologically complex game (if that is even an appropriate 
word for it) with which our project dealt, a networked, virtual environment based upon a 
client/server architecture involving multiple task-specific server components: 
 
? Login Server ? manages user authentication and login processes;  
? User Server ? manages instant messaging sessions;  
? Space Server ? manages the routing of messages between residents based upon their 
location in the virtual space;  
? Data Server ? manages connections to the various databases containing Second Life?s 
data and log information; and  
? Simulators ? each simulator manages the state of a single region in Second Life, 
including the state of both objects and the terrain and managing the simulation of 
physics for the region. 
 
Linden Lab has made the client software for Second Life available under the terms of the 
GNU General Public License (version 2.0), with the associated artwork licensed under the 
Creative Commons Attribution/Share-alike 3.0 license. All software for the server 
components is closed source. 
 
? 90?
 Second Life is organized into different 256 x 256 meter regions known as islands, or sims 
(short for ?simulators?). A region can be owned either by Linden Lab or by one of Second 
Life?s players, known as residents. The contents of a given region consist primarily of a 
combination of 3D objects, graphical texture files that can be overlaid on the 3D objects, 
audio files providing background noises and scripts that enable interactivity with the various 
objects within a region. All of this content is hosted on Linden Lab?s servers, with the 
different servers managing the interactions between residents? client software applications 
and the various regions. 
 
Objects in Second Life are created by linking various shape primitives (or prims). The 
primitives comprise eight basic shapes (box, prism, cylinder, sphere, torus, ring, tube and 
sculpted). Each of these can be further modified from their default shapes. A primitive is 
more complex than its name might imply, as each prim has an associated inventory that can 
contain scripts, notecards, textures and other items. Textures (image files applied to the faces 
of the primitives) give objects the illusion of possessing even more detail. Each object has a 
set of metadata elements associated with it, including the identity of the individual who 
originally created the object (the intellectual property rights holder), the Creator, and the 
individual who has possession of the object, the Owner. 
 
Object scripts, written in the Linden Scripting Language (LSL), are what transform Second 
Life from a beautiful, but static environment into a highly dynamic and interactive one, 
allowing any object to exhibit a wide variety of complex behaviors and accomplish a variety 
of functions within an island. Any island can have thousands of script-bearing objects, each 
object with its own creator and owner. Additionally, textures, scripts and other items in an 
object inventory may be the creations of someone other than the object creator (and may be 
impossible to identify). 
 
Preserving such an environment presents several obstacles. There are technological 
impediments to obtaining access to certain forms of content in Second Life. Due to the 
permissions system employed by Second Life, an object?s inventory (including scripts) is, by 
default, not available to anyone other than the object?s creator (and so cannot be accessed or 
copied by any other resident of Second Life). As a third party (not Linden Lab and not the 
owner of an object), we cannot realistically obtain access to the underlying database of object 
data on Linden Lab?s servers directly to make a complete copy, nor can we access the full 
contents of an island as a user on the system within that particular region. 
 
Intellectual property and contract law present additional obstacles to preservation activity. 
The Terms of Service agreement covering residents? access and use of the Second Life service 
specifically states ?You retain any and all intellectual property rights in content you submit to 
the service.? The Terms of Service also forbid any infringement of intellectual property 
through unauthorized copying of content available through the service, stating that ?You 
must obtain from the applicable Content Providers any necessary license rights in Content 
that you desire to use or access? and ?You agree that you will not copy, transfer, or 
distribute outside the Service any Content that contains any Linden In-World Content, in 
whole or in part or in modified or unmodified form, except as allowed by the Snapshot and 
Machinima Policy, or that infringes or violates any intellectual property rights of Linden Lab, 
other Content Providers, or any third parties.? The combination of Second Life?s contractual 
framework with copyright law make any effort to make a preservation copy of the contents 
? 91?
 of an island without the explicit permission of all of the content creators who have objects 
present within that island illegal. 
 
One final difficulty with archiving Second Life is determining the full scope of what must be 
archived. If we were to manage to archive all of the objects from a given region, including all 
scripts, animations and other inventory content that are typically protected, we would, in 
effect, have managed to archive a ghost town, an empty set of architecture and geography 
with no information about how the space was used or what its inhabitants were like. A static 
snapshot of a world such as Second Life may in some sense serve as a surrogate for the 
original, but it is a poor substitute. Part of the fundamental nature of Second Life is that it is a 
living, evolving, dynamic space. An archived copy of a region provides some documentation 
of what the world was like, but it is hardly a complete set of documentation, and in a very 
real sense it is not and cannot be a complete version of the original. 
 
Basics of Archiving a World 
 
Despite the issues mentioned above, our project decided that we would try to proceed with 
an approach that involved making an archival copy of our regions of interest. We decided 
that our approach should fulfill several criteria. First, it must respect the intellectual property 
of Second Life residents, as well as Linden Lab. It should, if at all possible, produce an archival 
copy that is not tied to the Second Life architecture, but which would allow a copy of a 
region?s contents to be instantiated in multiple different virtual environment platforms. 
While Linden Lab was a partner in our project, we also decided to approach the issue of 
archiving a region as a party without special permissions or access, in effect putting ourselves 
in the position that any library, archive or museum might occupy if they decided to make an 
archival copy of some region of Second Life. 
 
Creating an archival copy of a region in Second Life involved a multistage process. First, we 
created manifests of the various regions? contents and gathered metadata about the objects 
within those regions. Using the objects? Creator metadata, we next contacted each rights 
holder for objects in the regions asking for their permission to archive the content they 
owned. We then downloaded all available information for the objects we had received 
permission to archive, along with additional information about the island itself, and created a 
submission information package suitable for loading into repositories at the University of 
Illinois and Stanford University Libraries containing the complete set of object and region 
information. Details on our process are below. 
 
Creating a Manifest of an Island 
?
 One of the first issues we needed to deal with is identification of the objects to be archived. 
There are potentially hundreds if not thousands of them on each island, and they may be 
invisible, at a high altitude within the virtual space or otherwise difficult to identify. 
Fortunately, Linden Scripting Language (LSL) provides a fairly simple mechanism for 
detecting objects. However, LSL?s limited script memory presented some limitations that 
required a great deal more thought about algorithm design than was initially imagined. 
 
? 92?
 The primary function for sensing objects in LSL is llSensor() or llSensorRepeat() and the 
associated events ?sensor? and ?no_sensor?. The llSensor() function can be imagined as a 
form of virtual radar, sending out a ping, with the event sensor acting as the receptor 
(no_sensor is triggered when nothing matching the criteria is found). The llSensor() function 
has a maximum range of 96 virtual meters, so a single ping would sense a sphere with a 
radius of 96 meters. The spherical nature of llSensor()?s search capabilities presented us with 
a packing issue, however. If we used llSensor() to detect objects every 192 (2 x 96) meters, 
there will be gaps between the spheres (imagine packing ping pong balls into a box, and you 
can visualize the air space that would represent regions we could not detect). Cubes will pack 
perfectly, so we just compute the maximum size for a box bounded by the sphere and stack 
the boxes. The diagram below presents a 2D model of the 3D case. 
 
As you can see there are some overlaps 
in the scan areas when the scans are 
placed so as to focus our scans on 
adjacent cubes, rather than spheres, but 
checking for objects detected twice is 
straightforward. A more significant 
problem is that that the sensors only 
detect the sixteen objects closest to the 
center of the search sphere. So if there 
are seventeen or more objects within the 
space, the probe will not detect any 
object further away than the sixteenth 
object detected. In a case where we 
detect sixteen objects, we will not know 
whether this is because there actually are 
sixteen objects, or whether there are 
more within the space that simply are not 
being reported by the sensor. We can 
compensate for this by adjusting the 
sensor?s detection range down until we 
get less than sixteen objects per scan, but 
we cannot know what the reduced range 
necessary to ensure detection of all 
objects might be. 
 
We chose to adopt a method that involves a progressive subdivision of the search space in 
any case in which a scan detects sixteen objects. If a scan reports sixteen objects within the 
sensor?s range, the sides of the search box are decreased in size by half (so a 136 meter box 
would become 68) and we essentially stack 8 boxes into the original box with more than 16 
objects. We then scan each of those sub boxes. If any of them have 16 or more objects, we 
repeat the process. 
 
!"#$%&?
 (
 !")*?+,%&?
 ( !"#$%&?
 (
 !")*?+,%&?
 (
 !"#$%&?(!"#$%&?(
 !"#$%""&?(#)%$&(*#+",(--.&"#/012(%3(&4&05(678(9&3&0#:
 ;4&0-%))+",(--.&"#/012(#$%"#(?/"&(%3(
 <6=>:?@(9&3&0#3/(#$%"(%?A%$&"3($*B&#(
? 93?
 There is a potential flaw with this method. It is 
possible within Second Life to put more than one 
object in the same location. This makes it 
possible to have 16 or more objects that have 
the same exact coordinates. In these situations, 
if our scanning boxes reach a minimum sensor 
range of 0.01 meter (the minimum shape 
dimension permitted), we exit the function and 
our program returns an error alerting the user 
to the problem. If desired in those cases, a 
manual inventory of the objects can be 
performed. 
 
As each object is detected during our scanning 
process, information about the object?s owner 
and creator is obtained, and our program stores 
that information in an external MySQL 
database. Once all of the objects? intellectual property rights holders have been identified, we 
are ready to embark on the next phase of archiving, obtaining permissions to archive the 
objects from the appropriate rights holders. 
 
Obtaining Permission 
 
After having created a list of the objects within an island, we can then proceed to contact the 
creators of those objects to obtain permission to archive their materials. Given the 
potentially large number of content creators in any given region, we used a semi-automated 
system to obtain permission to archive objects. The avatar names and UUIDs (a unique 
identifier given to every asset in Second Life) of all content creators from a region are retrieved 
from our database, and sent to a Second Life object that acts as a portal for interaction with 
Second Life residents? avatars. The portal receives the avatar information and sends a notecard 
within Second Life to the creators, describing our project, what we want to do with the objects 
they created, and including an in-world teleport link they can use to travel to our portal, 
where they can obtain access to a web interface we have created for granting or denying 
permission to archive their materials.  
 
At the portal, creators use the object?s interface to access the web interface for permissions 
management. The creators are first asked to set up an account, and then they are presented 
with a list of all their objects we wish to archive. The creator may choose to allow us to 
archive all of the objects, some of the objects or none. Creators are also allowed to specify 
an embargo period during which their content will not be made public. The language for 
granting permission to archive is based upon a standard archival deed of gift used by 
Stanford University Libraries that was modified for use on this project (see Appendix G for 
the language used). Once we have a listing of the objects for which we have been granted 
permission to archive, we can embark on actually making an archival copy of those objects. 
?
 !"#$%&?($
 &%$))?($ &%$))?(
 $
 !"#$%&?()*"?$)*+,-,+"+?,#.%?$/011?()*"$?,#?,#$)&"?
 +"."(.,%#?%2?/%&"?.30#?45?%*6"(.$
? 94?
 Copying a World 
 
Overview of Building in Second Life 
 
The fundamental building block in Second Life is the shape primitive or prim. More complex 
objects are built by combining prims of different shapes together similar to a set of building 
blocks or Legos. There are eight prim types used: 
 
? box 
? cylinder 
? prism 
? sphere 
? torus 
? tube 
? ring 
? sculpted 
 
Most of these are self explanatory except for sculpted prims, a newer addition that allows 
builders to create more organic shapes using off-world 3D design software, and then 
importing them into Second Life as 2D UV maps (32 x 32 pixel .tga or .jpg files) that are polar 
coordinate maps of the sculpted objects. The low resolution of the UV maps limits the 
details that can be included and display costs associated with sculpted prims continue to 
make it advantageous to use the other prim types for many applications. 
 
There is a daunting array of parameters that can be used to alter the shape of these prims. 
Such modifications can often extend to such a degree that they render the base shape 
unrecognizable. Parameters cover such aspects of an object as its material composition, 
position, rotation with the coordinate space of the world, size, texture files overlaid on the 
prim, color, reflectivity and other physical parameters. There are also parameters having to 
do with ownership, creation, IP issues and contents of a prim?s inventory. Prims can contain 
other things such as other objects, scripts, animations and so on. 
 
Each face of a prim can be textured and colored independently. Second Life currently 
supports .tga, .jpg and .gif files for texturing with a maximum resolution of 1024 x 1024 
pixels. Textures can be sized and positioned on a prim face repeatedly as well. You can get 
the UUID for each texture, and capture the textures using the GLIntercept software. 
 
It is very common to find scripts inside Second Life prims. Without scripts, Second Life objects 
would just be pretty (or occasionally ugly) pictures. Prims can contain multiple scripts that 
function more or less independently of one another. A person can write scripts that allow 
objects to communication with one another. The Linden Scripting Language actually allows 
objects to do some surprisingly complex things, and given the ability to use HTTP and 
XML-RPC to contact off-world web servers, objects can exhibit extremely sophisticated 
behaviors. 
 
To make more complex shapes or items larger than 10 meters in any dimension, prims are 
linked together into objects. Linked prims can be moved as a unit and can communicate via 
scripts a bit more easily than unlinked prims. No more than 255 prims can be linked 
? 95?
 together in a single object. Many large buildings are therefore composed of many individual 
objects, each containing a number of prims. One prim in any object serves as the root prim. 
All other linked prims can be referred to by their link number. An unlinked prim has a link 
number of 0, the root prim of a linked set has link number of 1, and the rest are numbered 
consecutively based on the order they were added or selected. The root prim is important 
because position, rotation and velocity information for an object are all relative to the root 
prim. Child prim information is also given relative to the root prim. Additionally, the root 
prim is generally where the main scripts are placed, although child prims often contain 
scripts that communicate with scripts in the root prim. 
 
The largest independent building block in Second Life is the Island, Sim or Region. A sim 
measures 256 meters on the x and y axes. A region is defined by its terrain, stored in a .raw 
file that gives the height of the ground, ground textures and the water level. The ground 
textures display somewhat sequentially from lowest to highest terrain height and can be 
controlled semi-independently by specifying different heights for each corner of the sim. 
Additionally, there are sim parameters that determine what functions are permissible within 
the region, and privately held sims often have a covenant that functions like a legal 
document specifying how the sim is meant to be used. Sims can be divided into parcels that 
can also have names, descriptions and parameters that determine what is allowable (if the 
sim parameters allow that for smaller parcels). A standard sim supports a maximum of 
15,000 prims (not objects; the number of objects will depend on how prims are linked). 
 
The software we chose to copy the objects in a region is known as CopyBot. CopyBot is a 
text-based client for Second Life developed by a group of griefers (avatars who disrupt others? 
activities in Second Life) known in Second Life as the Patriotic Nigras (whose motto is ?Ruining 
Your Second Life Since 2006?). The group has a poor reputation in Second Life as its primary 
function was to make content theft and griefing easier. Despite their nefarious reputation, 
they developed what turns out to be a very useful tool for archiving Second Life content, as it 
allows us to export and import Second Life object data to and from an XML file. 
 
The export function downloads all object data into an XML file (see Appendix I for an 
example). It also downloads all of the textures used on a particular object. There are several 
important caveats. First, it breaks all links to intellectual property ownership. The names of 
the content creator and owner do not even appear in the file. Second, it does not gather the 
contents of an object?s inventory, such as scripts, sounds, etc. These have to be collected and 
inserted back into the objects manually, a process that is only feasible when you have the 
appropriate permissions to examine the objects? inventories. 
 
Another restriction of CopyBot is that it is designed to download a single object at a time. In 
order to download all of the objects on an island, we created a program that interacts with 
our database of objects created during the initial scan of an island. This program obtains the 
UUIDs for the objects we have permission to archive, and then sends a request via HTTP to 
a special attachment worn by the CopyBot avatar, directing the CopyBot to move to the 
object of interest and download it. The management program, combined with CopyBot, 
allows us to automate the downloading of all the objects we have permission to archive. 
 
In addition to the data for the individual objects, we need to store the information regarding 
the region and also provide facilities for relating the objects to each other (and individual 
? 96?
 information we may wish to store about the island). For this purpose we created an XML 
schema that stores all of the data required to reinstantiate the objects back into Second Life. 
Having the information stored in an XML document would also allow us to export the 
information into other virtual world platforms, although this would require some 
transformation of the data. The XML Schema is broken into six subdocuments and follows 
this basic structure: 
 
SLBuild.xsd ? the root file for the schema, which stores basic information about the 
island (e.g., the sim owner, a description of the sim) as well as information on all 
objects and their composition prims; 
 
SLShapes.xsd ? this file defines the parameters that influence the shape of each of 
the different prim types; 
 
SLPrimFeatures.xsd ? this file defines non-shape parameters that may be recorded 
for each prim type including color, flexibility, drag, the influence of gravity on the 
object, etc.; 
 
SLPrimTextures.xsd ? this file defines the parameters recording the texture 
information associated with each prim; 
 
SLPrimContents.xsd ? this file defines the information set which may be used to 
record information about the contents of a prim?s inventory; and 
 
SLBasicTypes.xsd ? this file defines basic element types used in other types in the 
schema, such as vectors and rotations. 
 
While our process often does not allow us to obtain complete information about the objects 
in an island (the inventories of primitives in particular are difficult to obtain), the schema 
does allow us to store a complete record of the contents of an island. 
 
Failures in Archiving 
 
While in theory the mechanisms defined above allow us to archive a fairly complete 
representation of an island in Second Life, separately from the Second Life system, and in a 
form that would potentially allow us to reinstantiate the island in another virtual 
environment platform, in practice our efforts can only be described as partially successful at 
best. We encountered a number of difficulties in attempting to archive our test cases in 
Second Life that severely impeded our ability to create a complete record of those worlds. 
 
The greatest impediment was the need to obtain permission from the intellectual property 
owners before we archived their objects. At its best, our response rate to requests for 
permission was 10% of the individuals we contacted on one island, and at worst, we received 
no responses at all. It is reasonable to ask whether an archive that contains only 10% of the 
objects within an island is worth creating at all; it is certainly not an accurate reflection of the 
island as it existed at that time. Moreover, some of our responses were actually hostile. Many 
Second Life builders earn a living from creating objects in Second Life, and they were not 
? 97?
 pleased with the notion that anyone might actually copy their creations and place them in an 
archive where others would be able to obtain access to them. 
 
In the case of one island, we experienced a complete failure in our attempts to archive not 
because of intellectual property issues, but due to a technical one. Democracy Island, a 
project of the New York School of Law, allows any and all Second Life residents to build 
within their island. As a result, the island has been packed with objects to the point where it 
has actually hit the limit on the maximum number of prims allowed within a given region. 
Our approach to archiving does require us to instantiate a probe object within an island to 
perform our initial scan for objects. Because we could not instantiate our probe, we were 
unable to obtain the necessary information to continue our attempts to archive that island. 
 
Another potential problem exists in that regions and parcels within Second Life can be 
declared ?no script? zones. Any scripts bound to an object in Second Life will immediate cease 
when they enter such an area. Our initial probe to identify objects within an island is a script-
 based object. If it accidently enters such a region, it immediately fails. 
 
A final major problem is that, as a third party, we have no access to the inventories of 
objects other than those we have created. While the information we can automatically extract 
from an island allows us to create a visually complete recreation of the objects in an island, 
any animations or scripts associated with the objects are lost. All of the behaviors of objects 
that make Second Life an interesting and dynamic environment are gone. While it is possible 
for the intellectual property owners of the objects to provide us with copies of the contents 
of the objects? inventories, it was difficult enough to simply obtain permission to archive. 
The odds of the IP owners being willing to manually copy all of their objects? inventories for 
us seem remote. 
 
In short, our experiments in trying to archive islands in Second Life at best resulted in 
extremely partial and static representations of the original. While the techniques we?ve 
developed may be useful in archiving some virtual world systems, at least in the case of 
commercial environments such as Second Life, there are severe limits to the preservation 
activity in which a third party can engage. Given the intellectual property and contractual 
restrictions governing Second Life, any hope for a complete archive of a Second Life world 
would rest on Linden Lab?s willingness to archive the content itself. 
 
 
 
? 98?
 8. Packaging Virtual Worlds 
 
Virtual Worlds and the OAIS Reference Model 
 
The Reference Model for an Open Archival Information System (OAIS) (Consultative 
Committee for Space Data Systems, 2002) has been accepted as one of the foundations for 
digital preservation efforts throughout the digital library community. One of the key aspects 
of the OAIS reference model is the concept of an archival information package, the 
complete set of content information and preservation description information necessary to 
maintain a data object as interpretable by the designated community a repository serves over 
the long-term. The OAIS reference model includes a data model for an archival information 
package (see the figure below), and later work by the Consultative Committee for Space 
Data Systems led to an XML packaging format for the creation of archival information 
packages, XML Formatted Data Units (XFDU) (Consultative Committee for Space Data 
Systems, 2008). 
OAIS Archival Information Package Data Model 
  
While the data model set forth in the OAIS reference model provides a structure that can in 
theory be applied to information housed in any archive, the OAIS reference model was 
designed to support the operations of aerospace agencies archives, including data archives. 
While the Consultative Committee for Space Data Systems (2002) did mention software as a 
type of content information that might be saved in an archive, the reference model?s 
examples, particularly in the case of representation information, tend to focus on simpler 
data types. It seems safe to assume that the use cases the OAIS reference model?s creators 
had in mind when designing the model did not include preserving World of Warcraft or 
DOOM 3. Our investigations have found that while the OAIS reference model as it stands 
? 99?
 today is capable of handling game software as an object of preservation, video games and 
interactive fiction are extremely complex digital objects, and packaging them in a manner 
consistent with the OAIS reference model can be a challenge, particularly with regards to the 
matter of providing adequate representation networks and adequate context information. 
 
The OAIS reference model specifies that an archival information package for digital content 
should contain the representation information necessary to convert the bit sequences in the 
digital information into more meaningful information by describing the data format used to 
encode the digital information. It recognizes that a given piece of representation may 
reference other pieces of representation information (as when Unicode references ISO/IEC 
Technical Report 19769 in its discussion of Unicode data types for the C programming 
language), and that representation information in digital form, as an object of preservation in 
its own right, may require its own representation information (e.g., if we stored the Unicode 
standard in PDF file format as representation information for an object, we would need a 
copy of Adobe?s PDF Reference specifying the PDF data format). The full set of 
representation information needed to ensure a digital object?s on-going interpretability can 
thus form a large and complex representation network. 
 
As an example, consider the game Mystery House from our case set. The original source code 
for Mystery House has apparently been permanently lost. Existing versions of the original 
game are encoded as files embedded within an Apple II disk image file. Our representation 
information for the game thus needs to start with a copy of Apple II: The DOS Manual, the 
documentation for the Apple II disk operating system. Fortunately, we have a copy of that 
as a PDF 1.6 file. Obviously, we will then need a copy of the Adobe PDF Reference for 
PDF version 1.6. The PDF Reference references 33 other documents published by Adobe, 
the vast majority of them technical documents needed to understand aspects of the PDF 
specification. It also references 60 documents from other agencies, including a large number 
of standards documents from agencies including the Institute of Electrical and Electronics 
Engineers, the International Color Consortium, the International Electrotechnical 
Commission, the International Organization for Standardization, the International 
Telecommunication Union, the Internet Engineering Task Force and the World Wide Web 
Consortium that define technologies relevant to implementation of the PDF specification. 
Many of these standards documents in turn reference other standards documents. We are 
well on our way to having a representation network that resembles a small technical library, 
and we only have the information needed for preservation of the disk image file format; we 
have not yet considered the files contained within the disk image. 
 
Of course, the size of this representation network could be truncated quite effectively if we 
keep our copy of Apple II: The DOS Manual in print form rather than in PDF format. This 
requires, however, that the wrapper for our archival information package for this game be 
able to successfully reference both digital and non-digital material. And that still leaves the 
issue of the representation information necessary to document the Apple II?s binary 
executable file format, including the documentation for the MOS 6502 instruction set used 
in the Apple II line. 
 
In addition to complicated representation networks, the games in our case set have shown 
that games can be extremely reliant on what the OAIS reference model calls ?context 
information? to aid in their interpretation. The reference model defines context information 
? 100?
 as ?information that documents the relationships of the Content Information to its 
environment. This includes why the Content Information was created and how it relates to 
other Content Information Objects.?  It also provides examples of context information for a 
software package: a help file for the software, user guides, related software, and 
documentation on the programming language used. Taken together, these examples indicate 
that the purpose of context information for software is to aid the individual examining the 
software to understand its purpose, its operation, and the technical environment necessary to 
enable the software?s use.  
 
All of this information is clearly necessary for the game software we examined in our case 
set. While we think of games such as Star Raiders for the Atari 2600 as relatively primitive in 
comparison to today?s games, players had to understand how to activate and deactivate their 
ship?s shields, how to use the attack computer in targeting enemy ships, how to use the 
game?s galactic map to identify their own location, their starbase location, and the location of 
enemy ships, how to assess their ship?s energy level (and how to recharge it), and how to 
determine the amount of damage their ship had taken, in addition to basics such as 
maneuvering and shooting. The game also tracks skill levels of the game players over time as 
they advance in missions. Without the game manual, it is extremely difficult to interpret the 
screen displays of the game and understand exactly how game play is supposed to proceed. 
And a game like Star Raiders, designed for an antiquated console platform, nicely 
demonstrates the necessity of maintaining complete technical documentation on the 
environment in which it was designed to run if we are to offer continuing access to the 
game. Without the documentation necessary to produce an emulation of the Atari 2600 
Console, the Star Raiders game as distributed is useless. 
 
Games, however, may also require context information beyond the forms documented 
within the OAIS reference model document. Games are not just technical phenomena, they 
are also social phenomena, and the explanation of ?why the content information was 
created? can be extraordinarily complex in the case of something like the International 
Spaceflight Museum in Second Life. There are entire books devoted to the origins of the Atari 
platform and games from our case set (for examples, see Montfort [2009], Montfort [2003], 
Barton [2008], Boellstorff [2008]). If context information is needed to understand why 
content information is created, then in the case of video games and interactive fiction 
context information is going to be far more extensive than in the case of a data file from a 
space mission, and will also involve more extensive links between the digital content 
information to be preserved and the standard bibliographic universe of published books and 
journal literature with which librarians have traditionally been concerned. 
? 101?
  
 
Spacewar! running on a PDP-1. Image courtesy of Kenneth Lu. 
 
Another obvious, but critical, aspect of games relating to context information is that they are 
highly interactive. And because they are interactive, the ?why? of their creation is not 
necessarily something established by the game designers, but is constructed on an on-going 
basis by the games? users. This is particularly obvious in the case of massively multiplayer 
online systems such as World of Warcraft or Second Life. The nature of these virtual worlds is 
constituted in significant part through their use. Without documentation of how players 
actually engage with these resources, individuals in the future studying these materials will be 
left with an incomplete answer as to the ?why? of their creation. But this is equally true of 
single-user games. The oldest game in our case set, Spacewar!, was developed in part as a 
mechanism to demonstrate the capabilities of the PDP-1 computer on which it was 
originally implemented, including the capabilities of the vector graphics display used by the 
computer (see figure above). Vector graphics display terminals, however, are almost unheard 
of today outside of a few niche applications. Video or photographic documentation of the 
game in use on its original platform is also needed if people studying the game Spacewar! are 
to gain a complete appreciation of the ?why? of its creation.  
 
Analysis of Packaging Requirements 
 
As the above discussion (as well as the previous discussions on FRBR) should make clear, 
we have found that, to the extent preservation of computer games and interactive fiction is a 
metadata problem, it is primarily a matter of structural metadata. More completely, the 
problem of preserving computer games and interactive fiction is primarily an issue of 
structural metadata and collection management, insuring that you have the complete set of 
? 102?
 representation information and context information necessary to render your content 
information both accessible and apprehensible, and that all of the necessary relationships 
between content, representation, and context information are appropriately recorded. 
Possessing a technical metadata record stating that the file named ?star_raiders.bin? contains 
a binary executable intended to run on a MOS 6502 processor is of limited value for 
preservation when compared to having copies of the MCS 6500 Family Programming 
Manual and the MOS 6500 Microprocessor data sheets necessary to interpret the contents of 
the file. In fact, if we possess structural metadata linking the ?star_raiders.bin? file to those 
documents, and asserting that they provide representation information for it, recording a file 
format in a technical metadata record is actually somewhat redundant. 
 
Both the relationships described in the Functional Requirements for Bibliographic Records Final 
Report (1997) and the Reference Model for an Open Archival Information System (2002) are critical to 
the successful preservation of video games and interactive fiction. Any packaging mechanism 
for preservation of these materials must therefore support expression of these relationships 
between different entities. Moreover, as the relationships that exist in FRBR are specific to 
particular classes of entities defined in the FRBR report, a packaging mechanism for digital 
games must allow support, either implicitly or explicitly, the identification of entities 
described within the package as belonging to a particular FRBR entity class. 
 
Beyond these requirements are ones common to any digital preservation effort. In addition 
to representation information and context information, a packaging format needs to be able 
to provide a wrapper for fixity, reference, and provenance information. The digital file 
format employed for any packaging file for metadata and content should itself demonstrate 
the sustainability factors identified in the Library of Congress?s Sustainability of Digital Formats 
(2007) website. 
 
Finally, while packaging games for preservation is primarily a structural metadata problem, it 
is clear from both our research (and from the OAIS reference model itself) that archival 
information packages for video games and interactive fiction will need to make reference to 
non-digital material (print documentation for software, books and journal literature 
containing context information, etc.). Such references need to be capable of providing 
sufficiently detailed descriptions of resources to allow for reliable identification of specific 
editions of print material.  
 
Metadata & Packaging Recommendations 
 
As noted above, packaging of video games and interactive fiction for preservation is in 
significant part a matter of making explicit the different relationships that exist between 
various resources. There are two common web technologies in use today that can both link 
resources and specify the exact nature of the link between them, and that also provide a 
wrapper format that we can consider as highly sustainable if evaluated using the Library of 
Congress?s sustainability criteria. The first is the XML Linking (XLink) (World Wide Web 
Consortium, 2001) standard, and the second is the RDF/XML syntax specification (World 
Wide Web Consortium, 2004). While these two standards have some significant differences, 
for our purposes, the similarities are more important. Both provide a mechanism to specify a 
relationship between two resources. Both provide a mechanism to record a URI that 
indicates the specific type of relationship between two resources (the predicate in an RDF 
? 103?
 triple in RDF/XML, the xlink:arcrole attribute in XLink). Both are capable of asserting 
relationships between resources when: 
 
? both resources are contained within the same document as the link itself; 
? one resource is contained within the document containing the link and the other is 
external; or 
? both resources are external to the document containing the link. 
 
The METS standard employs XLink, while OAI-ORE is expressible within the RDF/XML 
format; both of these provide the necessary linking mechanisms and do so using common 
standards. While MPEG-21 DIDL does not explicitly support XLink, its schema allows 
widespread use of attributes from non-MPEG namespaces (including XLink) and so is 
capable of encoding the necessary linking information. The XFDU specification from the 
CCSDS does not, unfortunately, currently conform to either XLink or to RDF, and so we 
cannot recommend it as a packaging standard at this time. 
 
We believe that a shared, formalized ontology defining the various FRBR and OAIS 
reference model relationships is the preferred mechanism for enabling the specification of 
relationships in digital preservation packaging standards. By providing standardized URIs for 
each of the relationships, ontology provides both the means of indicating the specific 
relationships between resources and provides a shared language for recording information 
on relationships between preserved resources that will help promote interoperability 
between institutions. For the Preserving Virtual Worlds project, we have created an OWL 
ontology (see Appendix C) that defines the FRBR Group 1, 2 and 3 entities as classes, and 
defines the various relationships mentioned in both FRBR and the OAIS reference model as 
properties between these FRBR-based classes. 
 
Having this ontology permits the expression of both FRBR and OAIS relationships between 
resources in packaging formats employing RDF/XML or that provide appropriate support 
for XLink extended links. In OAI-ORE, for example, defining a particular aggregation as a 
FRBR work and associating the work with a piece of context information can be 
accomplished relatively easily: 
 
<rdf:Description rdf:about="http://people.lis.illinois.edu/~jmcdonou/PVW/aggregation/Mysteryhouse/Work.1"> 
       <rdf:type rdf:resource="http://people.lis.illinois.edu/~jmcdonou/PVW.owl#Work" /> 
       <pvw:has_context_information rdf:resource="http://www.ifarchive.org/if-archive/solutions/mystery-house.txt" /> 
       <ore:aggregates rdf:resource="http://www.ifarchive.org/if-archive/solutions/mystery-house.txt" /> 
</rdf:Description> 
 
A similar description is possible in METS: 
 
<mets xmlns="http://www.loc.gov/METS/" 
    xmlns:xlink="http://www.w3.org/1999/xlink" 
    xmlns:pvw=http://people.lis.illinois.edu/~jmcdonou/PVW.owl?> 
      <fileSec> 
             <fileGrp USE="Data"> 
                 <file ID="F1"> 
                     <FLocat LOCTYPE="URL"  xlink:href="http://www.ifarchive.org/if-archive/solutions/mystery-house.txt"/> 
                 </file> 
           </fileGrp> 
     </fileSec> 
     <structMap> 
           <div xlink:label="W1" LABEL="Mystery House" TYPE="pvw:Work?></div> 
? 104?
            <div xlink:label="W2" LABEL="Mystery House walk through" TYPE="pvw:Work?> 
              <div xlink:label="E2" TYPE="pvw:Expression"> 
                         <div xlink:label="M2" TYPE="pvw:Manifestation"> 
                             <div xlink:label="I2" TYPE="pvw:Item"> 
                                  <fptr FILEID="F1"></fptr> 
                             </div> 
                         </div> 
                   </div> 
           </div> 
     </structMap> 
     <structLink> 
           <smLink xlink:from="W1" xlink:arcrole="pvw:has_context_information" xlink:to="W2"/> 
     </structLink> 
</mets> 
 
While we believe that an approach to game packaging relying on standardized ontologies for 
the description of relationship types is the best means for encoding the necessary linking 
information between resources in a preservation environment in a way that supports 
interoperability, the PVW ontology should be seen as the start of a discussion on 
standardized ontologies for this purpose and not the end. We do believe that an ontology or 
ontologies that describe both the FRBR and OAIS models and that integrate the two are an 
essential component for packaging of computer games and other software. But if ontologies 
are to become a component of the preservation community?s standard toolkit for packaging, 
development of those ontologies needs to be done in a way that involves the larger 
community as part of process that would feed into an official standards effort. 
  
With regards to descriptive metadata employed within archival information packaging, we 
have already mentioned the need for descriptions that are sufficiently detailed to enable 
identification of specific editions of resources. This is particularly true if we assume a 
distributed, collaborative preservation environment such as envisioned by the Library of 
Congress (2002) in their Plan for the National Digital Information Infrastructure and Preservation 
Program. In such an environment, it is conceivable that institutions may rely on copies of 
representation information and context information being held by other, remote institutions. 
In such a case, being able to describe a work in sufficient detail in order to enable its 
identification outside of your own institutional setting is of vital importance. 
 
Dublin Core does not provide this level of detail, and we would recommend against its use 
in a preservation setting. Description standards such as MARC/XML, MODS, CDWA and 
VRA Core that enable identification of specific editions of resources should be used instead. 
? 105?
 9. Steps for the Future 
 
Library of Congress National Collecting Plan 
 
The Library of Congress has based its collection policy statements on three fundamental 
principles: 
 
? The Library should possess all books and other library materials necessary to the 
Congress and the various officers of the Federal Government to perform their 
duties; 
? The Library should possess all books and other materials (whether in original form 
or copy) that record the life and achievement of the American people; and 
? The Library should possess in some useful form, the records of other societies, past 
and present, and should accumulate, in original or in copy, full and representative 
collections of the written records of those societies and peoples whose experience is 
of most immediate concern to the people of the United States. 
 
The Library has a number of separate collection policy statements and overviews for various 
specific topic fields and in some cases for particular formats. The collection overviews 
covering special materials and formats include one on computer files 
(http://www.loc.gov/acq/devpol/colloverviews/computer.pdf), which discusses the 
computer game holdings of the Motion Picture, Broadcast and Recorded Sound Division at 
the Culpepper facility. The Collection Policy Statement for Moving Image Materials includes 
video games within its scope, noting that: 
Video games have become an established, popular medium of moving image 
entertainment that demand inclusion in the collections of MBRS. The 
Division is developing new approaches for the more systematic acquisition 
of video games, including playback consoles and platforms, the multiplicity 
of formats and their equipment needs, and the technical challenges in 
preserving the digital source files. The collection will encompass a wide range 
of examples of video game culture, to allow historians decades hence to fully 
understand this as a popular phenomenon, and not have simply a few games 
that seemed significant at the moment of release. 
 
The Collection Policy Statement also states that the collecting level for video games should 
be at the research level, and that the Motion Picture, Broadcast and Recorded Sound 
Division ?both selects via copyright and purchases video games, their associated hardware, 
and magazines about them that reflect the breadth and depth of gaming culture.? 
 
We strongly endorse the maintenance of a research-level collection of published video games 
by the Library of Congress. Just as important, in addition to documenting the output of the 
game industry, the Library of Congress, in collaboration with other cultural institutions, 
should make a sustained effort to ensure that materials are collected that document game 
cultures and communities built around digital games within the United States, as well as 
international cultures and communities that have influenced those in the United States. 
Based on our research, we would like to recommend some steps that we believe will assist 
? 106?
 the Library in maintaining a research-level collection of video games, as well as suggesting 
some guidelines that might be useful in the selection process for materials relating to video 
games. 
 
Collection Policy Recommendations 
 
It is perhaps the least surprising of our findings that computer games and interactive fiction 
are among the most fragile of digital works, due to their extensive dependency and 
interconnection with computer operating systems and particular hardware configurations. 
We can safely assume that a PNG image will display equally well on a Mac or a PC, and that 
an update to a computer?s operating system is unlikely to mean that the image can no longer 
be displayed. Neither assumption is safe for software. Given this fragility, all software, 
including games, requires an extremely proactive approach to preservation, and this has 
implications for collection policy. 
 
One of the first implications that should be considered is the necessity of collecting materials 
other than the games themselves. Both related representation information and context 
information must be acquired and preserved as part of the Library?s collections if games 
added to the collection are to survive. We address issues around representation information 
below. Context information is important in two respects. First, it provides understanding of 
the relationship of specific content such as a digital game to its technical environment. For 
example, documents such as game manuals, developer-created technical support websites, 
and contributions by players to discussion forums provide information about hardware and 
software dependencies that must be understood to install and provide access to game 
content or instructions on how to operate a game or tools based on it (as in the case of 
replay or machinima production). Second, information about the context of a game often 
serves equally well as archival documentation of the technical, player, fan, or creative 
communities around a game, or about historical events and activities that took place in game 
or virtual worlds; we discuss historical collections at greater length in section 4 of this report. 
We note here that contextual information for games and virtual worlds contributes double 
duty by adding to historical as well as preservation documentation. 
 
With respect to context information generally, our research has shown that preservation of 
computer games requires an integrative approach to collection development that is not 
common in libraries, archives, and museums to date. Context information for games can be 
found in the form of traditional published print material in many forms. The range of 
published formats for game-related information extends from scholarly books and journal 
literature such as Montfort?s & Bogost?s (2009) Racing the Beam or Dennis Jerz?s (2007) article 
on the game Adventure, through trade and web-based publications such as game guides and 
magazines and the descriptions of play experiences characteristic of the ?new game 
journalism? (Jim Rossignol?s (2008) This Gaming Life, or Tom Chick?s columns and essays) to 
ancillary publications and ephemera such as printed game manuals or on-line player reviews. 
Digital games have been at the heart of scholarly discussion of ?trans-media? content, so it 
should come as no surprise that non-print media also provide contextual information about 
games. We can start with the growing number of game-based movies such as the DOOM, 
TombRaider, Final Fantasy, Resident Evil, BloodRayne, and Mortal Kombat, game-based machinima 
and replays, and documentaries on games and gaming culture like King of Kong: A Fistful of 
Quarters, Chasing Ghosts: Beyond the Arcade, Frag, and Second Skin. Even audio materials such as 
? 107?
 game soundtracks and popular music contribute to our understanding of game content. The 
history of video game music has resulted in significant collections of game music, whether in 
the form of soundtracks provided with games or produced by fans, or original live 
performances. Game composers such as George Sanger (The Fat Man, whose archive is 
preserved at the University of Texas) and Nobuo Uematsu (the main composer for the Final 
Fantasy series) have established game music as a genre. Well-known musicians and bands 
from several styles of music have contributed to the production of specific soundtracks 
(such as Nine Inch Nails for id Software?s Quake), composed and performed specific songs 
for games ranging from EA?s Fight Night to music games such as Dance, Dance, Revolution and 
Guitar Hero, or licensed songs used within video games (such as the large selection of music 
available via car radio within the game Grand Theft Auto IV, from classics such as Duke 
Ellington?s ?Take the ?A? Train? to contemporary music by artists including Ghostface 
Killah, Sepultura, and Seryoga). Focusing exclusively on materials that might fall within 
traditional publication chains, we can see that context information that we might collect 
crosses divisional boundaries within the Library of Congress, such as between book- and 
media-oriented divisions. Significant coordination on collection policies will be necessary if a 
comprehensive collection supporting preservation efforts is to be achieved. 
 
Context information for a new digital medium such as games and virtual worlds will often, if 
not usually, be found outside traditional publication channels. We suggest that the Library of 
Congress consider web archiving efforts focused on game and game community sites similar 
to those which can be found in the Archive-It collections we created as part of the 
Preserving Virtual Worlds project (preserved both at the Internet Archive and at Stanford). 
Web archiving efforts at the Library of Congress might focus on major sites devoted to 
gaming and the gaming industry as a whole, or industry-wide issues and movements, 
whether these sites are created by developers, publishers, trade groups, players, or advocacy 
groups; specific topical approaches might better be undertaken by collaborating research 
libraries and organizations deeply engaged with specific games, genres, or research directions 
(such as the emphasis on game world ?cartography? in our project). Examples of significant 
categories for websites of industry-wide significance might include: 
  
? Portals for information about game development or game publication and history. 
Examples: Gamasutra (http://www.gamasutra.com), IGN (http://www.ign.com) 
and Moby Games (http://www.mobygames.com) 
? Websites of game industry associations. Examples: the International Game 
Developers Association (IGDA, http://www.igda.org/) and the Entertainment 
Software Association (http://www.theesa.com/). It is worth noting here that the 
IGDA includes a Game Preservation SIG with its own website 
(http://www.igda.org/preservation); this SIG is chaired by Henry Lowood from our 
group and was a productive outreach channel for the Preserving Virtual Worlds 
project. 
? Websites devoted to retro-gaming and related attention to historical games and game 
collecting, including fan-created websites such as the Atari Museum 
(http://www.atarimuseum.com/mainmenu/mainmenu.html) and Retrogames 
(http://www.retrogames.com/). 
? Various sites devoted to development or support of emulation platforms useful for 
playing historical games and games from obsolete platforms, such as Stella, and Atari 
2600 emulator, and the MAME emulation platform. Examples: the Stella open-
? 108?
 source project site (http://stella.sourceforge.net/), the official site of the MAME 
development team (http://mamedev.org/) and the Emulator Zone 
(http://emulator-zone.com). 
? Fansites for games, many of which offer not only fan forums and fan-produced 
content, but also collections of game documentation, such as screenshots, walk-
 throughs, FAQs, reviews, replays, and videos produced from screen captures and 
replays. Examples: Koinup (http://www.koinup.com/), SL Universe 
(http://www.sluniverse.com/php/) and WCR for Warcraft III replays 
(http://www.wcreplays.com/). Blizzard Entertainment offers an excellent collection 
of such sites on its World of Warcraft fansites page, 
(http://www.worldofwarcraft.com/community/fansites.html). 
? Major sites devoted to collecting and distributing machinima provide extensive 
documentation both of player creativity and of the exploration and content of virtual 
and game worlds. Examples: Machinima.com (http://www.machinima.com), the 
Machinima Archive (curated by Henry Lowood and a PVW site for archiving 
content; http://www.machinima.org), and sites devoted to machinima from specific 
games or virtual worlds, e.g., Halo movies (http://halomovies.org) or Warcraft 
movies (http://www.warcraftmovies.com/). 
? Websites devoted to end-user game modifications. Examples: ModDB primarily for 
console games (http://wwww.moddb.com) or FilePlanet?s mod pages primarily for 
PC games (http://www.filplanet.com/mods) and historical FilePlanet sites devoted 
to specific games, such as Planet DOOM (http://planetdoom.gamespy.com/) 
? Websites devoted to alternative games and ?art games.?  Examples: 56KModern 
(http://www.56kmodern.com/), the Independent Game Festival?s Indie Games 
(http://www.indiegames.com/), the IndieCade festival website 
(http://www.indiecade.com/), or Gamescenes: Art in the Age of Videogames 
(http://www.gamescenes.org/). 
? Websites devoted to ?serious games? and educational uses of games, including 
military training. Examples: The Serious Games Initiative 
(http://www.seriousgames.org/), Games for Change 
(http://www.gamesforchange.org/), or Combatsim.com 
(http://www.combatsim.com/). 
 
Some of these sites, such as the FilePlanet family of websites, have barely weathered threats 
to their continued operation or are under duress in terms of their long-term survival. A 
detailed list of virtual world and massively-multiplayer game websites created as part of the 
PVW Project in 2008 (see Appendix H) includes many now-defunct websites, only a fraction 
of which we were able to preserve as part of our ongoing Archive-It crawl. 
 
In addition to electronic resources, media and web-based collections of documents, the 
Library of Congress might consider collecting in some categories of print materials that are 
not handled within traditional publication channels. These typically would be collected 
perhaps under the rubric of ?ephemera? in a Special Collections Department; we are 
proposing that these sorts of materials, along with other categories of near print or ?grey? 
publication play an important role in documenting digital media such as games. Examples 
include repair manuals and technical schematics for game platforms of many kinds, including 
arcade consoles; advertising materials for games; and production materials from gaming 
companies (design notebooks and sketches, development versions of game software and 
? 109?
 source code, storyboards, scripts, rule outlines, character descriptions, etc.). Clearly much of 
this latter material falls within the domain of archival and special collections as a specialized 
department or area of expertise within most libraries. It should be noted here that in libraries 
that are currently collecting game-related material (such as the University of Texas and 
Stanford University), substantial archival holdings are held in such departments, along with 
published books in circulating libraries and audio and video in media collections. However, 
we believe that increasingly preservation of digital media of all forms will require eliding the 
traditional distinctions between library collections and archival collections if it is to be 
successful. If the Library does not choose to collect such archival material on its own, 
coordination between its efforts and archival collections at other research institutions 
committed to documenting the history of gaming and interactive fiction (including Stanford 
University, the Strong Museum, the University of Maryland and the University of Texas) 
should form part of the Library?s efforts under the National Digital Stewardship Alliance. 
 
With regards to games themselves, obviously the Library will collect a large number of video 
games through mandatory deposit of published software with the Copyright Office. 
However, we would caution that, given that registration of copyright fulfills mandatory 
deposit requirements, and that registration of copyright for software programs does not 
require complete deposit of the software, it is possible for significant games to be registered 
with the Library but not actually reside within it. Considering our case set, for example, 
Mystery House from Sierra On-Line, Inc., is registered with the Copyright Office, but the 
Library does not actually hold a copy of the game. A videotape and textual description were 
submitted in lieu of the game itself. Given this reality of the copyright registration and 
mandatory deposit system, the Library will need to survey the games actually present within 
its collections and determine whether there are additional games which it might wish to add. 
In addition, the Library may wish to determine whether other classes of supporting 
materials, such as the above-mentioned video and description of Mystery House, should be 
identified as relevant and important contextual documentation for game preservation and 
history. These considerations raise questions about the criteria that the Library should use in 
selecting games for inclusion in their collections, and for evaluating materials already present 
for reasons such as copyright registration. 
 
In considering both the games in our case set and in the larger worlds of historical and 
contemporary games considered as a whole, there are a variety of factors that we believe 
should inform decisions regarding collecting of games: 
 
? Popularity/sales/distribution ? popularity, as measured by game sales, should be 
considered. A game such as Call of Duty: Modern Warfare 2, which sold 4.7 million 
copies on its release day, clearly represents an important piece of popular culture. 
However, it should also be noted that lists of top-selling titles require careful 
interpretation; for example, they often are crowded with a limited number of titles, 
due to dominant series (such as Madden Football), sequels, and titles that appear on 
multiple platforms. Games with a larger number of release platforms and ports are 
likely to be among the more popular game titles, to the extent that popular industry 
?top-10? lists and the like may only provide a narrow view of ?hit? titles. A better 
source of information on popularity and success might be review-based ranking such 
as Game Rankings (http://www.gamerankings.com/) or Metacritic?s game site 
? 110?
 (http://www.metacritic.com/games/). Stanford, for example, uses the latter, in its 
game approval plan. 
? Novelty (technical/artistic) ? certain games represent significant advances of the 
state of the art in gaming, either due to technical or artistic innovations. A game such 
as DOOM, which represented a vast leap in the use of immersive 3D graphics, or 
Myst, which influenced the aesthetics of many later computer games, should be 
considered for inclusion. More recently, specific movements oriented towards 
independent (?indy?) game development, art games, or news games have emphasized 
the use of games as a platform for creative expression; websites tracking these areas 
will provide guidance concerning game titles and designers who are influencing these 
trends in game development. 
? Intertextuality ? certain games may be of value in understanding and interpreting 
other games. A game such as David Smith?s The Colony from 1987, which utilized 
true 3D for a game involving an off-world colony overrun by bizarre creatures as a 
result of scientific experiments in teleportation, clearly helps contextualize DOOM, 
which was released six years later (and in fact some of his techniques, such as using 
ray casting to determine the visibility of objects within the 3D space, were later used 
by id Software). 
? Impact on the Industry ? This criterion is closely related to popularity and cultural 
impact. However, it also calls for collecting games that had a significant impact not 
just in terms of reception, but also in terms of such matters as marketing, packaging, 
and distribution. For example, it is important to ensure that a representative 
collection of games includes examples that introduced specific media (e.g., CD-
 ROM) or packaging formats (e.g., Electronic Arts? ?album cover? packaging). These 
impacts extend to the introduction of new forms of distribution and game genres, 
which may also lead to special difficulties in access to games that should be collected. 
For example, Valve?s Steam has had enormous significance as a digital distribution 
and rights management platform, but the specific nature of its technology will lead to 
difficulties in collecting games distributed via Steam. Likewise, web-based ?social 
network games? such as Farmville have had an enormous impact on casual game 
design and easily satisfy criteria such as popularity and business impact, but the task 
of ?collecting? such a game is made difficult by its close dependence on social 
network sites such as Facebook. 
? Cultural Impact ? some games have a significant cultural impact reflected in their 
influence on other forms of media or in their ability to spark public discussion and 
debate. Id Software?s DOOM has achieved significance by both of these markers. 
References to the game have appeared in a variety of other media (including episodes 
of The Simpsons, Friends, and Family Guy, as well as in a song by the Smashing 
Pumpkins). At the same time, the subject matter of DOOM (slaughtering demons in 
outer space), its moody graphics and audio, and the vocabulary (?shooters,? ?death 
match?) increased public attention to the levels of violence depicted in computer 
games. The game was released just as congressional hearings convened by Senators 
Herb Kohl and Joseph Lieberman were getting underway to examine media violence 
and its influence on children, and thus its release played a role in the national debate 
on these issues. In such cases, games become involved in public policy, and the 
Library should make a special effort to ensure that the games cited in debates such as 
those around DOOM are included within their collections. 
? 111?
 ? Creator Prestige ? A game may also be considered important enough to collect due 
to the identity of its creator. Mindwheel probably would not have received the 
attention it has were it not authored by U.S. Poet Laureate Robert Pinsky. Games by 
well-known designers such as John Romero, Will Wright, or Hironobu Sakaguchi 
may not necessarily always be the best examples of games, but they may still have a 
significant impact on the game industry. It should be noted that this is a criterion 
particularly for archival collecting; the collections at Stanford (Steve Meretzky, Hal 
Barwood) and Texas (Warren Spector) have already begun to collect the papers of 
noted game designers. Some games are significant due to the prestige of their 
developers as inventors or innovators, some due to the importance of computer 
technology and software innovation for this medium. John Carmack?s early games 
(before DOOM) are significant for this reason, and the Ralph Baer papers at the 
Smithsonian Institution provide an example of the impact of this criterion on 
archival collecting. 
? End-user Appropriation ? Some games offer features that make them particularly 
suited to end-user appropriation in various ways. This is a particularly important 
criterion for considering games not just as authored content per se, but as a creative 
platform for expression by fans, players, and others. One aspect of this creativity is 
the alteration of games or their use to create wholly new games. Games like DOOM 
and Quake were among the first to provide the means for users to create new maps 
for game play or make new assets such as creature ?skins? and monsters. Such games 
have since spawned a cottage industry of ?mods,? modified (in some cases heavily 
modified) versions of game. Users have also appropriated various multi-user online 
games (including Halo, World of Warcraft, Second Life and many others) for the 
production of machinima, videos created using game technology, assets, or the 
virtual world of a game as a staging and production platform. Finally, many players 
have created reputations as ?cyber-athletes? by virtue of their creativity in 
competitive game-play, to the extent that today professional and near-professional e-
 sports have attained some significance both nationally and internationally. All of 
these areas?mods, machinima, e-sports?represent activities through which players 
demonstrate their creativity. Games which have lead to the creation of significant 
user-produced content should also be given some priority in collecting activity, as 
should the mods and machinima produced using them, or replay movies that capture 
important moments in the history of competitive e-sports. 
? Impact on Game Design and Technology ? Particularly novel games may have a 
major impact on the development of the gaming industry. The game Crysis, for 
example, established a new benchmark for the whole industry in terms of 
photorealism in the display of the game world and in the complexity of the physics 
engine used by the game; the game was so demanding of computer hardware, in fact, 
that ?But can it run Crysis?? became a common trope when discussing the 
capabilities of a new computer platform. The original Pong game was obviously not 
noted for the sophistication of its game play, but it led to the development of the 
entire console game industry. 
 
In addition to the above issues, there are two other factors that the Library should consider 
in developing its game collection. Mandatory deposit for copyright will obviously help 
ensure that the Library has a strong collection of games produced and marketed within the 
United States. However, the gaming industry (and gaming community) is international, and a 
? 112?
 research collection such as the Library?s should include significant games from outside the 
United States. The factors listed above should be considered when selecting games from 
outside the United States, but the Library might also wish to consider making special efforts 
to collect representatives of genres of games that are not typical of the U.S. game market 
(the genres of visual novels and dating sims from Japan, for example), that have achieved 
prominence within their own countries (such as the various MMORPGs operated by Shanda 
in China), or that have had significant influence on games or the game industry in the United 
States. Likewise, the emphasis on collecting games that document creative play might point 
to documentation of competitive play cultures in countries such as Korea and regions such 
as Central and Northern Europe. Countries already possessing major video game industries 
include Japan, Korea, United Kingdom, Canada, Australia, Germany, Sweden, and France, 
and there are significant industries developing in China, Taiwan, Russia, the Ukraine, 
Finland, Denmark, and the Netherlands. 
 
On the other hand, game technology and the game industry have also been significantly 
shaped by the regional concentration of U.S. ?high-technology? industry, particularly in 
computing and web-related sectors. In recent years, the game industry has also been a driver 
in sectors such as the graphics and audio hardware industries and has influenced mainstream 
entertainment such as in the movie industry. Thus, it is no surprise that institutions currently 
active in game collecting activities tend to be located in or near prominent techno-regions or 
have themselves played prominent roles in producing relevant technologies (Stanford, Texas, 
Illinois). These comments regarding the collecting implications of both the global and the 
regional nature of the digital game industry underscore again our conviction that one 
collecting institution is unlikely to be able to do it all and lead us to conclude this section by 
noting again the importance of collaborative collecting efforts. 
 
Intellectual Property & Digital Preservation: Widening the Discussion 
 
There is a need for memory institutions interested in establishing video game repositories to 
support policy positions and offer services that are better aligned with the preservation and 
use practices of the game community, not only because of the potential for integrating the 
members of this community into the larger preservation network, but also because their 
attitudes and values may well influence those of professional archivists in years to come. 
With their deep curatorial investment in games, players have adopted a versatile set of 
approaches for collecting, managing, and providing long-term access to these cultural 
artifacts. While bitstream preservation and emulation are an essential part of the overall 
picture, so too are re-releases, remakes, demakes, ports, mods, and ROM hacks.3  As a result, 
game archives in the wild often reflect notions of authenticity that are different from those 
????????????????????????????????????????????????????????
 3??Remake? and ?demake? are game design terms coined by Ian Bogost.  As defined by Bogost, 
remakes ?are recreations of earlier works, irrespective of the hardware platform of original creation 
or recreation.? Conversely, demakes are ?retro-inspired reimaginings of modern games, as if they had 
been created on earlier hardware.  Demakes are not necessarily created to run on older machines, but 
their design and behavior are constrained by the real or perceived constraints of vintage systems.? 
See Bogost?s spring 2010 syllabus for ?Atari Hacks, Remakes and Demakes: Special Topics in Game 
Design and Analysis,? available at 
http://www.bogost.com/teaching/atari_hacks_remakes_and_demake.shtml. 
? 113?
 of memory institutions.4 While the dominant orientation of archivists is toward the 
stabilizing of cultural records, gamers can tolerate?indeed embrace?greater variability in 
the objects with which they interact. The Mario Brothers franchise, for example, has 
persevered largely because of the consumer appetite for new and altered game levels, power-
 ups, graphics, characters, and items. As another example, consider the case of Mystery House, 
the first graphical work of interactive fiction ever created. Released into the public domain in 
1987, Mystery House was recently reimplemented in the Inform programming language by the 
Mystery House Taken Over Occupation (MHTO) Force, comprised of Nick Montfort, Dan 
Shiovitz, and Emily Short (2004). The project team also commissioned ten contemporary 
digital artists to mod the game using a specially designed kit for the purpose. The end result 
is that MHTO oscillates between a preservation project and a remix project. As Jon Ippolito 
(2008, p. 106) has remarked, new media art ?can survive only by multiplying and mutating ... 
fixity is death.? Consequently, there is an argument to be made in favor of preservation 
strategies that involve reprogramming, reimplementation, and recreation. The ?softer? view 
of authenticity that underlies these strategies operates at what Seamus Ross (2010) would call 
a ?lower threshold of verisimilitude.? 
  
What policy initiatives and preservation services might be adopted in response to the needs, 
practices, and perspectives of players and player-archivists?  We propose the following: 
 
Digital Preservation Services: Comparative Methods and Stemmatics 
 
In addition to providing authenticated capture, ingest, hashing, and storage services for 
archival copies of games, digital repositories might also offer appropriate services for access 
copies of games in the wild. Because these copies are often modified rather than fixed 
representations?in line with the ?softer? canons of authenticity previously mentioned?
 repositories could provide users and player-archivists with the means to analyze, document, 
and measure their inter-relationships using similarity metrics and other approaches. A good 
example is Colossal Cave Adventure, created by Will Crowther in c. 1975, which has the 
distinction of being the first documented computer text-adventure game. Inspired by 
Kentucky?s Mammoth Cave system, the game inaugurated many of the conventions and 
player behaviors now associated with the genre, such as solving puzzles, collecting treasure, 
and interacting with the simulated fantasy world via short text commands (Jerz, 2007). 
Originally written in FORTRAN, the code has been repeatedly ported and expanded by 
hackers, fans, and programmers over the years, most influentially by Don Woods (c. 1977). 
The number of versions of Adventure is legion and continues to grow, with players mapping 
the evolution of the game using tree-like structures showing patterns of inheritance and 
variation (Dalenberg, 2004 and 2006). 
 
Applying the techniques of digital stemmatics, archivists could help users visualize and 
interpret these patterns in sophisticated ways. Developed in the 19th Century, stemmatics 
codified a set of methods for analyzing the filiation of literary manuscripts. Significantly, the 
tree structures representing these relationships have parallel importance in evolutionary 
biology and historical linguistics, where they are used to group genomes or languages into 
families; show how they relate to one another in genealogical terms; and reconstruct lost 
????????????????????????????????????????????????????????
 4??Archives in the wild? is a phrase coined by Jeremy John in the British Library?s Digital Lives 
Report. 
? 114?
 archetypes (Kraus, 2009). Speculating on the role of digital stemmatics (or phylogenetics, as 
the comparative method is called in biology) in the context of personal digital archives, 
Jeremy John of the British Library has postulated that ?future researchers will be able to 
create phylogenetic networks or trees from extant personal digital archives, and to determine 
the likely composition of ancestral personal archives and the ancestral state of the personal 
digital objects themselves? (John, 2010, p. 134). 
 
Stemmatic methods have already been applied to board games: Joseph Needham, a 
pioneering historian of East Asian science, technology, and culture, published a family tree 
of board games connecting divination, liubo, and chess through a long line of ancestry and 
descent (331); and biologist Alex Kraaijeveld has applied phylogenetics to variants of chess 
to help determine its place of origin. The methodology therefore shows great promise for 
the study of video and computer games in the wild, where variability rather than fixity of 
representation is often the norm.  
 
Two other tools cited recently by Jeremy John are also relevant in this context: 
 
? ccHost ? an open-source content management system developed under the auspices 
of Creative Commons that can be used to track and document how media content is 
used, reused, and transformed on the web. (ccMixter, the popular music site for 
remixing and sharing audio samples, is powered by ccHost.)5 
? Comparator ? a tool developed by Planets (Preservation and Long-term Access 
through Networked Services), a four-year project funded in part by the European 
Union. Comparator is designed to measure degrees of similarity between different 
versions of a digital object. 
 
Stemmatics might also be experimentally applied to the history of platform architectures.  
The system of hardware documentation developed by Walker Sampson for MITH?s vintage 
computers complements such an approach by modeling the collection ?through component 
pieces and parts? (Sampson, 2010). By decomposing each computer into discrete hardware 
components?microprocessors, disk drives, input devices, display technologies, and so 
forth?the documentation isolates possible units of variation. These units or parts can then 
be grouped together according to hereditary relationships (for example, partial pedigrees 
have already been constructed on Wikipedia and elsewhere for the MOS 6502 8-bit 
microprocessor, the ancestor of both the MOS 6510, which was used in the Commodore 64, 
and the W65C02S, which was used in the portable Apple II). Similar variations and 
derivations could be established for other processors and hardware components. Perhaps 
even more significantly, we could begin to model evolutionary degrees of similarity and 
difference not only between successive lines of personal computers or consoles developed 
by a single company (e.g., the Apple and Macintosh series), but also between entirely 
different families or brands of computers (e.g., the Atari ST, nicknamed ?Jackintosh? for its 
obvious mimicry of the original Macintosh. Both machines were also based on the Motorola 
68000 CPU). These findings would in turn have potential implications for computer 
restoration and console modding. 
????????????????????????????????????????????????????????
 5 Information about ccHost and ccMixter can be found online at 
http://wiki.creativecommons.org/CcHost and http://ccmixter.org/.  See also Jeremy John 56 and 
156n244. 
? 115?
  
Digital Preservation Services: Calculating Trust In Fan-Run Game Repositories 
 
Because game archives in the wild cannot usually be authenticated according to standard 
integrity checks, an alternative method for evaluating the authenticity of their holdings might 
involve the application of trust-based information. Jennifer Golbeck, for example, has 
demonstrated how the trust relationships expressed in web-based social networks can be 
calculated and used to develop end-user services, such as film recommendations and email 
filtering. Applying Golbeck?s insights, archivists could leverage the trust values in online 
game communities as the basis for judgments about the authority or utility of relevant user-
 run repositories, such as abandonware sites and game catalogs. Under this scenario, 
authenticity is a function of community trust in the content being provided. One 
consequence of this approach is that authenticity and mutability need not be considered 
mutually exclusive terms; on the contrary, fan-run game repositories that make provisions 
for transformational use of game assets?such as altering the appearance of avatars or 
inventory items?might in many instances increase trust ratings.  
 
IP and Public Policy: Reform of Contract Law 
 
While archivists and librarians have for the most part targeted copyright law for reform, 
gamers have influenced the civil code documents that govern virtual worlds, such as EULAs 
and TOS agreements. Because this class of documents is frequently updated to reflect the 
evolving nature of the relationships among different stakeholders (Grimes et al, 2008), 
gamers have arguably had greater success than archivists and librarians in advancing their 
goals over a relatively short period of time. We therefore recommend that archivists and librarians 
devote more effort to the reform of contract law and its associated document genres. To the extent that IP is 
becoming a matter of private policy rather than public policy, managed through licensing 
and contractual agreements as opposed to federal copyright law, then advocacy efforts 
increasingly need to be directed at the commercial industry rather than the copyright office 
or the legislative branch. In many respects?and perhaps somewhat counter-intuitively?the 
game industry is better positioned than Congress or the courts to respond in a progressive 
and timely fashion to emerging social, cultural, and technological forces that are radically 
restructuring the relationship between content creators and users. For example, within the 
last few years Microsoft and Blizzard have published Game Content Usage Rules that relax 
the exclusive authorial right of adaptation in order to permit players to generate and 
distribute machinima, video, and other derivative works of art (2004-2010). These rules in 
turn provide a nascent legal environment in which remix culture and Web 2.0 can flourish. 
That such an environment is the product of contract law rather than copyright law bears 
emphasis: the protracted processes through which copyright law has traditionally been 
emended increasingly provide inadequate grounds for meeting the challenges of our current 
cultural milieu, a milieu characterized by rapid technological upheavals that in turn redefine 
social attitudes toward concepts such as originality, copyright, and creativity. Case in point: 
the Section 108 Study Group, charged with re-examining the Copyright Act with an eye to 
updating it for the digital age, took nearly three years to reach agreement on a set of non-
 binding and comparatively tepid recommendations, which may or may not eventually result 
in actual policy change upon legislative review. By contrast, the example of Microsoft, in 
particular, demonstrates how contract law can help legitimate player innovation by rapidly 
legalizing and codifying it??rapidly? here being an admittedly relative term.  
? 116?
  
It is crucial to note, however, that the fluid state of contractual agreements is a double-edged 
sword: EULAs can just as easily be used to constrict or rescind rights as to expand or confer 
them. As Microsoft states in its content usage rules, the company may choose to revoke a 
license ?at any time and for any reason? (2004-2010). Although subject to some statutory 
constraints, it may act self-interestedly, magnanimously, or capriciously as it sees fit. A 
temporal content analysis would shed light on directional patterns of change: are user rights, 
once introduced into game licenses, really vulnerable to repeal, or do they tend to persist 
over time to the point that they become fully naturalized components of the game culture? 
 
Questions and caveats aside, it remains the case that civil code documents, such as EULAs, 
are more dynamic in nature than the U.S. Copyright Act.6   Between 2002 and 2007, for 
example, Second Life, the popular 3D virtual world developed by Linden Lab, published at 
least nineteen different versions of its EULA (Fitz, 2008). Just as revealingly, Microsoft 
published revisions to its Content Usage Rules less than two months after their initial release 
based on feedback from the machinima community (Hayes, 2008, p. 570). By contrast, 
Section 108 of the Copyright Act has been amended a mere four times since it was first 
published in 1976. The opportunities for citizen intervention into the machinery of contract 
law is therefore much greater than comparable opportunities for intervention into copyright 
law.7  
 
IP and Public Policy: Layered Licensing Agreements and Archival Endorsements 
 
As a way to address the impenetrability of many software license agreements, Jeremy John 
has recently proposed a layered approach to their implementation, citing Creative Commons 
deeds as a relevant model (56, 58). Under this example, the prohibitive length and jargon of 
many game licenses would be ameliorated by condensing them and restating their salient 
points in accessible language and machine-readable format. A set of icons would be 
prototyped as an additional layer encoding baseline user rights and exemptions, again 
following the CC precedent. The end result would be layered licenses, each layer targeting a 
different audience: lawyers and experts (original legalese), users (vernacular and visual 
versions), and machines (RDF version). Applying John?s suggestions to game EULAs, 
Content Usage Rules, and TOS agreements, we recommend that iconic and vernacular 
versions of these documents be developed and standardized to efficiently communicate the 
following information: 
 
? Whether or not the user has the right to reproduce game assets 
o The game assets created by the user 
o The game assets created by other users 
o The game assets created by the developers 
? Whether or not the user has the right to create derivative works, such as machinima 
? Whether or not the user has the right to transfer game assets (e.g., to a public 
repository) 
????????????????????????????????????????????????????????
 6?On the dynamism of civil code documents, such as EULAs, see Grimes et al., 2008. 
7?Such intervention is a function of the dynamism of the documents, and is by no means due to their 
being collaboratively authored by different communities of practice.  Machinimists and players 
influence the content of the documents indirectly rather than directly.?
? 117?
 ? Whether or not the user has the right to export his or her game assets (or those of 
others) in an open format 
? Whether or not an archival exemption clause exists that allows librarians and 
archivists to create preservation copies of game assets 
? Whether or not the license includes a ?competitive endorsement by official or public 
repositories? (John et al., 2010).  
 
The PVW team or LOC should consider using Appendix A: Virtual Worlds that Died 
During the Grant to pressure game companies to embrace use policies designed to secure 
the endorsement of libraries and archives; and to raise awareness among players about the 
importance of these policies for long-term access to and use of game assets. Additionally, the 
PVW team recommends convening a working group of librarians, archivists, and researchers 
to draft an archival exemption clause and conduct outreach events with game developers and 
players to promote its adoption. 
 
IP And Public Policy: Expanding User Rights In Adaptation 
 
Archivists and librarians have spent most of their political capital in lobbying for legal 
exceptions to the exclusive authorial right of reproduction, largely ignoring the fact that they 
and their users also have a tremendous stake in limiting the exclusive authorial right of 
adaptation. Gamers, by contrast, have significantly expanded their rights to produce derivative 
works based on the creative assets of game publishers (Hayes, 2008). We therefore recommend 
that archivists, librarians, and game scholars work to expand user rights in the area of adaptation, following 
the example of gamers. Section 108 of U.S. copyright law privileges librarians and archivists as a 
special class of users for whom reproduction is the single most valuable non-exclusive right 
that may be accorded to them for purposes of preserving the media objects in their care.8 
  Conversely, game content usage rules privilege players and the commercialization of 
primary content, which is potentially supported by the production of derivative works, such 
as machinima, that boost brand value. Granting the user the right to prepare derivative 
works is therefore fundamental to realizing what Lawrence Lessig (2008) calls a ?Read-
 Write? (RW) culture. Defined in opposition to a ?Read-Only? (RO) culture, which assumes 
a clear division between creators and consumers, a RW culture sees the two as deeply 
intertwined: RW creativity is one in which a society?s ?ordinary citizens? not only passively 
?read? their culture, but also actively transform it, producing content such as fan fiction, 
game mods, machinima, and audio remixes.  
 
Within the RW context, it is the right of adaptation, not reproduction, that achieves 
preeminence among the five established and inter-related pillars of the copyright code (the 
reproduction, adaptation, distribution, performance, or display of the original work). 
Moreover, the model of creativity that underwrites RW culture shades almost imperceptibly 
into an emergent model of preservation. Because players are already an integral part of the 
preservation system for video games, any adaptation rights they acquire as content creators 
also helps them in their capacity as content preservers (Kraus, Donahue & Winget, 2009). 
????????????????????????????????????????????????????????
 8?The word ?reproduce? and its cognates (e.g., ?reproducing?) appear no less than twenty-four times 
in Section 108 of the U.S. Copyright Act.  Conversely, the word ?derivative? never appears at all, nor 
does any related terminology (such as ?adaptation?).  Additionally, the executive summary report 
issued by the Section 108 Study Group makes no recommendations regarding the right of adaptation.?
? 118?
 This crossover advantage stems from the fact that the player as modder and the player as 
preservationist both produce a transformed digital object. In short, because citizen 
preservation methods that transmit culture via ?version streams? are becoming increasingly 
prevalent,9 it is essential that archivists, librarians, and curators advocate strongly for the 
right to prepare derivative works.  
 
Representation Information and Format Registries/Tools 
 
The Library has endeavored in creating its digital preservation systems to ensure that they 
comply with the Open Archival Information System (OAIS) Reference Model. This carries 
with it the obligation to ensure that any digital content acquired is matched by the necessary 
representation information to interpret the digital content. Our research into our case set 
indicates that the representation information necessary for preservation of software will fall 
into several major classes: 
 
1.  Source Code Representation Information 
a. Structure Information ? documentation of text standard (e.g., Unicode) 
employed by the source code documents 
b. Semantic Information 
i. Documentation of the programming language employed 
ii. Documentation of any OS libraries invoked by the code which may 
not be part of the standard libraries for a language 
iii. Documentation of any external APIs invoked by the code 
2. Binary Executable Representation Information 
a. Structure Information 
i. Documentation of the executable file format employed by the binary 
(e.g., COFF, ELF, Mach-O, etc.) 
b. Semantic Information 
i. Documentation of the instruction set employed by the processor the 
binary executable was created to run on (e.g., Intel 64 and IA-32 
Architecture reference manuals). 
ii. Documentation of the standard hardware architecture for the 
computing platform the executable was designed to run on, including 
information on memory, buses used to interconnect parts of the 
system, standard I/O devices and any special I/O devices required 
by the program. 
3. Game Data File Representation Information (images, databases, audio tracks, etc.) 
a. Structure Information 
b. Semantic Information 
4. Media Representation Information (disk image formats, ROM organization, etc.) 
a. Structure Information 
b. Semantic Information 
5. Recursive Representation Information (rep. information for all preceding forms of 
rep. information) 
????????????????????????????????????????????????????????
 9??Version stream? is a concept defined by Jon Ippolito, et al., in the context of The Pool, an online 
collaborative environment for designing, sharing, and disseminating variable media art: 
http://bit.ly/57hhE0.?
? 119?
 a. Structure Information 
b. Semantic Information 
 
While in many ways the representation information for software is similar to that for any 
other complex data object which may employ a variety of file formats, there are several 
aspects of representation information for computer games that we believe deserve 
highlighting and that suggest possible future steps that the Library might wish to consider. 
 
The first is that the representation information for computer software consists in significant 
part of standards documents or public specifications created by companies. Standards for 
character encoding and programming languages form the greatest part of representation 
information for source code versions of programs (augmented by documentation regarding 
device specific APIs that might be used in a program). For binary executables, the formats 
employed for executables on different platforms are well documented, as are the instruction 
sets for processors for particular machines; companies producing hardware typically have 
every incentive to provide developers this information in order to encourage them to create 
software for that platform. While representation information for source code and binary 
executable formats is typically readily available for modern platforms, it should be noted that 
this information can be 1) quite voluminous (the documentation for Intel 64 and IA-32 
architectures runs to several thousand pages); 2) difficult to obtain once the technologies 
supporting a particular game become antiquated; 3) unlikely to be found in typical library 
collections, particularly standards documents; and 4) highly repetitive for different pieces of 
software. 
 
Given these facts, and given that the OAIS Reference Model suggests that the end points of 
a representation network for a data object must be decipherable without use of computing 
equipment, there appear to be clear cost benefits to maintaining a separate collection of 
printed representation information with the Library of Congress. This would allow 
information packages that the Library of Congress creates for any games it holds to 
reference print copies of representation information for these items. Given that so much of 
this information is in the form of standards documents, there might be grounds for some 
collaboration with the National Institute of Standards and Technology on the creation of 
such a collection. 
 
Another crucial point about representation information for gaming materials that should be 
highlighted is that it may include more information about storage media than might be 
required for some other forms of data. Some of the games that we have investigated would 
require an emulation strategy for preservation due to the fact that the original source code 
was lost or unavailable, and in those instances, the binary executable versions of games that 
we were able to obtain were contained within a disk image file for the original game 
platform. Such instances can be seen as examples of the onion-model set forth in the 
PREMIS metadata dictionary, where one file may contain bitstreams, which may in turn 
contain further bitstreams, each of which may require separate representation information. 
For disk images, the representation information will include documentation on the original 
media format and use. 
 
As noted in the OAIS Reference Model, representation information can itself require 
representation information, leading to the creation of a representation network. We found 
? 120?
 that representation networks for games could quickly become quite large both in terms of 
the number of works and the byte count of the representation information files, and that the 
representation information for a game could easily outsize the game itself by several orders 
of magnitude. In the case of Star Raiders for the Atari 2600 platform, the game itself required 
only 8K of storage, while the representation information for the game consumed over 100 
MB. This provides yet another incentive to compile a central store of representation 
information that information packages for games (and probably other materials) would 
reference at the Library. 
 
Having the representation information, however, is not enough. One of the aspects of 
representation networks that has not been sufficiently commented upon in the case of 
standards documents is that they are stable (at least until a standards document is 
significantly revised), and that much of the recursively referenced representation information 
for a standards document will be other standards documents cited in the original standard. A 
database tracking these representation network links between documents would be of great 
benefit to the Library and others. At the moment, we do not believe that the data models 
being proposed for the Unified Digital Format Registry (UDFR) will support recording this 
type of information linking the pieces of documentation associated with a particular data 
format. The Library may wish to either work with the UDFR to ensure that support for this 
type of information is added into the UDFR data model, or consider creating its own 
database for representation network information. 
 
Digital Game Canon 
 
This outreach activity coincided with the initial development of the Preserving Virtual 
Worlds proposal. The creation of the Digital Game Canon was undertaken under the 
auspices of the Game Preservation Special Interest Group (SIG) of the International Game 
Developers Association (IGDA); Henry Lowood of the PVW team has chaired the Game 
Preservation SIG since 2006 and organized the original Digital Game Canon activity, 
namely, the announcement of the initial selection of 10 games for inclusion in the Digital 
Game Canon. 
 
The goals of the Digital Game Canon are two-fold: 1) recognition for the importance of 
digital game culture, including raising awareness of the impact and responsibilities of this 
importance for the game industry; and 2) establishment of a basis for decisions about the 
historical value of specific game titles that reflect a mix of academic, industry, and 
journalistic perspectives. 
 
The Canon provides a starting-point for the difficult task of preserving this history inspired 
by the role of that the U.S. National Film Registry has played for film culture and history. 
The scope of the Canon has thus far been international. Our argument: We could do worse 
than to start by making sure these games and archival material related to them are available 
to future developers, players, and scholars. 
 
At the 2007 Games Developers Conference, five panelists (Matteo Bittanti, Christopher 
Grant, Henry Lowood, Steve Meretzky, and Warren Spector) revealed and discussed their 
choices for the first 10 games on this list 
? 121?
 (https://www.cmpevents.com/GD07/a.asp?option=C&V=11&SessID=3885). These were 
the 10 game titles put on the Canon at this event: 
 
? Spacewar! (MIT, 1962) 
? Star Raiders (Atari, 1979) 
? Zork I: The Great Underground Empire (Infocom, 1980; PDP-11 version) 
? Tetris (Alexey Pajitnov, 1985) 
? Sim City (Maxis, 1989) 
? Super Mario Brothers 3 (Nintendo, 1990) 
? Civilization I/II (MicroProse, 1991-1996) 
? DOOM (id Software, 1993) 
? Sensible World of Soccer (Sensible, 1994) 
? Warcraft I/II/III (Blizzard, 1994-2003) 
 
Full audio and slides from the event can be downloaded from the IGDA Preservation SIG 
website (http://www.igda.org/preservation/files/dgc_gdc2007/).  
The Preservation SIG also maintains a web page for the Digital Game Canon 
(http://wiki.igda.org/Game_Preservation_SIG/Digital_Game_Canon/). The Gamasutra 
game design website has commissioned articles on games in the Canon, such as this essay on 
Spacewar: 
http://www.gamasutra.com/view/feature/1433/down_the_hyperspatial_tube_.php. The 
Canon project was originally conceived as an ongoing project, much like the National Film 
Registry. Plans have been discussed to revive the project in 2010 or 2011 as a continuing 
series of annual additions of 10 or so games at a suitable public venue, possibly as a joint 
venture with a website such as the Joystiq blog community. Just as the National Film 
Registry is linked to the work of the Library?s National Film Preservation Board, we propose 
that the Digital Game Canon be coordinated in a similar fashion by a board with ties to 
Library of Congress. If this were of interest to the Library, this would be a topic for further 
investigation with Henry Lowood and the IGDA Preservation SIG, as well with potential 
project partners in industry and the player community. It could also provide the basis for 
future preservation activities based on the work accomplished by the Preserving Virtual 
Worlds project. 
Reimagining Videogame Asset Management & Preservation (ReVAMP) 
Symposium 
 
The challenges of digital preservation require early intervention -- a requirement that has 
brought archivists, librarians, information technology professionals, and scientists from a 
range of fields (notably in the space and geospatial communities) together to find solutions. 
To date, video game designers have been relatively uninvolved in these collaborations.  
 
Through our work with the Preserving Virtual Worlds project, a survey of game developers, 
and discussions with game industry professionals, we have realized that there is a large gap 
of understanding and experience between people in the industry and those in cultural 
institutions, with scholars sitting somewhere in the middle ? perhaps able to act as bridging 
agents. As video games grow in cultural importance, the need to preserve them and the 
? 122?
 materials generated during their development becomes more evident. It is essential that these 
diverse communities understand each other?s goals and perspectives to ensure that future 
students, scholars, and game developers are able to access the rich history of video games? 
creation and use. The natural first step in creating such understanding is to bring 
representatives of each community into one room for discussion. 
 
To impress upon the video game industry the importance of preserving its own history, as 
well as for the purpose of educating its talent pool, we propose to start with a two-day 
Symposium styled after the 1960 Conference on Science Manuscripts (Woolf, 1962).10 The 
event will bring video game developers, information professionals, and game scholars 
together to discuss the challenges and value of instituting formal preservation programs 
through presentations and discussion. It is hoped that such a forum will lead to the 
development of a roadmap for the future, generate follow-up activities, encourage cross-
 community collaboration, and motivate the video game developers to take action for 
preservation. The planning details, agendas, and activity summaries could be used as a model 
to jump-start preservation and collaboration in other industries that may currently be 
struggling with (or ignoring) the same issues.  
  
The goal at the heart of our proposal is inspired by the history of science documentation 
efforts in the United States; we intend nothing less than to instigate an intensive series of 
meetings and jump-start serious efforts toward the adequate documentation and archival 
preservation of the many worlds involved in video games: design, technology, business, and 
culture. To do so, we propose to host a series of symposia styled after the Conference on 
Science manuscripts mentioned above. The conference could be scheduled around a major 
industry event, such as the Game Developers Conference (GDC), to ensure maximum 
attendance. Stanford would be an ideal venue due to its location in the heart of Silicon 
Valley, the presence of significant game-related holdings in its archives, and its proximity to 
the GDC site in San Francisco.  
  
We believe that a series of small symposia would be best suited towards influencing key 
members of the game industry to build support for future preservation work. To accomplish 
this goal, the first ReVAMP conference will bring together scholars, information 
professionals, and developers to present and discuss the benefits of preserving game 
development materials, including providing a general awareness of game history, reuse of 
game assets, training new developers, preserving culturally important records and artifacts, 
and building brand awareness. Additionally, we will strive to ease the understandable 
concerns of business managers and legal offices regarding respect for licenses and 
intellectual property rights. Attendance at the first ReVAMP conference will be by invitation 
only, and we expect to include no more than 30?40 individuals.  
   
Ultimately, we seek to partner with the game development industry in the development of 
one or more disciplinary centers with archival repositories to support all sectors of the video 
game world in much the same way that the Center for History of Physics (CHP) was 
established in 1965 by the American Institute of Physics to ensure adequate documentation 
????????????????????????????????????????????????????????
 10?See also, Joseph Anderson?s write up on the development of the Center for History of Physics at 
the American Institute of Physics, ?Difficult to Document: Physics in Government and Industry.? 
Available at http://www.bath.ac.uk/ncuacs/FP_Anderson.htm.?
? 123?
 and preservation of physics materials. Gaining the support of the creators themselves is a 
vital first step along this ambitious path, and a conference of this nature represents the best 
way to open discussion with game developers and get them thinking about preserving their 
own history. As Joseph Anderson, head of the CHP?s archival program has put it, ?All the 
stakeholders in archival records?the people who create them, the researchers who use, and 
the archivists who preserve them?should work together to decide what can and should be 
preserved and to develop appraisal guidelines.? The ReVAMP conference will represent our 
best effort to date to create such a consensus among parties concerned with the records of 
game development, technology, business, and culture. We therefore propose a follow-up 
symposium, somewhat smaller than the first, for the organization of an appropriate effort to 
create a disciplinary history center for the preservation of these records. 
 
A Research Agenda for Preservation of Video Games & Interactive 
Fiction 
 
While the Preserving Virtual Worlds project has managed to answer a number of research 
questions regarding the issues which make preservation of computer games and interactive 
fiction problematic, as well as demonstrating the feasibility of building upon existing 
packaging mechanisms with the digital preservation community for the storage of these 
unique materials, any such investigation will in turn spark new questions and lines of inquiry 
to be pursued. Based on research, there a number of questions with respect to the 
preservation of games and related content that we believe need further investigation. 
 
One of the foremost of these arises out of the work we have done on emulation as a 
preservation strategy for computer games. As discussed in chapter 6, emulation proved to be 
only a partially successful strategy. Aspects of the original game which we would, in the best 
case, wish to preserve (e.g., the music accompanying game play in DOOM, differences in 
graphic output in Star Raiders) proved difficult to maintain under emulation, and we expect 
similar problems to arise under migration approaches. The operation of any piece of 
software involves a complex interplay with a computer?s hardware and its operating system, 
and changes to any component in this dynamic can have an impact on the software?s 
performance and appearance. To expect software to stay utterly unchanged in the face of 
changing computing platforms may not be realistic, but we clearly need more research into 
how emulators might improve their performance, particularly with respect to their ability to 
successfully reproduce the behaviors of I/O devices.  
 
Preservationists have accepted the necessity of changing the form and appearance of stored 
information in order to insure its on-going accessibility; newspapers in microform are not 
identical to the paper copies they replace. But we have little experience to date with users? 
desires and expectations regarding preserved software, or with their unstated criteria for 
what constitutes ?good enough? or ?better than good enough? preservation. In the language of 
the preservation community, we do not know what properties of games our users consider 
significant. Given that the preservation strategies we have evaluated with respect to 
computer games and interactive fiction all suffer from some degree of imperfection, a first 
major research question for the preservation community is ?what are the significant 
properties of digital games which we should seek to maintain and what are there relative 
? 124?
 degrees of importance??  Without this knowledge, we are poorly positioned to select an 
appropriate preservation strategy from the options available. 
 
As demonstrated by our severe difficulties in attempting to create a preservation copy of 
islands in Second Life, massive, multiplayer virtual environments are extraordinarily difficult to 
preserve in any meaningful way. Even at the most basic level of creating a complete replica 
of the physical environment, our efforts were only (very) partially successful. Clearly further 
research must be undertaken regarding not only how we might preserve these environments, 
but how we might provide access to what we?ve preserved. The How They Got Game project 
at Stanford University has been investigating the use of an open source virtual environment 
software package, Sirikata, as a means of display 3D worlds originally built on other 
platforms. The project has already managed to export portion of the environment from id 
Software?s game Quake and re-instantiate it within Sirikata. Further research is needed, 
however, not only on how to achieve more complete translation of an original environment 
in the Sirikata platform (including such features as scripts, animations and other dynamic 
aspects of the virtual world) but also on how to present a recreated virtual world as part of a 
virtual museum of 3D worlds. 
 
While our project was able to develop mechanisms for packaging games for preservation 
that building upon existing standards within the preservation community, our work also 
revealed the practical limits of using these technologies. Generation of XML packaging files 
for complex compilations of material such as games is simply prohibitively time-consuming 
using current technologies. Research and development of tools that speed the process of 
generating packaging files for such materials, including recording the necessary relationships 
between content, representation information and context information, will be absolutely 
critical to the long-term success of the preservation community. 
 
Another recurring theme of our investigations has been that preservation activity 
surrounding computer games and interactive fiction could benefit significantly from the 
involvement of the gaming community. Gamers have documentation regarding games? use 
and history that will be incredibly valuable for scholars in the future; they often possess 
technical documentation regarding the hardware and software necessary to run a game that 
the game companies themselves may no longer possess; and they have demonstrated their 
willingness to invest long hours in the creation of emulators, websites on games and other 
activities. If this energy could be harnessed by the digital preservation community, it would 
be of immense value. 
 
Whether this is possible or not, and if so how, are questions that remain to be answered and 
that we believe deserve further investigation. Some specific questions that have arisen in the 
course of our research have been: 
 
? How might metadata systems run by the cultural heritage sector be opened up to 
contributions from the gaming community to allow them to provide enhanced 
description information, as well as context and representation information, for 
games?  Are there legal or social issues that might impede this? 
? What legal and social issues might impede libraries collaborating with the gaming 
community on development and maintenance of emulation software? 
? 125?
 ? Websites created by the gaming community documenting particular aspects of games 
have demonstrated the same longevity as other parts of the web, which is to say, they 
are highly perishable. If the library community were to provide infrastructure for the 
gaming community for some of this activity, it might help promote the survival of 
some of this material. What factors might impede this, and how can they be 
addressed? 
 
Through studying these questions with regards to the gaming community, digital 
preservationists might also gain further insight into their relationships with other 
communities of practice. 
 
 
? 126?
 10. Conclusion 
 
In 1996, as the first generation of online, 3D 
multi-user virtual environments were becoming 
increasingly known and popular on the Internet, 
the VRML Consortium created a working group 
known as ?Living Worlds.?  The Living Worlds 
group was to create specifications that would 
allow the virtual worlds created by different 
companies to be truly interoperable, with avatars 
and objects being able to move freely between 
environments provided by different companies. 
Through cooperation, the companies involved in 
the working group hoped to spur further user 
interest and participation in these evolving virtual 
worlds. 
 
Unfortunately, while draft specifications were 
developed and some test implementations were 
performed, the Living Worlds work never really 
came to fruition. The initial burst of development in multi-user virtual environments died 
quickly as companies struggled to find a sustainable business model for running a virtual 
world, and several companies involved in the Living Worlds effort went bankrupt or were 
acquired by other firms. Later entries into the multi-user virtual environment sphere, such as 
the early MMORPGs Ultima Online and Everquest, had little motivation for pursuing 
interoperability. 
 
There are several lessons that the preservation community can take from the early history of 
3D virtual worlds. The first is that without a sustainable business model, any effort to try to 
keep a virtual world alive is doomed to failure. The second, perhaps less obvious lesson is 
that involvement of the user community is a vital component of sustainability. While it has 
struggled financially over the years, ActiveWorlds (the company responsible for Alphaworld, a 
3D virtual world in operation since 1995) has managed to stay in business, and this may in 
part be ascribed to the strong community of users that has inhabited the space and their 
active involvement in its development. It does not seem coincidental that one of the very 
few companies producing 3D virtual worlds to have survived since 1995 is also the one with 
many users willing to undertake the expense of traveling to face-to-face annual ?Reunion? 
events. A final lesson is that, given the time necessary for collaborative efforts to develop, it 
is usually better to start them sooner rather than later. 
 
Our project has documented a number of significant issues impeding the preservation of 
virtual worlds, including obsolescence of original platforms, difficulties in identification and 
description of the objects of preservation, problematic rendition of games using common 
preservation strategies of migration and emulation, and numerous difficulties resulting from 
intellectual property and contract law. But we have also found several reasons to be hopeful 
that these problems can be overcome. Existing packaging formats employed by the digital 
preservation community, such as METS, OAI-ORE and BagIt, can be employed for 
A map of Alphaworld , an early, 
multi-user 3D world, in 1996 
? 127?
 packaging game materials for preservation, particularly when supplemented with ontologies 
which delineate the data models set forth in the Functional Requirements for Bibliographic Records 
Final Report and the Reference Model for an Open Archival Information System. While we 
documented several problems with existing emulation technology?s ability to successfully 
render older games, we also note that there have been successful efforts to modify emulators 
to enable them to more perfectly reproduce the original experience of game play (Bogost, 
2009). Interaction between the gaming community and game companies such as Microsoft 
and Blizzard Entertainment have resulted in those companies re-crafting their licensing 
regimes for their MMORPGs to provide machinima creators the legal permissions they need 
to create new video works using these virtual worlds. Our own collaborations with the 
gaming community and the Internet Archive have resulted in the creation of a repository of 
moving image documentation of virtual worlds within the Internet Archive that will allow 
user-created documentation of worlds to persist within a stable preservation repository. 
 
While these initial efforts provide hopeful signs that the preservation of virtual worlds is, in 
fact, a tractable problem, continuing work and collaboration by librarians, archivists, 
curators, game developers, authors of electronic literature, and the communities they all 
serve will be needed to insure the problem of preserving virtual worlds finds a permanent 
solution. Libraries, archives and museums need to start long-overdue conversations about 
how they might more effectively collaborate with each other on collection management, 
description and preservation of virtual worlds. Game companies and the digital preservation 
community need to collaborate on changes to intellectual property law that will allow game 
developers to profit from their work while insuring their work is also known to future 
generations. Cultural heritage organizations also need to actively engage the gaming 
community. A good starting-point would be a discussion of how libraries, archives and 
museums might provide stable, preservation infrastructures that the gaming community 
might use to document the history and culture of virtual worlds and to collaborative develop 
tools to insure on-going access to those worlds. 
 
In October of 2007, Linden Lab, our commercial collaborator on the Preserving Virtual 
Worlds project, and IBM announced their intention to develop open standards to enable the 
interoperability of 3D virtual worlds. By July of 2008, Linden Lab and IBM were able to 
demonstrate their first successful case of interoperability, moving avatars back and forth 
between a Second Life Preview Grid and an OpenSim virtual world server. Through their 
collaboration, the two companies were able to finally realize the dream that sparked the 
Living Worlds project over a decade ago, and through the collaboration of the cultural 
heritage community, game companies and gamers, we will be able to insure that virtual 
worlds remain living worlds in the future. 
? 128?
 Bibliography 
 
Aliaga-Buchenau, Ana-Isabel (2003). The ?Dangerous? Potential of Reading: Readers and the 
Negotiation of Power in Nineteenth Century Narratives. New York: Routledge. 
 
Avendon, Elliott M. & Brian Sutton-Smith (1971). The Study of Games. New York: J. Wiley. 
 
Bainbridge, William Sims (July 27, 2007). ?The Scientific Research Potential of Virtual 
Worlds.? Science 317(5837), pp. 472-476. 
 
Bartle, Richard (1996). Hearts, Clubs, Diamonds, Spades: Players Who Suit MUDs. Retrieved 
August 31, 2010 from http://www.mud.co.uk/richard/hcds.htm 
 
Barton, Matt (2008). Dungeons & Desktops: The History of Computer Role-Playing Games. 
Wellesley, MA: A. K. Peters. 
 
Becker, et al. (1997). Dublin Core Element: Coverage. Retrieved Aug. 30, 2010 from 
http://www.alexandria.ucsb.edu/public-documents/metadata/dc_coverage.html 
 
Bethesda Softworks (Firm) (2008). Fallout 3. Rockville, MD: Author. 
 
Boellstorff, Tom (2008). Coming of Age in Second Life: An Anthropologist Explores the Virtually 
Human. Princeton, NJ: Princeton University Press. 
 
Bogost, Ian (2007). Persuasive Games: The Expressive Power of Videogames. Cambridge, MA: The 
MIT Press. 
 
Bogost, Ian ([2009]). A Television Simulator: CRT Emulation for the Atari VCS. Retrieved 
August 10, 2010 from http://www.bogost.com/games/a_television_simulator.shtml. 
 
Bohannon, J. (2008). ?Slaying Monsters for Science.?  Science 320 (20 June), 1592. 
 
Book, Betsy (2006). ?What is a Virtual World??  Virtual Worlds Review. Retrieved August 10, 
2010 from http://www.virtualworldsreview.com/info/whatis.shtml 
 
Bryant, John (2002). The Fluid Text: A Theory of Revision for Book and Screen. Ann Arbor: 
University of Michigan Press. 
 
Burke, T. (2006). The History of Virtual Worlds (weblog comment), Terra Nova. Retrieved 
March 1, 2009 from 
http://terranova.blogs.com/terra_nova/2006/12/the_history_of_.html. 
? 129?
  
Candlelight Vigil. (2001). re: Candlelight Vigil on Luclin.  Posts by Nirrian and Keeter, Sept. 12 
2001. Retrieved Sept. 15, 2001 from http://www.everlore.com. 
 
comScore (Jan. 28, 2009). Game On! Online Gaming Surges as Gamers Seek Out Free Alternatives in 
Tight Economy. Retrieved July 17, 2010 from  
http://www.comscore.com/layout/set/popup/Press_Events/Press_Releases/2009/1/Onli
 ne_Gaming_Grows. 
 
Consultative Committee for Space Data Systems (2002). Reference Model for an Open Archival 
Information System (OAIS) (CCSDS 650.0-B-1). Washington, DC: CCSDS Secretariat, National 
Aeronautics and Space Administration. 
 
Consultative Committee for Space Data Systems (2008). XML Formatted Data Unit (XFDU) 
Structure and Construction Rules. (CCSDS 661.0-B-1). Washington, DC: CCSDS Secretariat, 
National Aeronautics and Space Administration. 
 
Coyle, Karen (2004). ?Future Considerations: The Functional Library Systems Record.? 
Library Hi Tech 22(2), pp. 166-174. 
 
Dalenberg, Russel (2004). ?Adventure Family Tree.? Russel Dalenberg?s Home Page [personal 
website]. Retrieved August 31, 2010 from http://www.io.com/~ged/www/family.html. 
 
Dalenberg, Russel (2006). ?Versions and Ports of Adventure known to exist, v. 2.6.7.? Russel 
Dalenberg?s Home Page [personal website]. Retrieved August 31, 2010 from 
http://www.io.com/~ged/www/advelist.html. 
 
Debord, Guy-Ernest (November 1956). ?Th?orie de la D?rive.? Les Levres Nues 9. 
 
Easter egg (June 24, 2010). The Doom Wiki. Retrieved August 31, 2010 from 
http://doom.wikia.com/wiki/Easter_egg. 
 
Easter Eggs and cheats (n.d.). ONLINE SIERRA ? The Sierra Online Fan Site. Retrieved 
August 31, 2010 from http://tawmis.com/onlinesierra/eggsandcheats.html. 
 
Electronic Frontier Foundation (2010). Coders? Rights Project Reverse Engineering FAQ. San 
Francisco, CA: Author. 
 
Entertainment Software Association (2010). Essential Facts about the Computer and Video Game 
Industry: 2010 Sales, Demographic and Usage Data. Retrieved Aug. 17, 2010 from 
http://www.theesa.com/facts/pdfs/ESA_Essential_Facts_2010.PDF. 
 
Fitz, Kate (2008). ?Terms of service (Second Life) - History of changes.? Second Life LawSpot. 
Retrieved March 24, 2008 from http://www.lawspotonline.com/lawspot/termsofservice.jsp. 
 
GamePolitics.com (Oct. 9, 2008). Report: Obama Ads in Burnout Paradise. Available at: 
http://www.gamepolitics.com/2008/10/09/report-obama-ads-burnout-paradise. 
 
? 130?
 Golbeck, Jennifer (2005). Computing and Applying Trust in Web-based Social Networks. Ph.D. 
Dissertation. Retrieved Aug. 30, 2010 from 
http://trust.mindswap.org/papers/GolbeckDissertation.pdf. 
 
Grimes, Justin, Paul Jaeger and Kenneth Fleischmann (2008). ?Obfuscatocracy: A 
Stakeholder Analysis of Governing Documents for Virtual Worlds,? First Monday 13(9) 
http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/2153/2029. 
 
Hayes, Christina (2008). ?Changing the Rules of the Game: How Video Game Publishers 
Are Embracing User-Generated Derivative Works,? Harvard Journal of Law & Technology 
21(2): 567-587.  
 
IFLA Study Group on the Functional Requirements for Bibliographic Records (1997). 
Functional Requirements for Bibliographic Records: Final Report. The Hague: International 
Federation of Library Associations and Institutions. 
 
Ippolito, Jon (2008) ?Death by Wall Label.? In Christiane Paul (ed.), New Media in the White 
Cube and Beyond. Berkeley: University of California Press.  
 
Ippolito, Jon, et al. The Pool.  Retrieved January 30, 2010 from 
http://pool.newmedia.umaine.edu/index.php. 
 
J. Paul Getty Trust and the College Art Association (2009). Categories for the Description of 
Works of Art. Patricia Harpring and Murtha Baca (Eds.). Los Angeles, CA: Author. 
 
Jerz, Dennis (2007). ?Somewhere Nearby is Colossal Cave: Examining Will Crowther?s 
Original ?Adventure? in Code and in Kentucky? DHQ 1 
http://digitalhumanities.org/dhq/vol/001/2/000009/000009.html. 
 
John, Jeremy L., et al. (2010). Digital Lives / Personal Digital Archives for the 21st Century: An 
Initial Synthesis. March 3, 2010. Beta Version 0.2. 
 
Juul, Jesper (2005). Half-Real Video Games Between Real Rules and Fictional Worlds. Cambridge, 
MA: MIT Press. 
 
Kirschenbaum, Matthew, Richard Ovenden, and Gabriela Redwine (2010). Digital Forensics 
and Born-Digital Content in Cultural Heritage Collections. Washington, DC: Council on Library 
and Information Resources. 
 
Kraus, Kari (2009). ?Conjectural Criticism: Computing Past and Future Texts.? Digital 
Humanities Quarterly 3 http://digitalhumanities.org/dhq/vol/3/4/000069/000069.html. 
 
Kraus, Kari, Rachel Donahue, Megan Winget (2009). ?Game Change: The Role of Amateur 
and Professional Cultures in Preserving Virtual Worlds,? Digital Humanities Conference, College 
Park, MD (June 22-29 2009). Retrieved Aug. 30, 2010 from 
http://www.mith2.umd.edu/dh09/index.html%3Fpage_id=99.html. 
  
Lessig, Lawrence (2008). Remix: Making Art and Commerce Thrive in the Hybrid Economy. New 
York: Penguin Press. 
? 131?
  
 
Library of Congress (2002). Preserving our Digital Heritage: Plan for the National Digital Information 
Infrastructure and Preservation Program. Washington, DC: Author. 
 
Library of Congress (2007). Sustainability of Digital Formats: Planning for Library of Congress 
Collections. Washington, DC: Author. 
 
Library of Congress Network Development & MARC Standards Office (2008). MODS 3.3. 
Washington, DC: Author. 
 
Library of Congress Network Development & MARC Standards Office (2009). 
MARCXML: The MARC 21 XML Schema. Washington, DC: Author. 
 
Lowood, H. (2008). ?Impotence and Agency: Computer Games as a Post-9/11 Battlefield,? 
in Andreas Jahn-Sudmann & Ralf Stockmann (Eds.), Games Without Frontiers - War Without 
Tears. Computer Games as a Sociocultural Phenomenon. London: Palgrave Macmillan: 78-86. 
 
McDonough, Jerome, Matthew Kirschenbaum, Doug Reside, Neil Fraistat & Dennis Jerz 
(Fall 2010). ?Twisty Little Passages Almost All Alike: Applying the FRBR Model to a Classic 
Computer Game.?  Digital Humanities Quarterly 4(2). 
 
McDonough, Jerome & Mona Jimenez (April 2007). ?Video Preservation and Digital 
Reformatting: Pain and Possibility.?  Journal of Archival Organization 4(1/2), pp. 167-191. 
 
Microsoft Corp., Inc. (2009). Xbox.com, Game Content Usage Rules. Retrieved August 31, 2010 
from http://www.xbox.com/en-US/community/developer/rules.htm. 
 
Monnens, Devin, Andrew Armstrong, Judd Ruggill, Ken McAllister, Zach Vowell & Rachel 
Donahue (2009). Before It?s Too Late: A Digital Game Preservation White Paper. Henry Lowood 
(Ed.). Mt. Royal, NJ: International Game Developers Association. Retrieved Aug. 11, 2010 
from http://www.igda.org/wiki/images/8/83/IGDA_Game_Preservation_SIG_-
 _Before_It%27s_Too_Late_-_A_Digital_Game_Preservation_White_Paper.pdf. 
 
Montfort, Nick, (2003). Twisty Little Passages: An Approach to Interactive Fiction. Cambridge, MA: 
MIT Press. 
 
Montfort, Nick & Ian Bogost (2009). Racing the Beam: The Atari Video Computer System. 
Cambridge, MA: MIT Press. 
 
? 132?
 Montfort, Nick, et al. (2004). Mystery House Taken Over. Commissioned by New Radio and 
Performing Arts, Inc. Retrieved January 30, 2010 from 
http://www.turbulence.org/Works/mystery/. 
 
NoSkill Memorial Site. Retrieved Dec. 1, 2004 from 
http://www.doom2.net/noskill/index.htm. 
 
Pearce-Moses, Richard (2005). A Glossary of Archival and Records Terminology. Chicago, IL: The 
Society of American Archivists. Retrieved Aug. 11, 2010 from 
http://www.archivists.org/glossary/index.asp. 
 
Pinsky, R., & Synapse Software Corporation (1984). Mindwheel. Richmond, CA: Synapse 
Software Corp. & Broderbund. 
 
Ross, Seamus (2010). Personal Communication. Computer Forensics and Cultural Heritage 
Symposium, University of Maryland.  
 
Rossignol, Jim (2008). This Gaming Life: Travels in Three Cities. Ann Arbor, MI: University of 
Michigan Press. 
 
Rubin, Stephen (2010). ?Intellectual Property Content, Law, and Practice.? In Steve Rabin 
(Ed.) Introduction to Game Development, 2nd Edition, Course Technology.  Charles River Media. 
 
Ruecker, Stan, St?fan Sinclair & Milena Radzikowska, (2007). ?Confidence, Visual Research 
and the Aesthetic Function.? Partnership: the Canadian Journal of Library and Information Practice 
and Research. 2(1). 
 
Sampson, Walker (July 9, 2010).  "Modeling a Computer on Omeka [blog post].? iSchool 
Digital Humanities Network Blog. Retrieved Aug. 31, 2010 from 
http://www.ischooldh.org/network/. 
 
Section 108 Study Group (2008). Executive Summary Report, Retrieved August 30, 2010 from 
http://www.section108.gov/docs/Sec108ExecSum.pdf. 
 
Siwek, Stephen E. (2007). Video Games in the 21st Century: Economic Contributions of the US 
Entertainment Software Industry. Entertainment Software Association. Available at: 
http://www.theesa.com/facts/pdfs/VideoGames21stCentury.pdf . 
 
Visual Resources Association Data Standards Committee (2007). VRA Core 4.0. Kansas City, 
MO: Visual Resources Association. 
 
Williams, D. (2006). ?Groups and Goblins: The Social and Civic Impact of Online Gaming.? 
Journal of Broadcasting and Electronic Media 50,  pp. 651-670. 
 
Woolf, Harry (Ed.) (March 1962). ?The Conference on Science Manuscripts.?  ISIS 53(171), 
pp. 1-158. 
 
? 133?
 World Wide Web Consortium (27 June 2001). XML Linking Language (XLink) Version 1.0. 
Steve DeRose, Eve Maler and David Orchard (Eds.). Cambridge, MA: Massachusetts Inst. 
of Technology. Retrieved Aug. 31, 2010 from http://www.w3.org/TR/xlink/. 
 
World Wide Web Consortium (10 February 2004). RDF/XML Syntax Specification (Revised). 
Dave Beckett (Ed.). Cambridge, MA: Massachusetts Inst. of Technology. Retrieved Aug. 31, 
2010 from http://www.w3.org/TR/REC-rdf-syntax/. 
 
WorldofWarcraft.com, Letter to the Machinimators of the World. Retrieved May 12, 2008 
from http://www.worldofwarcraft.com/community/machinima/letter.html. 
 
Wu, Jia (Feb. 5, 2010). Global Video Game Market Forecast. Strategy Analytics. 
 
 
? 134?
 Appendix A - Virtual Worlds that Died During the Grant 
 
Virtual 
World 
Publisher Description DoB Date 
Died 
Cause of 
Death 
Links 
1 Aradath/Drag
 on's Gate 
Mythic 
Realms 
(when 
closed) 
Gamer's 
World, AUSI, 
Genie, AOL, 
Mythic 
Realms 
1984 10 
February 
2007 
  
2 Championshi
 p Manager 
Online 
Beautiful 
Game 
Studios/Ja
 destone 
Publisher: 
Eidos. 
Cancelled by 
Square Enix 
after it 
acquired 
Eidos. 
2005 30 April 
2010 
 http://www.cm-
 online.com/main/acti
 on/forum/thread/232
 78?
 ?
 http://www.computer
 andvideogames.co
 m/article.php?id=23
 3969?
 ?
 http://www.edge-
 online.com/news/ch
 ampionship-
 manager-online-to-
 close 
3 Cities XL Monte 
Cristo 
Games 
a Sim City-
 esque game 
with MMO 
and social 
networking 
components. 
9 
Octo
 ber 
2009 
8 March 
2010 
Low 
subscriber 
numbers are 
leading Monte 
Cristo to pull 
back the 
online 
component 
and focus on 
its single-
 player side. 
http://www.virtualwo
 rldsnews.com/2010/
 01/cities-xl-to-close-
 mmo-due-to-small-
 subscriber-
 base.html?
 ?
 http://www.massivel
 y.com/2010/01/27/ci
 tiesxl-to-close-
 multiplayer-features/ 
4 EA-Land 
(AKA The 
Sims Online) 
Electronic 
Arts 
MMOG 
based on the 
video game 
The Sims 
17 
Dece
 mber 
2002 
4.35am 
PST on 1 
August 
2008 
?Widely seen 
as a failed 
attempt to port 
the single-
 player game 
to an online, 
multiplayer 
environment. 
Still, EA kept 
TSO running, 
even as it was 
eclipsed by 
other social 
virtual worlds, 
and it limped 
along with a 
small 
http://www.stanford.
 edu/group/htgg/cgi-
 bin/drupal/?q=node/
 239?
 ?
 http://news.cnet.com
 /8301-13772_3-
 9931757-52.html?
 ?
 http://terranova.blog
 s.com/terra_nova/20
 08/04/ea-to-close-
 ea.html 
? 135?
 membership.? 
5 Faketown Identity 
Play 
A free MMO 
created in a 
pixel art 
aesthetic, 
Faketown 
features 
drawing and 
animation 
tools, along 
with photo 
and mp3 
hosting, as 
well as 
YouTube 
videos. 
 23 June 
2008 
"The site and 
underlying 
technology 
are currently 
on auction to 
the highest 
bidder," said 
Yeary via 
email. "We 
still believe in 
the product 
and we are 
optimistic that 
someone will 
see the value 
in re-
 establishing 
the 
community." 
http://www.zazzle.co
 m/rip_faketown_tshi
 rt-
 2359978227933893
 36?
 ?
 http://www.virtualwo
 rldsnews.com/2008/
 07/faketown-
 closes.html?
 ?
 http://www.virtualwo
 rldsnews.com/2008/
 07/interview-
 lesso.html 
6 GoPets Zynga Virtual pet 
site where 
owners could 
customize 
3D pets. 
4 
Augu
 st 
2005 
8 
Novemb
 er  2009 
Chose to 
focus on 
Petville 
instead? 
http://en.wikipedia.or
 g/wiki/GoPets?
 ?
 http://www.virtualwo
 rldsnews.com/2009/
 12/this-week-zynga-
 unleashed-its-latest-
 social-game-petville-
 -as-the-name-
 implies-petville-is-
 zyngas-entry-into-
 the-time-test.html 
7 Hellgate: 
London 
Flagship 
Studios / 
Hanbitsoft 
Fantasy RPG 31 
Octo
 ber 
2007 
31 
January 
2009 
Hanbitsoft has 
acquired the 
US/EU 
territory rights 
and will be re-
 releasing back 
to the US/EU 
territories with 
its sequel 
Hellgate 
London: 
Resurrection. 
?
 ?
 http://en.wikipedia.or
 g/wiki/Hellgate:_Lon
 don 
8 Lively Google Web-based 
virtual 
environment 
integrated 
with 
YouTube and 
Picasa. 
8 
July 
2008 
31 
Decemb
 er 2008 
?But we've 
also always 
accepted that 
when you take 
these kinds of 
risks not every 
bet is going to 
pay off.? 
http://googleblog.blo
 gspot.com/2008/11/l
 ively-no-more.html?
 ?
 http://www.lively.co
 m/goodbye.html?
 ?
 http://arstechnica.co
? 136?
 m/old/content/2008/
 11/google-to-shut-
 down-lively-its-
 interactive-3d-
 world.ars 
9 The Matrix 
Online 
Sony 
Online 
Entertain
 ment 
A MMOG 
developed by 
Monolith 
Productions. 
It was the 
official 
continuation 
of the 
storyline of 
the Matrix 
series of 
films. 
22 
Marc
 h 
2005 
31 July 
2009 
?doubts about 
the game 
circled within 
the industry, 
based on the 
lacklustre 
reception of 
the later two 
Matrix films 
and an 
overcrowded 
MMORPG 
market.? 
http://en.wikip
 edia.org/wiki/T
 he_Matrix_On
 line 
http://thematrixonlin
 e.station.sony.com/i
 ndex.vm?
 ?
 http://www.joystiq.co
 m/2009/08/04/the-
 matrix-onlines-final-
 moments-
 documented/?
 ?
 http://thematrixonlin
 e.station.sony.com/g
 ame_basics.vm 
10 Meridian 59 Near 
Death 
Studios 
Sword and 
Sorcery 
combat RPG 
15 
Dece
 mber 
1995 
Announc
 ed 5 
January 
2010 
"Unfortunately
 , M59 never 
really grew," 
Green admits 
in his personal 
blog. "We 
were lucky 
that we got a 
lot of attention 
for keeping an 
old game alive 
from the 
press. We 
also had a 
small and 
dedicated 
group of fans 
willing to keep 
the game 
alive. But, the 
press didn't 
really care 
about our 
attempts to 
improve the 
game, and the 
fans weren't 
interested in 
trying to 
attract new 
players." 
http://meridian59.ne
 ardeathstudios.com/?
 ?
 http://www.gamasutr
 a.com/view/news/26
 677/Meridian_59_D
 eveloper_Near_Dea
 th_Studios_Closes.p
 hp?
 ?
 http://www.gamepro.
 com/article/news/21
 3452/near-death-
 gives-up-the-ghost-
 after-nine-years-of-
 meridian-59/ 
? 137?
 11 Metaplace Areae The 
Metaplace 
website 
promises that 
the platform 
will integrate 
smoothly into 
our current 
web 
standards, 
allowing for 
integration of 
Metaplace 
elements into 
websites, 
RSS feeders, 
and more. 
 1 
January 
2010 
It just hasn?t 
gotten traction 
http://www.raphkost
 er.com/2009/12/21/
 metaplace-com-
 closing/?
 ?
 http://www.massivel
 y.com/2008/02/25/g
 dc08-raph-kosters-
 reinventing-mmos-a-
 metaplace-
 antemortem/?
 ?
 http://www.metaplac
 eveterans.com/foru
 m/viewforum.php?f=
 3&sid=3d45064c8a0
 c864cac00a0cc20fa
 1812 
12 Phantasy 
Star Online 
    
Sega     On 
Dreamcast. 
21 
Nove
 mber 
2000 
31 March 
2007 
  
13 Tabula Rasa NCsoft A MMORPG 
developed by 
Destination 
Games about 
humanity's 
last stand 
against a 
group of 
aliens called 
"The Bane". 
http://eu.play
 nc.com/eu/ga
 mes/overvie
 w/tabula_ras
 a/ 
2 
Nove
 mber 
 200
 7 
28 
February 
2009 
"Unfortunately
 , the fact is 
that the game 
hasn't 
performed as 
expected."?
 ?
 http://www.ga
 masutra.com/
 php-
 bin/news_inde
 x.php?story=2
 1224 
http://eu.plaync.com
 /eu/games/info/tabul
 a_rasa/?
 ?
 http://www.gamasutr
 a.com/php-
 bin/news_index.php
 ?story=22510?
 ?
 http://www.gamepoli
 tics.com/2009/05/06
 /richard-garriott-
 sues-nc-soft-over-
 millions-stock-
 options?
 ?
 http://arstechnica.co
 m/gaming/news/200
 9/03/does-a-game-
 have-to-fail-to-have-
 an-ending-tabula-
 rasa.ars?
 ?
 http://en.wikipedia.or
 g/wiki/Tabula_Rasa
 _%28video_game%
 29 
14 There.com Makena 
Technolog
 ies 
A 3D online 
virtual world. 
From home 
page: "...an 
9 
Janu
 ary 
2003 
9 March 
2010 
Recession: "... 
at the end of 
the day, we 
can't cure the 
http://www.prod.ther
 e.com/info/announc
 ement?
 ?
? 138?
 online 
getaway 
where you 
can hang out 
with your 
friends and 
meet new 
ones..." 
recession, 
and at some 
point we have 
to stop writing 
checks to 
keep the 
world open," 
says Wilson. 
"There's 
nothing more 
we would like 
to avoid this, 
but There is a 
business, and 
a business 
that can't 
support itself 
doesn't work." 
http://news.cnet.com
 /8301-13772_3-
 10462627-52.html?
 ?
 http://www.virtualwo
 rldsnews.com/2010/
 03/therecom-
 shutting-down-on-
 march-9th11.html 
15 Virtual Magic 
Kingdom 
Walt 
Disney 
Parks and 
Resorts 
Online 
MMOG 
created as 
part of their 
50th 
anniversary 
with areas 
and games 
based on 
real park 
scenery and 
attractions 
open to the 
public daily 
between 
7:00am?
 10:00pm 
PST. 
23 
May 
2005 
21 May 
2008 
Created for 
50th Ann and 
promotion 
ended long 
after originally 
planned. 
http://www.intercot.c
 om/discussion/show
 thread.php?t=13054
 8?
 ?
 http://familyinternet.
 about.com/od/websit
 es/tp/most-
 influential-virtual-
 worlds.htm?
 ?
 http://www.worldsin
 motion.biz/2008/04/
 disney_closes_gate
 s_to_virtual.php 
16 Vivaty Vivaty Vivaty is a 
next 
generation 
3D virtual 
world in the 
browser 
where you 
can meet 
new friends 
and express 
yourself by 
dressing up 
your avatar 
and 
personalizing 
your own 
virtual scene 
for free. 
Marc
 h 
2008 
16 April 
2010 
?Vivaty.com is 
a rather 
expensive site 
to run, much 
more than a 
regular web 
site, and 
Vivaty the 
company has 
been running 
out of money 
for some 
time.??
 Jay Weber, 
Co-founder 
and Chief 
Technical 
Officer 
http://blog.vivaty.co
 m/2010/03/31/vivaty
 -shutdown-party/?
 ?
 http://www.raphkost
 er.com/2010/03/31/v
 ivaty-is-closing-
 down/?
 ?
 http://games.venture
 beat.com/2010/03/3
 1/vivaty-shuts-down-
 virtual-world/?
 ?
 http://blog.vivaty.co
 m/2008/03/31/welco
 me-to-vivaty/ 
? 139?
 17 Weblin Zweitgeist A software 
that turns the 
internet into 
a virtual 
world where 
you can chat 
with people 
from all over 
the world 
 Announc
 ed 7 Aug 
2009 
?Media reports 
attribute the 
closure to a 
lack of funds.? 
http://www.virtualwo
 rldsnews.com/2009/
 08/weblin-ceases-
 operations.html?
 ?
 http://www.raphkost
 er.com/2009/08/08/
 weblins-
 closing/?utm_source
 =feedburner&utm_m
 edium=feed&utm_c
 ampaign=Feed%3A
 +RaphsWebsite+%2
 8Raph%27s+Websit
 e%29 
?
  
? 140?
 Appendix B ? Media Coverage of  PVW 
?
 Kotaku      October 2007                      
?The Library of Congress Loves Video Games? 
http://kotaku.com/313328/the-library-of-congress-loves-video-games 
 
Kotaku      November 2008 
?On the LOC Preserving Virtual Worlds Project? 
http://kotaku.com/5096782/on-the-loc-preserving-virtual-worlds-project 
 
Special Collections 2.0: New Technologies 2009 
For Rare Books, Manuscripts and Archival 
Collections 
On p. 14: ?Creative born-digital works in Second Life are also earmarked for long term 
preservation; the Library of Congress recently provided a grant to several universities, 
spearheaded by the University of Illinois at Urbana-Champaign, to develop digital 
preservation standards for Second Life, as part of their Preserving Creative America 
program.? 
 
Crispy Gamer     January 2009 
?Saved Games: Preserving the New TV? 
http://www.crispygamer.com/features/2009-04-08/saved-games-preserving-the-new-
 tv.aspx 
 
MetaverseTV     April 2009 
ttp://blip.tv/file/2003603 
 
Blog post: Crispy Gamer   April 2009 
http://www.crispygamer.com/features/2009-04-08/saved-games-preserving-the-new-
 tv.aspx 
 
Salon      April 2009 
?How to Make Machinima Without Getting Sued Blind? 
http://www.salon.com/technology/the_gigaom_network/online_video/index.html?section
 =online_video&blog=/tech/giga_om/online_video/2009/04/28/how_to_make_machinim
 a_without_getting_sued_blind 
 
Kotaku      April 2009 
?What AJ Learned About Machinima Law Today? 
http://kotaku.com/5226851/what-aj-learned-about-machinima-law-today 
 
IEEE Spectrum    April 2009 
?Stanford Boots Up Machinima? 
http://spectrum.ieee.org/sandbox/consumer-
 electronics/gaming/stanford_boots_up_machinima 
 
? 141?
 Future Tense, American Public Media  April 2009 
?Machinima lowers barrier of entry to filmmaking, but raises legal questions? 
http://www.publicradio.org/columns/futuretense/2009/04/machinima-lower.html 
 
Blog post: Pixel Vixen    May 2009 
http://www.pixelvixen707.com/?p=1743/#content 
 
Stanford Magazine    November/December 2009      
?Saving Worlds Preserving the digital and virtual?    
http://www.stanfordalumni.org/news/magazine/2008/novdec/farm/news/virtual.html      
 
Voice of America with Art Chimes  March 2010 
http://www1.voanews.com/english/news/science-technology/Our-World--13-March-
 2010-87482377.html   at ~15:53 
 
Atlantic Monthly    March 2010           
?Pac Rat?     
http://www.theatlantic.com/doc/201003/archiving-video-games 
 
Kojo Nnamdi Show    March 2010 
http://thekojonnamdishow.org/shows/2010-03-30/preserving-video-games-and-virtual-
 worlds 
 
University of Maryland Diamondback  April 2010 
?A Second Life? 
http://www.diamondbackonline.com/news/a-second-life-1.1434797 
http://thekojonnamdishow.org/shows/2010-03-30/preserving-video-games-and-virtual-
 worlds 
 
Baltimore Public Radio   May 2010 
http://stream.publicbroadcasting.net/production/mp3/wypr/local-wypr-899636.mp3 
http://mdmorn.wordpress.com/2010/05/03/53102-the-perplexing-task-of-archiving-
 virtual-worlds/ 
 
Ars Technica     June 2010 
?Saving Virtual Worlds from Extinction? 
http://arstechnica.com/gaming/news/2010/06/the-art-of-archiving-virtual-worlds.ars 
 
WILL-AM 580?s Sidetrack   July 2010 
http://will.illinois.edu/sidetrack/program/sidetrackjul10/ 
 
USA Today     August 2010 
?Video Game Hall of Fame Inducting Pac-Man and Pals? 
http://www.usatoday.com/tech/gaming/2010-08-05-gamearchive05_ST_N.htm 
 
Christian Science Monitor   August 2010 
?Video game museum gives arcade classics extra lives? 
http://www.csmonitor.com/Innovation/Tech/2010/0805/Video-game-museum-gives-
? 142?
 arcade-classics-extra-lives 
 
This Way Up, Radio New Zealand  August 2010 
http://www.radionz.co.nz/national/programmes/thiswayup 
 
ABC News Technology site   August 2010 
?Hall of Fame playing up video games? 
http://abcnews.go.com/Technology/hall-fame-playing-video-
 games/story?id=11328263&page=1 
http://stream.publicbroadcasting.net/production/mp3/wypr/local-wypr-899636.mp3  
? 143?
 Appendix C ? Preserving Virtual Worlds Ontology 
?
 <?xml version="1.0"?> 
<rdf:RDF 
    xmlns="http://people.lis.uiuc.edu/~jmcdonou/PVW.owl#" 
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" 
    xmlns:protege="http://protege.stanford.edu/plugins/owl/protege#" 
    xmlns:xsp="http://www.owl-ontologies.com/2005/08/07/xsp.owl#" 
    xmlns:p1="http://www.owl-ontologies.com/assert.owl#" 
    xmlns:owl="http://www.w3.org/2002/07/owl#" 
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" 
    xmlns:swrl="http://www.w3.org/2003/11/swrl#" 
    xmlns:swrlb="http://www.w3.org/2003/11/swrlb#" 
    xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#" 
  xml:base="http://people.lis.uiuc.edu/~jmcdonou/PVW.owl"> 
  <owl:Ontology rdf:about=""/> 
  <owl:Class rdf:ID="Item"> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Item"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasReproduction"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf rdf:resource="http://www.w3.org/2002/07/owl#Thing"/> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Manifestation"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Work"/> 
    </owl:disjointWith> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Item"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isPartOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="exemplifies"/> 
        </owl:onProperty> 
        <owl:cardinality rdf:datatype="http://www.w3.org/2001/XMLSchema#int" 
        >1</owl:cardinality> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Object"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Concept"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Expression"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Person"/> 
    </owl:disjointWith> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Item"/> 
        <owl:onProperty> 
? 144?
           <owl:ObjectProperty rdf:ID="hasPart"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith> 
      <owl:Class rdf:ID="Event"/> 
    </owl:disjointWith> 
  </owl:Class> 
  <owl:Class rdf:about="#Concept"> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >an abstract notion or idea</rdfs:comment> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Object"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Event"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Expression"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Item"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Manifestation"/> 
    </owl:disjointWith> 
  </owl:Class> 
  <owl:Class rdf:about="#Object"> 
    <owl:disjointWith rdf:resource="#Item"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Manifestation"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Event"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >a material thing</rdfs:comment> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Expression"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
  </owl:Class> 
  <owl:Class rdf:about="#Event"> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >an action or occurrence</rdfs:comment> 
? 145?
     <owl:disjointWith rdf:resource="#Item"/> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Expression"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Manifestation"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
  </owl:Class> 
  <owl:Class rdf:about="#Manifestation"> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Expression"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Manifestation"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#hasPart"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith rdf:resource="#Item"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Manifestation"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#isPartOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >the physical embodiment of an expression of a work</rdfs:comment> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <rdfs:subClassOf rdf:resource="http://www.w3.org/2002/07/owl#Thing"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isReproductionOf"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Manifestation"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
  </owl:Class> 
  <owl:Class rdf:about="#Expression"> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >the intellectual or artistic realization of a work)</rdfs:comment> 
    <rdfs:subClassOf rdf:resource="http://www.w3.org/2002/07/owl#Thing"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#hasPart"/> 
        </owl:onProperty> 
? 146?
         <owl:allValuesFrom rdf:resource="#Expression"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Expression"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#isPartOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Place"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:cardinality rdf:datatype="http://www.w3.org/2001/XMLSchema#int" 
        >1</owl:cardinality> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="realizes"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith rdf:resource="#Manifestation"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <owl:disjointWith rdf:resource="#Item"/> 
  </owl:Class> 
  <owl:Class rdf:about="#Place"> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Person"/> 
    </owl:disjointWith> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Manifestation"/> 
    <owl:disjointWith rdf:resource="#Item"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >a location; encompasses a comprehensive range of locations: terrestrial and extra-
 terrestrial; historical and contemporary; geographic features and geo-political 
jurisdictions</rdfs:comment> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <owl:disjointWith rdf:resource="#Expression"/> 
  </owl:Class> 
  <owl:Class rdf:about="#Person"> 
    <owl:disjointWith rdf:resource="#Place"/> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Manifestation"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >an individual; encompasses individuals that are deceased as well as those that are 
living.</rdfs:comment> 
    <owl:disjointWith rdf:resource="#Expression"/> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Work"/> 
? 147?
     </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <owl:disjointWith rdf:resource="#Item"/> 
  </owl:Class> 
  <owl:Class rdf:about="#Work"> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasSupplement"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasAdaptation"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith> 
      <owl:Class rdf:about="#Corporate_Body"/> 
    </owl:disjointWith> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <owl:disjointWith rdf:resource="#Item"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >a distinct intellectual or artistic creation</rdfs:comment> 
    <owl:disjointWith rdf:resource="#Expression"/> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith rdf:resource="#Manifestation"/> 
    <rdfs:subClassOf rdf:resource="http://www.w3.org/2002/07/owl#Thing"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isImitationOf"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasTransformation"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasSuccessor"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#hasPart"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isSuccessorOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
? 148?
       <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="complements"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasSummary"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasImitation"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isSummaryOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith rdf:resource="#Place"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:about="#isPartOf"/> 
        </owl:onProperty> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isAdaptationOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <owl:disjointWith rdf:resource="#Person"/> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isTransformationOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="isSupplementOf"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
    <rdfs:subClassOf> 
      <owl:Restriction> 
        <owl:allValuesFrom rdf:resource="#Work"/> 
        <owl:onProperty> 
          <owl:ObjectProperty rdf:ID="hasComplement"/> 
        </owl:onProperty> 
      </owl:Restriction> 
    </rdfs:subClassOf> 
? 149?
     <owl:disjointWith rdf:resource="#Concept"/> 
  </owl:Class> 
  <owl:Class rdf:about="#Corporate_Body"> 
    <owl:disjointWith rdf:resource="#Expression"/> 
    <owl:disjointWith rdf:resource="#Work"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >an organization or group of individuals and/or organizations acting as a 
unit</rdfs:comment> 
    <owl:disjointWith rdf:resource="#Place"/> 
    <owl:disjointWith rdf:resource="#Object"/> 
    <owl:disjointWith rdf:resource="#Event"/> 
    <owl:disjointWith rdf:resource="#Person"/> 
    <owl:disjointWith rdf:resource="#Concept"/> 
    <owl:disjointWith rdf:resource="#Manifestation"/> 
    <owl:disjointWith rdf:resource="#Item"/> 
  </owl:Class> 
  <owl:ObjectProperty rdf:ID="hasGamePlayData"> 
    <rdfs:subPropertyOf> 
      <owl:ObjectProperty rdf:ID="hasContextInformation"/> 
    </rdfs:subPropertyOf> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isGamePlayDataFor"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isExemplifiedBy"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
    <rdfs:range rdf:resource="#Item"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#exemplifies"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasContextInformation"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isContextInformationFor"/> 
    </owl:inverseOf> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Item"/> 
          <owl:Class rdf:about="#Manifestation"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasRuntimeContext"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isRuntimeContextFor"/> 
    </owl:inverseOf> 
    <rdfs:subPropertyOf rdf:resource="#hasContextInformation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasSemanticRepresentationInformation"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isSemanticRepresentationInformationFor"/> 
    </owl:inverseOf> 
    <rdfs:subPropertyOf> 
      <owl:ObjectProperty rdf:ID="hasRepresentationInformation"/> 
    </rdfs:subPropertyOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isCreatedBy"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
? 150?
           <owl:Class rdf:about="#Person"/> 
          <owl:Class rdf:about="#Corporate_Body"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain rdf:resource="#Work"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isTransformationOf"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#hasTransformation"/> 
    </owl:inverseOf> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#exemplifies"> 
    <owl:inverseOf rdf:resource="#isExemplifiedBy"/> 
    <rdfs:domain rdf:resource="#Item"/> 
    <rdfs:range rdf:resource="#Manifestation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasSuccessor"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#isSuccessorOf"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isRealizedThrough"> 
    <rdfs:range rdf:resource="#Expression"/> 
    <rdfs:domain rdf:resource="#Work"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#realizes"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasTransformation"> 
    <owl:inverseOf rdf:resource="#isTransformationOf"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
? 151?
         <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isContextInformationFor"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <owl:inverseOf rdf:resource="#hasContextInformation"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Item"/> 
          <owl:Class rdf:about="#Manifestation"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasAlternate"> 
    <rdfs:range rdf:resource="#Manifestation"/> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isAlternateOf"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isArrangementOf"> 
    <rdfs:range rdf:resource="#Expression"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="hasArrangement"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isSupplementOf"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#hasSupplement"/> 
    </owl:inverseOf> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasSupplement"> 
    <owl:inverseOf rdf:resource="#isSupplementOf"/> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
? 152?
     <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasAsSubject"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Event"/> 
          <owl:Class rdf:about="#Concept"/> 
          <owl:Class rdf:about="#Person"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Corporate_Body"/> 
          <owl:Class rdf:about="#Place"/> 
          <owl:Class rdf:about="#Object"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain rdf:resource="#Work"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasRevision"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Expression"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isRevisionOf"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="embodies"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isEmbodiedIn"/> 
    </owl:inverseOf> 
    <rdfs:range rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasComplement"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#complements"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#complements"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
? 153?
       <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf rdf:resource="#hasComplement"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasArrangement"> 
    <rdfs:range rdf:resource="#Expression"/> 
    <owl:inverseOf rdf:resource="#isArrangementOf"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isSemanticRepresentationInformationFor"> 
    <rdfs:subPropertyOf> 
      <owl:ObjectProperty rdf:ID="isRepresentationInformationFor"/> 
    </rdfs:subPropertyOf> 
    <owl:inverseOf rdf:resource="#hasSemanticRepresentationInformation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isPartOf"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#hasPart"/> 
    </owl:inverseOf> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasAnalyticInformation"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isAnalyticInformationFor"/> 
    </owl:inverseOf> 
    <rdfs:subPropertyOf rdf:resource="#hasContextInformation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasSyntacticRepresentationInformation"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isSyntacticRepresentationInformationFor"/> 
    </owl:inverseOf> 
    <rdfs:subPropertyOf> 
      <owl:ObjectProperty rdf:about="#hasRepresentationInformation"/> 
    </rdfs:subPropertyOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isEmbodiedIn"> 
    <owl:inverseOf rdf:resource="#embodies"/> 
    <rdfs:range rdf:resource="#Manifestation"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasReproduction"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#isReproductionOf"/> 
    </owl:inverseOf> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
? 154?
         </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isRealizedBy"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Person"/> 
          <owl:Class rdf:about="#Corporate_Body"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasSummary"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#isSummaryOf"/> 
    </owl:inverseOf> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isReconfigurationOf"> 
    <rdfs:domain rdf:resource="#Item"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="hasReconfiguration"/> 
    </owl:inverseOf> 
    <rdfs:range rdf:resource="#Item"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isImitationOf"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#hasImitation"/> 
    </owl:inverseOf> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasRepresentationInformation"> 
? 155?
     <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#isRepresentationInformationFor"/> 
    </owl:inverseOf> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isAnalyticInformationFor"> 
    <owl:inverseOf rdf:resource="#hasAnalyticInformation"/> 
    <rdfs:subPropertyOf rdf:resource="#isContextInformationFor"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasAbridgement"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Expression"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="isAbridgementOf"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="hasProvenanceInformation"> 
    <rdfs:domain rdf:resource="#Item"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isOwnedBy"> 
    <rdfs:domain rdf:resource="#Item"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Person"/> 
          <owl:Class rdf:about="#Corporate_Body"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isAlternateOf"> 
    <owl:inverseOf rdf:resource="#hasAlternate"/> 
    <rdfs:range rdf:resource="#Manifestation"/> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isSummaryOf"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf rdf:resource="#hasSummary"/> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasImitation"> 
? 156?
     <owl:inverseOf rdf:resource="#isImitationOf"/> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isAbridgementOf"> 
    <owl:inverseOf rdf:resource="#hasAbridgement"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasReconfiguration"> 
    <rdfs:domain rdf:resource="#Item"/> 
    <rdfs:range rdf:resource="#Item"/> 
    <owl:inverseOf rdf:resource="#isReconfigurationOf"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasAdaptation"> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:about="#isAdaptationOf"/> 
    </owl:inverseOf> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isAdaptationOf"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <owl:inverseOf rdf:resource="#hasAdaptation"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isGamePlayDataFor"> 
    <owl:inverseOf rdf:resource="#hasGamePlayData"/> 
    <rdfs:subPropertyOf rdf:resource="#isContextInformationFor"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#realizes"> 
? 157?
     <owl:inverseOf rdf:resource="#isRealizedThrough"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Work"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isSyntacticRepresentationInformationFor"> 
    <rdfs:subPropertyOf> 
      <owl:ObjectProperty rdf:about="#isRepresentationInformationFor"/> 
    </rdfs:subPropertyOf> 
    <owl:inverseOf rdf:resource="#hasSyntacticRepresentationInformation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isRuntimeContextFor"> 
    <rdfs:subPropertyOf rdf:resource="#isContextInformationFor"/> 
    <owl:inverseOf rdf:resource="#hasRuntimeContext"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasPart"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf rdf:resource="#isPartOf"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isTranslationOf"> 
    <rdfs:range rdf:resource="#Expression"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <owl:inverseOf> 
      <owl:ObjectProperty rdf:ID="hasTranslation"/> 
    </owl:inverseOf> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isRepresentationInformationFor"> 
    <owl:inverseOf rdf:resource="#hasRepresentationInformation"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Item"/> 
          <owl:Class rdf:about="#Manifestation"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:ID="isProducedBy"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Person"/> 
          <owl:Class rdf:about="#Corporate_Body"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isReproductionOf"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
? 158?
         </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf rdf:resource="#hasReproduction"/> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isRevisionOf"> 
    <owl:inverseOf rdf:resource="#hasRevision"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Expression"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#hasTranslation"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="#Expression"/> 
    <owl:inverseOf rdf:resource="#isTranslationOf"/> 
  </owl:ObjectProperty> 
  <owl:ObjectProperty rdf:about="#isSuccessorOf"> 
    <rdfs:range> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:range> 
    <owl:inverseOf rdf:resource="#hasSuccessor"/> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Work"/> 
          <owl:Class rdf:about="#Expression"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:ObjectProperty> 
  <owl:DatatypeProperty rdf:ID="Scheduled_Treatment"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Scale"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Context_for_Work"> 
    <rdfs:domain rdf:resource="#Work"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Special_Characteristic"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Extensibility"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Projection"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Source_for_Acquisition_or_Access_Authorization"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Date_of_Creation"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Expression"/> 
? 159?
           <owl:Class rdf:about="#Work"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Rights_Metadata"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Item"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Work"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Context_for_Expression"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Language"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Condition"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Form_of_Expression"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Extent"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Expression"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Terms_of_Availability"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Revisability"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Technical_Metadata"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Fingerprint"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Intended_Termination"> 
    <rdfs:domain rdf:resource="#Work"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Form_of_Work"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
    <rdfs:domain rdf:resource="#Work"/> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >the class to which the work belongs (e.g., novel, play, poem, essay, biography, 
symphony, concerto, sonata, map, drawing, painting, photograph, etc.)</rdfs:comment> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Use_Restrictions"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Exhibition_History"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Dimensions_of_Carrier"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Date_of_Publication_or_Distribution"> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
? 160?
     <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Extent_of_Carrier"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Critical_Response_to_Expression"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Summarization_of_Content"> 
    <rdfs:domain rdf:resource="#Expression"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Uniform_Title"/> 
  <owl:DatatypeProperty rdf:ID="Provenance_Metadata"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Title"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Expression"/> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Work"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
    <rdfs:comment rdf:datatype="http://www.w3.org/2001/XMLSchema#string" 
    >A word, phrase, or group of characters naming an intellectual or artistic 
creation</rdfs:comment> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Series_Statement"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Intended_Audience"> 
    <rdfs:domain rdf:resource="#Work"/> 
    <rdfs:range rdf:resource="http://www.w3.org/2001/XMLSchema#string"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Edition_Issue_Designation"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Fixity_Information"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Publisher_or_Distributor"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Access_Restrictions"> 
    <rdfs:domain> 
      <owl:Class> 
        <owl:unionOf rdf:parseType="Collection"> 
          <owl:Class rdf:about="#Manifestation"/> 
          <owl:Class rdf:about="#Item"/> 
        </owl:unionOf> 
      </owl:Class> 
    </rdfs:domain> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Identifier"/> 
  <owl:DatatypeProperty rdf:ID="Statement_of_Responsibility"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Form_of_Carrier"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Place_of_Publication_or_Distribution"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Capture_Mode"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Physical_Medium"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
? 161?
   </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Marks_and_Inscriptions"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Fabricator_or_Manufacturer"> 
    <rdfs:domain rdf:resource="#Manifestation"/> 
  </owl:DatatypeProperty> 
  <owl:DatatypeProperty rdf:ID="Treatment_History"> 
    <rdfs:domain rdf:resource="#Item"/> 
  </owl:DatatypeProperty> 
  <rdf:Description> 
    <rdf:first rdf:resource="#Work"/> 
  </rdf:Description> 
</rdf:RDF> 
 
? 162?
 Appendix D ? Multi-Institutional Collaboration: Lessons 
Learned 
 
When partners are slow in providing contributions, enlist a strike force of elite otter 
mercenaries for encouragement. 
 
 
 
 
? 163?
 Appendix E ? Publications 
 
Forthcoming 
 
Kraus, K. ? ?A Counter-Friction to the Machine?: What Game Scholars, Librarians, and 
Archivists Can Learn from Machinimists about User Activism.?  Journal of Visual Culture 
(special Machinima-themed issue guest-edited by the Stanford Humanities Lab), 
forthcoming spring 2011. 
 
Lowood, H. & M. Nitsche. The Machinima Reader. Cambridge, Mass.: MIT Press, forthcoming 
2011. (Includes ?Video Capture: Machinima, Documentation, and the History of Virtual 
Worlds, by Lowood.) 
 
Lowood, H., ?Memento Mundi: Are Virtual Worlds History??  In: Digital Media: 
Interdisciplinary Perspectives on History, Preservation, and Ontology, eds. Megan Winget and William 
Aspray. Scarecrow Press, forthcoming 2011. 
 
Lowood, H., ?Perfect Capture: Three Takes on Replay, Machinima and the History of 
Virtual Worlds.?  Journal of Visual Culture (special Machinima-themed issue guest-edited by 
the Stanford Humanities Lab), forthcoming spring 2011. 
 
McDonough, J. ?Packaging Videogames for Long-Term Preservation: Integrating FRBR and 
the OAIS Reference Model.?  Journal of the American Society for Information Science and Technology. 
Forthcoming. 
 
2010 
 
Kraus, K. (March 2010). Quoted in ?Pac Rat: The Fight to Preserve Old Video Games from 
Bit Rot, Obsolescence, and Cultural Oblivion.? Atlantic Magazine. 
 
Kraus, K. & Donahue, R. (May 3, 2010). ?The Perplexing Task of Archiving Virtual 
Worlds.? Interview. Maryland Morning Show with Sheilah Kast, broadcast on Baltimore?s NPR 
member station, WYPR 88.1 FM. http://mdmorn.wordpress.com/2010/05/03/53102-the-
 perplexing-task-of-archiving-virtual-worlds/ 
?
 Lowood, H. (2010). ?The Future of Virtual Worlds,? (with William Sims Bainbridge, Wayne 
Lutters and Diana Rhoten,? in: Online Worlds: The Convergence of the Real and the Virtual, ed. 
William Sims Bainbridge (London: Springer, 2010): 289-302. 
 
? 164?
 McDonough, J., Kirschenbaum, K., Reside, D., Fraistat, N. & Jerz, D. (Fall 2010). ?Twisty 
Little Passages Almost All Alike: Applying the FRBR Model to a Classic Computer Game.?  
Digital Humanities Quarterly 4(2). Retrieved Sept. 13, 2010 from 
http://www.digitalhumanities.org/dhq/vol/4/2/000089/000089.html. 
 
2009 
 
Kirschenbaum, M. & Farr, E., et al. (October 2009). ?Digital Materiality: Preserving Access 
to Computers as Complete Environments.? iPres 2009. 
 
Kirschenbaum, M., Tabbi, J., Grigar, D., Tata, M.A., Heckman, D., Gibbs, A. & Angel, M. 
(December 2009).  ?E-Ject: On the Ephemeral Nature, Genres, and Criticism of Electronic 
Objects.? Digital Arts and Culture 2009. 
 
Lowood, H., ed. (2009). Before It?s Too Late: A Game Preservation White Paper (written by the Game 
Preservation SIG of the International Game Developers Association). Available at: 
http://wiki.igda.org/Game_Preservation_SIG/White_Paper/Before_It%27s_Too_Late:_A
 _Digital_Game_Preservation_White_Paper. 
 
Lowood, H., McDonough, J., Frick, C. & Renear, A. (2009). ?Digital Curation of 
Humanistics, Multimedia Materials: Lessons Learned and Future Directions? (Panel 
abstract), Proceedings of DigCCurr 2009: Digital Curation ? Practices, Promise & Prospects, ed. Helen 
R. Tibbo, et al. (Chapel Hill: School of Library and information Science, Univ. of North 
Carolina, 2009; distributed by Lulu Press): 42-43. 
 
Lowood, H. (2009). ?Game Counter,? in Fiona Candlin and Raiford Guins (Eds.),The Object 
Reader. Abingdon, Eng., and New York: Routledge: 466-69. 
 
Lowood, H. (2009). ?Players are Artists,? in Debora Ferrari and Luca Traini (Eds.), The Art 
of Games: Nuove Frontiere tra Gioco e Bellezza. Aosta: Regione Auonoma Valle d?Aosta: 190-93. 
Also in Italian trans. as ?I giocatore come artisti,? pp. 194-97. 
 
Lowood, H. (Summer 2009). ?Putting Politics Into Play: Three Recent Books on Virtual 
Worlds,? American Journal of Play 2(1). 
 
Lowood, H. (July-Sept. 2009). ?Video Games in Computer Space: The Complex History of 
Pong,? in IEEE Annals in the History of Computing: 5-19. 
 
Lowood, H. (2009). ?Warcraft Adventures: Texts, Replay and Machinima in a Game-Based 
Story World,? in Pat Harrigan and Noah Wardrip-Fruin (Eds.), Third Person: Authoring and 
Exploring Vast Narratives. Cambridge: MIT Press: 407-27. 
 
Lowood, H. guest editor, July-Sept. 2009 issue of IEEE Annals of the History of Computing. 
 
McDonough, J. (2009). ?How will we preserve virtual worlds?? in Bart De Nil & Jeroen 
Walterus (Eds.), Nieuwe Perspectieven voor Digitaal Erfgoed. Brussel: Pharo Publishing. 
 
McDonough, J., Kirschenbaum, M., Reside, D., Freistat, N., Jerz, D., Lowood, H., Kraus, 
? 165?
 K., Donahue, R. & Winget, M. (2009). ?Preserving Virtual Worlds: Models & Community? 
(panel & paper abstracts). In Digital Humanities 2009. Conference Abstracts. University of Maryland, 
College Park. June 22-25, 2009, pp. 22-29. Available at: http://www.mith2.umd.edu/dh09/wp-
 content/uploads/dh09_conferencepreceedings_final.pdf. 
 
2008 
 
Lowood, H. (2008). ?La cultura del replay. Performance, spettatorialit?, gameplay,? in Schermi 
interattivi. Il cinema nei videogiochi, ed. Matteo Bittanti. Rome: Meltimi: 69-94. 
 
Lowood, H. (Spring 2008). ?Game Capture: The Machinima Archive and the History of 
Digital Games,? in Mediascape: Journal of Cinema and Media Studies. 
 
Lowood, H. (2008). ?Replay Culture: Performance and Spectatorship in Gameplay,? in 
Carlos A. Scolari (Ed.), L?homo videoludens: Videojocs, textualitat i narrativa interactiva. Vic: Eumo 
Editorial: 167-87. 
? 166?
 Appendix F ? Museogames Exhibit at the Mus?e des 
arts et m?tiers 
Implications for Collecting Repositories 
 
The Museogames exhibit?on display from 22 June-7 November 2010 at the Mus?e des Arts 
et M?tiers in Paris?showcases the history of videogames from the classic arcade games of 
the 1980s through the sixth-generation console games that became available at the turn of 
the 21st century. Filling three rooms of the museum?s lower level, the exhibit includes home 
consoles, arcade cabinets, peripherals, miscellaneous game culture documents and artifacts, 
and recorded interviews with game developers and scholars. The exhibit?s main attraction is 
the game table in the second room, where visitors are invited to sit down and experience 
retro games and platforms spanning nearly three decades, from the Magnavox Odyssey, the 
first home gaming console, to Sony?s PlayStation 2, the world?s best-selling console. Real-
 time video capture of visitors playing Pac-Man, Super Mario Bros, and other vintage games 
is projected onto the walls of the exhibition space. 
   
Left: A view of the game table. Right: Magnavox Odyssey game console. Photos courtesy of Matthew 
Kirschenbaum. July 2010. 
 
For collecting institutions such as the Library of Congress, Museogames serves as an object 
lesson in managing the underlying tensions between the preservation and access functions of 
archives, particularly where interactive media are concerned. While many of the works in the 
exhibition are drawn from the museum?s permanent collections, the Mus?e des Arts et 
M?tiers also partnered with MO5.com, a non-profit French association dedicated to the 
preservation and public dissemination of digital culture and videogames. Founded in 1996, 
the association is comprised of a diverse set of stakeholders, including private collectors, 
hobbyists, journalists, historians, researchers, and gamers. The organization boasts a 
collection of some 30,000 objects, which run the gamut from antiquated computers, 
peripheral devices, and other hardware to software, source code, technical manuals, and rare 
press kits.  
 
In its information leaflets the museum takes pains to distinguish its video game collection 
from that of MO5.com along two axes: selection and use. 
? 167?
 ? Selec t ion  ? ?Toward the end of the 90s, a few private collectors took it upon 
themselves to safeguard videogame heritage. Their motivation was essentially 
sentimental and the objective was to keep everything. The approach adopted by the 
museum is somewhat different. All pieces that join its collections represent a 
landmark in the history of videogames. The objects acquired by the Mus?e des Arts 
et M?tiers bear witness to the technical developments of their time; they have played 
an important role in our economic, social, and cultural history and represent an 
essential link between the inventions that preceded them and those that were to 
come after them. This policy involves selectivity when acquiring pieces for the museum? (?Two 
Collections?; emphasis added). 
? Use  ? ?The objective is [for the museum] to preserve [its] collections for future 
generations. For this reason, unlike the pieces belonging to private collections and 
which do not have this perenniality requirement, handling of the video games is 
strictly limited. This difference is illustrated throughout this exhibition with the 
MO5.com collection featuring videogames that can be played, and the collection that 
belongs to the Mus?e des arts et m?tiers, which is protected for preservation purposes? 
(?Two Collections?; emphases added).  
 
These distinctions are visually reinforced throughout the exhibit, with game consoles from 
the museum?s collection sequestered behind wire cages and those from the MO5.com 
collection set up in the open as game stations.11  One obvious consequence of this functional 
division of collections is that the role of the museum visitor changes from that of passive 
observer to active participant over the course of the exhibition. Likewise, the status of the 
game artifact itself shifts from one room to the next?from antiquarian relic to source of 
hands-on entertainment. 
   
 
Contrasting experiences: on the left Kraus views the museum?s collection of game consoles displayed within 
wireframe cages; on the right, kids play with classic console games lined up in the second gallery room. Photos 
courtesy of Matthew Kirschenbaum. July 2010. 
????????????????????????????????????????????????????????
 11?While we cannot categorically state that every item behind the cages is drawn from the museum?s 
permanent collection rather than MO5.com, this inference is supported by the museum?s 
information leaflet, quoted above.  All the games on the game table are from the MO5.com 
collection.?
? 168?
  
The example of Museogames prompts two observations with special relevance to collecting 
institutions: 
 
? In an age of interactive media, archivists and curators need a richer typology of user 
behaviors with which to work. Access models to cultural heritage have been overly 
determined by the stewardship of paper documents, with the researcher variously 
understood as reader, transcriber, or viewer of content. In the case of videogames 
and variable media art, however, a more accurate depiction of the user might be that 
of player or interactor. 
? In a user-centered milieu, it is especially important that collecting institutions solicit 
the input and consider the needs and expectations of the user community when 
developing selection and appraisal models. These expectations need to be 
thoughtfully and creatively balanced against other archival responsibilities, notably 
preservation. 
 
With these observations in mind, we propose the following ways MO5.com might serve as a 
resource and model for collecting institutions: 
 
? An analysis of the scope and diversity of the MO5.com holdings would allow a 
repository to get a purchase on the publishing, documentary and artifactual universe 
of videogame production and reception: from concept art and storyboards to 
marketing materials to vintage software and hardware (and everything in between). 
Determining the full spectrum of products of videogame culture would pave the way 
for the archive to adopt a systematic approach to selection and appraisal rather than 
be disadvantaged by a partial view of the collecting landscape. 
? The Museogames exhibition demonstrates the value of establishing strong 
partnerships with private collectors in order to design interactive game experiences 
for the public without compromising the preservation mission of the collecting 
institution. In the absence of a North American counterpart to MO5.com, the 
Library of Congress might consider facilitating the formation of one, perhaps under 
the auspices of (or in cooperation with) the IGDA Game Preservation SIG. 
? Finally, a collecting institution such as the Library of Congress might consider 
internalizing some of the practices of MO5.com by adopting a dual-track strategy for 
collection development. This strategy would entail acquiring and maintaining both 
preservation and access copies of software and hardware?such as early 
microcomputers and home game consoles?rather than following the example of the 
Mus?e des Arts et M?tiers in assuming a noncustodial role with respect to playable 
versions of games.  Such a collection model has been adopted by the International 
Center for the History of Electronic Games (ICHEG) at the Strong National 
Museum of Play.  As described in the ?Analysis of Hardware Preservation? section 
of this report, the ICHEG collects multiple copies of the same artifact, allowing the 
museum to source hardware components for installations and exhibits from 
duplicates in its own collection.  
? 169?
  
Bibliography 
 
?Two Collections Brought Together for Museogames,? Paris: Mus?e des Arts et M?tiers, 22 
June-7 November 2010. Print. 
 
Museogames exhibition, Paris: Mus?e des Arts et M?tiers, 22 June-7 November 2010. 
http://museogames.com/ 
? 170?
 Appendix G ? Second Life Deed of  Gift 
 
LICENSE AGREEMENT 
This is a license agreement between DONOR NAME (Donor) and The Board of Trustees 
of Leland Stanford Junior University (Stanford) regarding the preservation of virtual 
property in the Stanford University Libraries' Preserving Virtual Worlds Second Life Archive 
(Archive).  
Donor is the creator and copyright owner of material developed for Second Life, more 
particularly described to the right of the webpage. 
Donor hereby grants Stanford a non-exclusive, royalty-free, irrevocable, worldwide license 
to preserve the Object(s), and to display and distribute the Object(s) for educational and/or 
not-for-profit purposes in all media now known or hereafter created, including but not 
limited to print, audio, electronic, video, optical disk, photographic, digital, and film, subject 
to the following.  
1. Donor warrants and represents that s/he is the exclusive creator of and sole 
rightsholder for the Object(s) and has the full authority to enter into this Agreement. 
Donor agrees to indemnify the Board of Trustees of the Leland Stanford Junior 
University against any claims (including attorneys? fees and court costs) made against 
the University relating to this Agreement. 
2. Stanford is licensed to use and manipulate the Object(s) in all ways necessary to 
preserve the work, including changing the format of the Object(s).  
3. Stanford may display and distribute the Object(s) on its own website, and may also 
share the Object(s) with partner organizations for purposes of display and 
distribution.  
4. Except as set forth herein, copyrights held by the Donor in the Object(s) will remain 
with the Donor, including the rights to sell, assign or otherwise exploit or dispose of 
copyrights, subject to this irrevocable license agreement. Donor may commercially 
exploit the copyrighted materials by any means now known or hereafter known, 
including but not limited to print, motion picture, audio, audio-visual, visual, 
electronic, film, internet, or similar means of exploitation and dissemination.  
5. Stanford University Libraries is under no obligation to accept the Object(s) or to 
maintain the Object(s) if accepted. Stanford reserves the right to refuse to accession 
Objects, and may delete Objects from the Archive at any time for any reason. 
Stanford is under no obligation to maintain the Object(s) or the Archive. Stanford is 
in no way liable for third party use of the Object(s).  
6. Stanford will provide access to the Object(s) in accordance with its own policies and 
procedures. The Object(s) will not become accessible to researchers until archival 
processing has been completed, and the time required for such processing is at 
Stanford's discretion.  
? 171?
 7. Optional: Stanford agrees to restrict access to Object(s) as requested by the Donor 
below, and will not make the Object(s) available on publicly accessible media until 
that time. Stanford will, however, preserve the Object(s) in the Archive during that 
time.  
8. This Agreement will be subject to the laws of California and any action arising out of 
this Agreement will be exclusively heard in a court of competent jurisdiction in Santa 
Clara County, California.  
? 172?
 Appendix H ? Gaming Websites Identified in PVW 
Project 
 
? = Archive-It seed has been made 
  
Virtual Words, MMOs, Game Worlds in general 
  
? ?http://terranova.blogs.com/  
          Terra Nova : Collaborative blog on online communities and worlds. 
? ?http://my.binhost.com/lists/listinfo/mud-dev2   
          MUD-Dev2 Info Page : discussion list of MUD system design, development and 
implementation  
? ?http://my.binhost.com/pipermail/mud-dev2/  
          MUD-Dev2 archives 
? ?http://www.raphkoster.com/2007/02/02/full-mud-dev-archive-for-download/   
          Raph's Website : MUD-Dev1 archive can be downloaded here  
? ?http://mmorpg.qj.net/ 
          QJ.net : MMORPG blog "24/7 Coverage of Your Favorite MMORPG News"  
? ?http://www.fudco.com/habitat/ 
          Habitat Chronicles : Chip Morningstar and Randy Farmer 
? ?http://www.dsgames.net/index.htm 
          DSGames history, including Qlink.net and Habitat screenshots 
? ?http://www.guildcafe.com/ 
          Guild Cafe a social network site with blogs and forums, supporting player guilds and 
clans of many MMORPGs and online games 
  
  
 Second Life: 
  
? ?http://secondlife.com 
        Second Life : official website  
? http://mariogerosa.blogspot.com/ 
        Played in Italy: Curation of History of Second Life  
? http://second-life.com 
        Second Life : tips, tricks, news, and everything you want to know  
? http://blog.secondlife.com/   
          Official Linden Blog 
? http://secondlifeherald.com 
          The Second Life Herald : presents the game?s events, news, and author?s tips  
? http://secondlife.reuters.com 
          Reuters Second Life News Portal: company portal  
? http://lindenlab.com 
        Linden Lab : official website of Linden Lab, makers of Second Life  
? http://secondlifeinsider.com 
? 173?
           Second Life Insider : weblog with news and tips about both the players and the 
game  
? http://valleywag.com/tech/second-life/ 
          Valleywag : gossip rag on Second Life  
? http://secondlifegrid.net/programs  
          Second Life Grid : open source platform used for 3D virtual world development in  
        business,education,and nonprofit organizations 
? http://forums.secondlife.com 
          Second Life Forums 
? http://sundancechannel.com/secondlife 
          Sundance Channel: Blogs: Second Life : virtual screening room  
? http://edition.cnn.com/2007/TECH/11/12/second.life.irpt/ 
          CNN Enters the Virtual World of Second Life : CNN to open i-report hub in 
Second Life  
? http://secondlifevideo.com 
        Second Life Video  
? http://wiki.secondlife.com/wiki/Video_Tutorials   
          Second Life Video Tutorials : list of video tutorials  
? http://crn.com/it-channel/205101362 
          Accenture Scientist Predicts Death of Second Life 
? http://slurl.com   
          SLurl : location-based linking in Second Life  
? http://somethingawful.com/d/second-life-safari/   
          Second Life Safari 
? http://slprofiles.com 
          Second Life Profiles : online community for residents to meet  
? http://getafirstlife.com 
          Get A First Life: one page satire of Second Life  
? http://abc.net.au/services/secondlife   
          ABC Online: Second Life : ABC island in Second Life  
? http://dell.com/html/global/topics/sl/index.html 
          Dell Island In Second Life 
? http://wiki.secondlife.com/wiki/Main_Page 
        Second Life Wiki 
? http://slexchange.com 
          SL Exchange : Second Life commerce  
? ?http://infoisland.org 
          Second Life Library 2.0 
? http://usd.auctions.secondlife.com 
          Second Life Land Auctions 
? http://slcc2007.wordpress.com 
          The Official SLCC Blog 
? http://reperes-secondlife.com 
          Reperes : first market research institute on Second Life  
? http://alpha.cbs.com/primetime/csi_ny/second_life/ 
          Virtual CSI NY : CSI NY on Second Life  
? http://secondlla.googlepages.com 
? 174?
           Second Life Liberation Army 
? http://slnn.com 
          Second Life News Network 
? http://mmorpg.qj.net/category/Second-Life/cid/850 
          MMORPG News : MMORPG news on Second Life  
? http://roughtype.com/archives/2006/12/avatars_consume.php 
          Rough Type : Nicholas Carr?s blog  
? http://secondlife.vodafone.com 
          Vodafone  
? http://freshtakes.typepad.com/sl_communicators   
          Business Communicators on Second Life 
? http://slgames.wordpress.com/ 
          Second Life Games : blog  
? http://sleducation.wikispaces.com 
          Second Life in Education : wiki  
? http://teen.secondlife.com/whatis   
          Teen Second Life 
? http://slvid.com 
          SLVid : share Second Life videos  
? http://nonprofitcommons.org 
          Nonprofits in Second Life  
? http://ibm.com/developerworks/spaces/secondlife 
          Second Life 
? http://flickr.com/groups/secondlife 
          Flickr: Second Life : upload pictures from Second Life  
? http://2lifeblog.com 
          Second Life Scripts 
? http://massively.com/category/second-life 
          Posts From The Second Life Category At Massively : the daily news about 
MMOs  
? http://crunchbase.com/company/secondlife 
          Crunchbase Company Profile 
? http://secondlife.intellagirl.com 
        Second Life Education Research : blog   
? http://sldrama.com 
        SecondLife Drama 
? http://web.ics.purdue.edu/~mpepper/slbib 
          Second Life Annotated Bibliography 
? http://secondlifelibrary.blogspot.com 
          Second Life library 2.0 
? http://slpodcast.com 
          Second Life Podcast : the podcast and blog about Second Life  
? http://sl-art-news.blogspot.com 
        Second Life Art News 
? http://secondlifenotes.com 
          Second Life Notes: podcast dedicated to exploring the music, art, and culture of 
Second Life  
? 175?
 ? http://jira.secondlife.com/secure/Dashboard.jspa 
          Second Life Issues  
? http://secondlife.blogs.cnn.com 
        Second Life I-Reports : news of a virtual world  
? http://slfuturesalon.blogs.com/ 
        Second Life Future Salon 
? http://secondlifeupdate.com 
        Second Life Update 
? http://slcn.tv 
        Second Life Cable Network 
? http://second411.com 
         Second 411 : Second Life search engine   
? http://iste.org/Content/NavigationMenu/Membership/Member_Networking/IST
 E_Second_Life.htm 
          Iste Second Life: Iste island on Second Life  
? http://foureyedmonsters.com/secondlife 
          Four Eyed Monsters : blog  
? http://taotakashi.wordpress.com 
Tao?s Thoughts On Second Life : blog  
? http://molotovalva.com 
        My Second Life 
? http://bloghud.com 
        BlogHUD : blogging system for the residents of Second Life  
? http://rootscamp.org/RootsCampSL 
        Roots Camp 
? http://secondlife.crowneplaza.com 
        Crowne Plaza?s Second Life Meeting Room Reservation 
? http://apps.facebook.com/second-life 
          Second Life Facebook Application 
? http://slballet.org 
          Second Life Ballet 
? http://lists.secondlife.com/cgi-bin/mailman/listinfo 
        Second Life Mailing Lists 
? http://slcreativity.org/blog 
          SL Creativity 
? http://slleftunity.blogspot.com 
        Second Life Left Unity 
? http://secondliferesearch.blogspot.com 
        Second Life Research : blog about research on Second Life  
? http://slambling.blogspot.com 
          Ambling In Second Life : blog  
? http://secondlifehowto.com 
          Second Life HowTo: free guide to content creation in Second Life  
? http://virtualaloft.com 
          Virtual Aloft 
? http://sltree.com 
          Second Life Tree : the metaverse directory  
? 176?
 ? http://secondlife.wikia.com 
          Second Life Wikia 
? http://www.myfirstsecondlife.com 
          My First Second Life 
? http://secondlife.podcast.com 
        Second Life Podcasts 
? http://secondlifeproject.com 
          Second Life Project 
? http://bethssecondlife.blogspot.com 
        Beth?s Second Life : a teacher's blog  
? http://libsecondlife.org 
          Libsecondlife : project trying to understand the game from a technical aspect  
? http://lindenlifestyles.com 
          Linden Lifestyles: unofficial Second Life fashion and shopping blog  
? http://mynameiskate.typepad.com/secondlife 
Kate?s So-Called Second Life : blog  
? http://slbusinesscommunicators.pbwiki.com 
        Second Life Business Communicators Wiki : a collaborative resource for anyone 
interested in business and communications applications of Second Life  
? http://sltutorials.net 
          SLTutorials.net 
? http://blog.secondstyle.com 
          Second Style Fashionista : fashionista blog for Second Life  
? http://mercedes-benz-secondlife-infos.com  
          Mercedez-Benz : Mercedes-Benz in Second Life  
? http://siobhantaylor.wordpress.com 
Sio?s Second Life : blog 
? http://nwn.blogs.com/nwn/2007/12/second-life-as.html 
          New World Notes : Wagner James Au?s blog  
? http://sweden.se/templates/cs/CommonPage____18195.aspx  
          Second Life: The House of Sweden : the official gateway to Sweden  
? http://storyofmysecondlife.com 
          The Story of My ?Second Life? : educator?s grant-funded exploration into Second 
Life  
? http://slfashionpolice.wordpress.com 
          Second Life Fashion Police : fashion faux pas of Second Life  
? http://slhandbook.com 
        Second Life handbook : guide to people, places, and things in Second Life 
? http://esl-secondlife.blogspot.com 
          Secondlifeenglish.com Blog 
? http://slsolutions.org 
          Second Life Solutions : home of the metaverse stock exchange  
? http://slsailing.org 
        Second Life Sailing Federation 
? http://aeneaideas.wordpress.com 
        Aenea?s Second Life : blog 
? http://caterin.wordpress.com 
? 177?
           Girl Meets Second Life : Caterin?s Second Life blog  
? http://www.secondlifeize.com 
          Second Life Linden Dollars 
? http://secondlifesearch.ourtoolbar.com 
          Second Life Search Toolbar 
? http://kellysecondlife.com/eprise/main/web/us/customers/secondlife/index.html 
        Kelly?s second life 
? http://secondlifeenglish.com 
        Do You Speak SLEnglish? 
? http://slpulse.com 
          Second Life Pulse : second life press releases, announcements, and events  
? http://gridgrind.com 
          The Second Life Grid Grind : Second Life news, info, gossip and gab  
? http://secondstyle.com 
          Second Style Magazine 
? http://secondlife.blogs.com/babbage/2005/08/second_life_in_.html 
          The Creation Engine: Second Life in Mono : Babbage?s blog  
? http://ausslers.com 
          AusSLers : Australian Second Life Educators and Researchers  
? http://sl.nmc.org 
        NMC Campus Observer  
? http://slrecord.typepad.com 
        Second Life Record : blog 
? http://fizzysecondlife.blogspot.com 
        Fizzy?s Second Life 
? http://nicolaescher.com 
          Nicola Escher : virtual world fashion designer 
? http://secondlifenow.com 
          Mr. P?s : bringing Second Life to life 
? http://secondlifetraveler.com 
        Second Life Traveler 
? http://scottsecondlife.blogspot.com 
          Oh Second Life : blog 
? http://sluniverse.com/pics 
          SLUniverse 
? http://sl-sexualhealth.org.uk 
          A Sexual Health SIM in Second Life 
? http://sltweets.com 
          Second Life Twitter HUD Plus : metaverse to real life gateway  
? http://sldnug.net 
          Second Life .Net Developers User Group: a microsoft.net user community in 
Second Life 
? http://www.metaversemessenger.com/ 
          Metaverse Messenger: an SL in-world publication; Tagline: ?A real newspaper for a 
virtual world? 
  
  
? 178?
 World of Warcraft: 
  
? ?http://worldofwarcraft.com  
          World of Warcraft Community Site : official site  
? http://wow-europe.com 
          World of Warcraft Europe 
? http://wow.allakhazam.com 
Allakhazam?s magical realm  
? http://bestofwarcraft.com   
          The Best of Warcraft 
? http://forums.worldofwarcraft.com 
          WOW Forums 
? http://thottbot.com 
          Thottbot : World of Warcraft database  
? http://worldofwar.net 
          World of Warcraft the Unofficial Site : offers news, information, movies, screen 
shots, forums, journals, live chats, and item database  
? http://blizzard.com 
          Blizzard Entertainment 
? http://wow.warcry.com 
          World of Warcraft Warcry 
? http://worldofwarcraft.filefront.com 
          World of Warcraft Downloads: World of Warcraft files for download 
? http://mapwow.com   
          World of Warcraft Maps 
? http://wowwiki.com 
        WoWWiki : guide to the World of Warcraft  
? http://wowvault.ign.com 
          World of Warcraft Vault : World of Warcraft news, trailers, screeenshots, previews, 
reviews, guides 
? http://wow.incgamers.com  
        World of Warcraft the Unofficial Site 
? http://wowguru.com 
          World of Warcraft Guru : database 
? http://wowhead.com 
          Wowhead : database website for World of Warcraft  
? http://wow.qj.net 
          QuickJump : World of Warcraft news, announcements, patches, downloads, and videos  
? http://wowarmory.com 
          The World of Warcraft Armory : vast searchable database of information for World 
of Warcraft  
? http://wowinsider.com 
          Wow Insider 
? http://worldofwconline.com 
          WoW World of Warcraft Online 
? ?http://warcraftmovies.com 
          World of Warcraft Movies: large database of World of Warcraft movies  
? 179?
 ? http://almostgaming.com/wow 
          World of Warcraft Strategy Guides 
? http://warcraft.trei.ro/world_of_warcraft.php 
          Warcraft Trei Ro 
? http://wowdetox.com  
          WOW Detox: detox center for wow addiction  
? http://war3world.com 
          War3World : contains downloads, replays, maps, and wallpapers  
? http://wow.curse.com 
          World of Warcraft AddOns, Downloads : World of Warcraft news, articles, files, 
screenshots, and videos  
? http://wow.gameamp.com 
          World of Warcraft GameAmp: maps, gold, guides, quests, downloads, guilds, demos, 
and videos  
? http://warcraftrpg.com 
          World of Warcraft the Roleplaying Game 
? http://gamesites200.com/wow 
          World of Warcraft Top 200 : top 200 WOW sites as ranked by users  
? http://wowui.incgamers.com   
          WoWUI @ IncGamers : World of Warcraft mods, addons, and more  
? http://wow.stratics.com 
          World of Warcraft Stratics: news and information coverage for World of Warcraft  
? http://askapadwe.com 
          World of Warcraft Help 
? http://worldofwarcraftguru.blogspot.com 
          World of Warcraft : blog 
? http://worldofwarcraftempire.com 
          WoW Empire : information and resources for World of Warcraft  
? http://wow-europe.com/en/burningcrusade 
          World of Warcraft: The Burning Crusade 
? http://warcrafthelp.com 
          World of Warcraft Community: discussion forums for the World of Warcraft gaming 
community  
? http://wowchron.com 
          World of Warcraft Chronicles Podcasts and News 
? http://freewarcraftguides.com 
          World of Warcraft Guides 
? http://stormscale.org 
          Stormscale : news about World of Warcraft from official websites and boards  
? http://wowforum.com 
          World of Warcraft Forum 
? http://worldofwarcraftboard.com 
          World of Warcraft Board: discussion forum powered by vBulletin  
? http://forum.igsky.com/forum-3-1.html 
          World of Warcraft IGSky Game Forum 
? http://blizzard.co.uk/wow/forums.shtml 
          World of Warcraft Forum 
? 180?
 ? http://goblinworkshop.com 
          The Goblin Workshop: forum  
? http://mmotricks.com/forums/world-of-warcraft 
          Massive Multiplayer Online Gaming Forum 
? http://graffe.com/forums/forumdisplay.php?f=53 
          Graffe Forums 
? http://forum.ragezone.com 
          MMORPG development forum 
? http://ubuntuforums.org/showthread.php?t=120615 
          World of Warcraft with Wine : forum 
? http://broadbandreports.com/forum/wowetc 
          World of Warcraft Forum 
? http://moncom.net/moncomwow.asp 
          World of Warcraft Forum 
? http://forums.gameaxis.com/forumdisplay.php?f=236 
          The Unofficial World of Warcraft Forum 
? http://worldofwarcraft-game.com 
          WoW World of Warcraft : description and links 
? http://uberwow.com 
          Uber WoW Fansite : uber WOW fansite; WOW forum; WOW guides  
? http://mmoverload.com 
          MMOverload : strategy website 
? http://abcwow.net  
          ABC World of Warcraft  
? http://wowvillage.com 
          World of Warcraft Village : "all things world of warcraft"  
? http://leeroyjenkins.net 
          Leeroy Jenkins 
? http://wowfix.com 
          WoWFix 
? http://www.almostgaming.com/wow  
          World of Warcraft Strategy Guides 
? http://wowstatus.com 
          World of Warcraft Private Server Status 
? http://exploitsrus.com/wow.html  
          Exploits R Us : wow cheats, gold, and hacks 
? http://worldofwarcraft100.com 
          World of Warcraft 100 : source for WOW sites  
? http://worldofwarcraft.bz 
          World of Warcraft BZ : free private EMU wow server 
? http://frostbolt.com 
          Frostbolt : WOW blog  
? http://worldofwarcraft.areblogs.com 
          World of Warcraft Addict's Blog 
? http://hogit.org 
Hogit?s Story : blog  
? http://wowgrrl.com 
? 181?
           World of Warcraft Blog by WoWGrrl 
? http://casualwow.blogspot.com 
          Casual WoW : blog  
? http://world-of-warcraft-gold.com/blog  
          World of Warcraft News Blog  
? http://1up.com/do/my1Up?publicUserId=5829624 
          Legendary Thread : blog  
? http://lamthara.blogspot.com 
          Lamthara : blog  
? http://marama-wow-corner.blogspot.com 
Marama?s WoW Corner : blog  
? http://the-wow-blog.com 
          The World of Warcraft Blog 
? http://aeonity.com/wow 
          Alliance World of Warcraft Blog 
? http://wowguy.wordpress.com 
          World of Warcraft Guy : blog  
? http://worldofwarcraftblog.wordpress.com  
          World of Warcraft Blog 
? http://wowdaily.info 
          WoWDaily 
? http://worldofwarcraftblog.info 
          World of Warcraft Blog : tips on gold making and power leveling  
? http://wowgold.shoutpost.com 
          ShoutPost : blog 
? http://wow-gold-reviews.blogspot.com 
          World of Warcraft WoW : blog 
? http://noggaddicts.com 
          Noggaddicts : blog 
? http://wow-tips.blogspot.com 
          World of Warcraft Tips, Guides, and Exploits 
? http://warcraftbot.blogspot.com 
          World of Warcraft Bot 
? http://aeigelus.com 
          Adventures of Aeigelus : blog  
? http://warcraft.topplayers.com 
          World of Warcraft Blog 
? http://squidoo.com/worldofwarcraftblogs 
          List of World of Warcraft Blogs on Squidoo 
? http://wow.mmodb.com 
          World of Warcraft Database 
? http://wow.gamepressure.com 
          GamePressure Network: database- maps, locations, quests, items, etc  
? http://warcraftworlds.net  
          World of Warcraft Resource Site 
? http://hiddenstuff.com  
          World of Warcraft Guides Database  
? 182?
 ? http://wowecon.com 
          World of Warcraft Auction House Price Database 
? http://worldofwconline.com/database/items 
          World of Warcraft Items Database 
? http://warcraftpets.com/wow.pets/index.asp 
          Warcraft Small Pets : index of all non-combat pets  
? http://wow-item.info 
          Wow Item Compare Database 
? http://wowdigger.com 
        WoWDigger : database  
? http://gankbang.com/a/ 
        Gankbang - Armory data search and ranking  
? http://www.wowjutsu.com/world/ 
        WoWJutsu -- Guild data and ranking 
? http://www.blogazeroth.com/ 
        Blog Azeroth -- forum for WoW bloggers 
? http://www.killerguides.com/ 
        Killer Guides -- strategy guides for MMOs, incl. WoW (commercial site) 
? http://elitistjerks.com/ 
        Elitist  Jerks -- forums and strategy for WoW (guild site) 
  
  
Lord of the Rings Online: 
? http://lotro.com 
          The Lord of the Rings Online : official community site  
? http://lotro.allakhazam.com 
          Allakhazam's Magical Realm : database with quest walkthroughs, item lists, a full 
bestiary, spells, etc.  
? http://lotrovault.ign.com 
          The Lord of the Rings Online Vault : Lord of the Rings Online news, trailers, 
screenshots, previews, etc.  
? http://lotrmmorpg.com 
          LOTR MMORPG : community site; features forums, screenshots, updated news, 
guild forums, user photos  
? http://lotro.warcry.com 
          Lord of the Rings Online WarCry 
? http://lordoftheringsonline.net 
          The Madhouse Tavern : European fansite of the MMORPG LOTR Online  
? http://lotro.mmodb.com  
          Lord of the Rings Online Database  
? http://gamespot.com/pc/rpg/middleearthonline/index.html 
          Gamespot : news, previews, images, videos, links, and a forum  
? http://lotrtalk.com 
          Lord of the Rings Online : LOTR Online Community Talk  
? http://lotronline.com 
          Lord of the Rings Online : information database fan site  
? http://forums.lotro.com 
? 183?
           Lord of the Rings Online Forums 
? http://lordoftheringsonlinenews.blogspot.com 
          Lord of the Rings Online : blog 
? http://lotro.curse.com 
          Curse : LOTR Online news, articles, files, screenshots, and videos  
? http://lordoftherings.filefront.com 
          Lord of the Rings Files : LOTR downloads  
? http://lotro.us 
          Lord of the Rings Online Fansite : information on classes, races, and creatures  
? http://guides.ign.com/guides/12112 
          Lord of the Rings Online Guide  
? http://lord-of-the-rings.org/lotr_game.html  
          The Lord of the Rings Online Game 
? http://massively.com/category/lord-of-the-rings-online  
          Posts from the Lord of the Rings Online Category at Massively  
? http://lotronlineguides.com   
          Lord of the Rings Online Leveling & More Guide 
? http://mmorpg.qj.net/category/The-Lord-of-the-Rings-Online-Shadows-of-
 Angmar/cid/1905 
          QuickJump : MMORPG blog  
? http://lordoftherings.gameamp.com 
          GameAmp : betas, release dates, maps, guides races, classes and PVP information  
? http://games.slashdot.org/games/07/06/01/0816244.shtml 
          Slashdot : LOTR Online review  
? http://community.codemasters.com/forum/forumdisplay.php?f=417 
          Codemasters Forum 
? http://lotrovideo.blogspot.com 
          Lord of the Rings Online Videos 
? http://lorebook.lotro.com 
          Lord of the Rings Online Lorebook 
? http://gamesites200.com/lotro 
          Lord of the Rings Online Top 200 : top 200 sites for LOTR Online  
? http://lotro.turbine.com/index.php?page_id=73&siid=3 
          Lord of the Rings Online Gameplay 
? http://arda-online.com/map/ 
          Lord of the Rings Online Game Database on Google Maps 
? http://lotro.tentonhammer.com 
          Lord of the Rings Online @ TenTon Hammer : LOTRO community site  
? http://mapslotro.com 
          Lord of the Rings Online Maps 
? http://lotro-wiki.com 
          Lord of the Rings Online Wiki 
? http://lotroedge.com 
          LOTRO Edge : lotr online guide 
? http://lotrofaces.com 
          Character Faces 
? http://lotrointerface.com 
? 184?
           LotRO Interface : comprehensive site dealing with anything Interface related in 
LOTRO  
? http://lotrolife.com 
          Lord of the Rings Online MMORPG web comic 
? http://lotrorphaven.com  
          Lord of the Rings Online Role Player?s Haven 
? http://lotro.stratics.com 
          Lord of the Rings Online Community Resource 
? http://lotro.wikia.com/wiki/Main_Page 
          Lord of the Rings Online Wikia Wiki 
? http://thebrasse.com/lotro 
          The Brasse 
? http://weathertopradio.com 
          Weathertop Radio : podcasts with news and updates on LOTRO  
? http://lotroalliance.com 
          Lord of the Rings Online Alliance 
? http://gamertales.com/lotrotales.php 
          Gamer Tales : collection of LOTRO gaming tales  
? http://kismetbp.com 
          Kismet Lord of the Rings Online : news and fansite  
? http://guildcafe.com/LordOfTheRingsOnline.php 
          Guildcafe : kinships, guilds, profiles, screenshots  
? http://lotrocraft.com 
          Lord of the Rings Online Crafters of Middle Earth : fansite for crafters  
? http://lotroguilds.net 
          LOTRO Guilds : guild listing, screenshots, videos, interviews  
? http://lotronotes.com 
          LOTRO Notes : searchable tips database  
? http://lotropolis.com 
          LOTROpolis : LOTRO community  
? http://middleearthcenter.com 
          Middle Earth Center 
? http://virginworlds.com/podcast.php?show=3&ep=15 
          Ringcast : weekly podcast covering LOTRO  
? http://lickanear.com/corcoffee.html 
          Lick An Ear : LOTRO related fansite  
? http://underthebanner.com 
          Under the Banner : information blog  
? http://visionsofthering.com 
          Visions of the Ring: website for LOTRO suggestions, concepts, ideas  
? http://arda-online.com 
          Arda Online : LOTRO community  
? http://lotromovies.net 
          Lord of the Rings Online Movies Database 
? http://onlinelotr.com 
          Lord of the Rings Online : database of quests, deeds, items  
? http://lotro-game.com 
? 185?
           LOTR Online 
? http://www.lotro-game.com/resources/databases 
          Lord of the Rings Online Databases 
? http://lotrotraits.com 
          LotRO Traits : filterable trait database  
? http://lotrocrafters.com 
          Lord of the Rings Online : crafters forum  
? http://pooh.cz/lotro  
          Lord of the Rings Online : fansite and blog  
? http://lotrovideo.vsocial.com 
          Lord of the Rings Online Videos 
? http://annonamarth.eu 
          Annon Amarth 
? http://lotroinsider.blogspot.com 
          Lord of the Rings Online Insider : blog  
? http://blogtoplist.com/rss/lord-of-the-rings-online.html 
          Blog Toplist 
? http://lordoftheringsonline.wordpress.com 
          Lord of the Rings Online : LOTRO blog  
? http://lotr-online-leveling-guides.blogspot.com 
          Free Lord of the Rings Online Leveling Guide  
? http://female-gamer.com/lotro 
          LOTRO Insider Blog : female gamer blog  
? http://lotroblog.cn 
          Lotro Blog 
? http://lotro.mmorpgedge.com 
          LOTRO Edge 
? http://thesafehouse.org/forums/forumdisplay.php?f=64 
         The Safehouse Forums 
? http://ubuntuforums.org/showthread.php?t=386480 
          Lord of the Rings Online : forum 
? http://lotroquests.com 
        LOTRO Quests : guides, pictures, maps 
  
  
There.com: 
  
? http://there.com 
          There : official site 
  
  
Sony's Home 
  
? http://www.homebetatrial.com/  
        PlayStation Home Trial 
? http://www.playstationhome.com/  
        PlayStation Home Forums (unofficial) 
? 186?
 ? http://www.youtube.com/watch?v=oInP2DAa3BA  
        PlayStation Home GDC 2007 presentation video 
  
  
Entropia Universe  
? http://www.entropiauniverse.com/  
        Entropia Universe : Official website 
? http://www.euturnpike.com/  
        EU Turnpike : Ore Rates, Crafting Costs, Charts, Maps, guides, tools and more  
? http://www.entropedia.info/Page.aspx?page=Main%20Page  
        Entropedia : Charts, maps and guides  
  
 
The Sims Online 
? http://player.thesimsonline.ea.com/index.jsp 
        The Sims Online : Official website 
? http://sims.stratics.com/ 
        Sims Stratics: Sims Online Community 
? http://ea-land.ea.com/ 
        EA Land: The Sims Online, free version, launched in Feb 2008 
  
Digital Games (generally) 
? ?http://www.mobygames.com/home 
? ?http://grandtextauto.org/ 
? ?http://www.gamasutra.com/ 
? ?http://www.ludology.org/ 
? ?http://vgmaps.com/ 
 
? 187?
 Appendix I ? Sample Output from CopyBot 
 
<llsd> 
    <map> 
        <key>120692409</key> 
        <map> 
            <key>name</key> 
            <string>test object 2</string> 
            <key>description</key> 
            <string/> 
            <key>phantom</key> 
            <boolean>0</boolean> 
            <key>physical</key> 
            <boolean>0</boolean> 
            <key>position</key> 
            <array> 
                <real>136.809585571289</real> 
                <real>67.7011108398438</real> 
                <real>2000.5</real> 
            </array> 
            <key>rotation</key> 
            <array> 
                <real>0</real> 
                <real>0</real> 
                <real>0</real> 
                <real>1</real> 
            </array> 
            <key>scale</key> 
            <array> 
                <real>0.5</real> 
                <real>0.5</real> 
                <real>0.5</real> 
            </array> 
            <key>material</key> 
            <integer>3</integer> 
            <key>shadows</key> 
            <boolean>0</boolean> 
            <key>textures</key> 
            <array> 
                <map> 
                    <key>colors</key> 
                    <array> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                    </array> 
                    <key>scales</key> 
                    <real>1</real> 
                    <key>scalet</key> 
                    <real>1</real> 
                    <key>offsets</key> 
                    <real>0</real> 
                    <key>offsett</key> 
                    <real>0</real> 
                    <key>imagerot</key> 
                    <real>0</real> 
? 188?
                     <key>bump</key> 
                    <integer>0</integer> 
                    <key>shiny</key> 
                    <integer>0</integer> 
                    <key>fullbright</key> 
                    <boolean>0</boolean> 
                    <key>media_flags</key> 
                    <integer>0</integer> 
                    <key>mapping</key> 
                    <integer>0</integer> 
                    <key>glow</key> 
                    <real>0</real> 
                    <key>imageid</key> 
                    <uuid>89556747-24cb-43ed-920b-47caed15465f</uuid> 
                </map> 
            </array> 
            <key>volume</key> 
            <map> 
                <key>path</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>16</integer> 
                    <key>end</key> 
                    <real>1</real> 
                    <key>radius_offset</key> 
                    <real>0</real> 
                    <key>revolutions</key> 
                    <real>1</real> 
                    <key>scale_x</key> 
                    <real>1</real> 
                    <key>scale_y</key> 
                    <real>1</real> 
                    <key>shear_x</key> 
                    <real>0</real> 
                    <key>shear_y</key> 
                    <real>0</real> 
                    <key>skew</key> 
                    <real>0</real> 
                    <key>taper_x</key> 
                    <real>0</real> 
                    <key>taper_y</key> 
                    <real>0</real> 
                    <key>twist</key> 
                    <real>0</real> 
                    <key>twist_begin</key> 
                    <real>0</real> 
                </map> 
                <key>profile</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>1</integer> 
                    <key>hole</key> 
                    <integer>0</integer> 
                    <key>end</key> 
? 189?
                     <real>1</real> 
                    <key>hollow</key> 
                    <real>0</real> 
                </map> 
            </map> 
            <key>light</key> 
            <map> 
                <key>color</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                </array> 
                <key>intensity</key> 
                <real>0</real> 
                <key>radius</key> 
                <real>0</real> 
                <key>cutoff</key> 
                <real>0</real> 
                <key>falloff</key> 
                <real>0</real> 
            </map> 
            <key>flex</key> 
            <map> 
                <key>simulate_lod</key> 
                <integer>0</integer> 
                <key>gravity</key> 
                <real>0</real> 
                <key>air_friction</key> 
                <real>0</real> 
                <key>wind_sensitivity</key> 
                <real>0</real> 
                <key>tension</key> 
                <real>0</real> 
                <key>user_force</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                </array> 
            </map> 
            <key>sculpt</key> 
            <map> 
                <key>texture</key> 
                <uuid>00000000-0000-0000-0000-000000000000</uuid> 
                <key>type</key> 
                <integer>0</integer> 
            </map> 
        </map> 
        <key>120692410</key> 
        <map> 
            <key>name</key> 
            <string/> 
            <key>description</key> 
            <string/> 
            <key>phantom</key> 
            <boolean>0</boolean> 
? 190?
             <key>physical</key> 
            <boolean>0</boolean> 
            <key>position</key> 
            <array> 
                <real>0</real> 
                <real>-1.05799865722656</real> 
                <real>0</real> 
            </array> 
            <key>rotation</key> 
            <array> 
                <real>0</real> 
                <real>0</real> 
                <real>0</real> 
                <real>1</real> 
            </array> 
            <key>scale</key> 
            <array> 
                <real>0.5</real> 
                <real>0.5</real> 
                <real>0.5</real> 
            </array> 
            <key>material</key> 
            <integer>3</integer> 
            <key>shadows</key> 
            <boolean>0</boolean> 
            <key>textures</key> 
            <array> 
                <map> 
                    <key>colors</key> 
                    <array> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                    </array> 
                    <key>scales</key> 
                    <real>1</real> 
                    <key>scalet</key> 
                    <real>1</real> 
                    <key>offsets</key> 
                    <real>0</real> 
                    <key>offsett</key> 
                    <real>0</real> 
                    <key>imagerot</key> 
                    <real>0</real> 
                    <key>bump</key> 
                    <integer>0</integer> 
                    <key>shiny</key> 
                    <integer>0</integer> 
                    <key>fullbright</key> 
                    <boolean>0</boolean> 
                    <key>media_flags</key> 
                    <integer>0</integer> 
                    <key>mapping</key> 
                    <integer>0</integer> 
                    <key>glow</key> 
                    <real>0</real> 
                    <key>imageid</key> 
? 191?
                     <uuid>89556747-24cb-43ed-920b-47caed15465f</uuid> 
                </map> 
            </array> 
            <key>volume</key> 
            <map> 
                <key>path</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>16</integer> 
                    <key>end</key> 
                    <real>1</real> 
                    <key>radius_offset</key> 
                    <real>0</real> 
                    <key>revolutions</key> 
                    <real>1</real> 
                    <key>scale_x</key> 
                    <real>1</real> 
                    <key>scale_y</key> 
                    <real>1</real> 
                    <key>shear_x</key> 
                    <real>0</real> 
                    <key>shear_y</key> 
                    <real>0</real> 
                    <key>skew</key> 
                    <real>0</real> 
                    <key>taper_x</key> 
                    <real>0</real> 
                    <key>taper_y</key> 
                    <real>0</real> 
                    <key>twist</key> 
                    <real>0</real> 
                    <key>twist_begin</key> 
                    <real>0</real> 
                </map> 
                <key>profile</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>1</integer> 
                    <key>hole</key> 
                    <integer>0</integer> 
                    <key>end</key> 
                    <real>1</real> 
                    <key>hollow</key> 
                    <real>0</real> 
                </map> 
            </map> 
            <key>parentid</key> 
            <integer>120692409</integer> 
            <key>light</key> 
            <map> 
                <key>color</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
? 192?
                     <real>0</real> 
                    <real>0</real> 
                </array> 
                <key>intensity</key> 
                <real>0</real> 
                <key>radius</key> 
                <real>0</real> 
                <key>cutoff</key> 
                <real>0</real> 
                <key>falloff</key> 
                <real>0</real> 
            </map> 
            <key>flex</key> 
            <map> 
                <key>simulate_lod</key> 
                <integer>0</integer> 
                <key>gravity</key> 
                <real>0</real> 
                <key>air_friction</key> 
                <real>0</real> 
                <key>wind_sensitivity</key> 
                <real>0</real> 
                <key>tension</key> 
                <real>0</real> 
                <key>user_force</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                </array> 
            </map> 
            <key>sculpt</key> 
            <map> 
                <key>texture</key> 
                <uuid>00000000-0000-0000-0000-000000000000</uuid> 
                <key>type</key> 
                <integer>0</integer> 
            </map> 
        </map> 
        <key>120692411</key> 
        <map> 
            <key>name</key> 
            <string/> 
            <key>description</key> 
            <string/> 
            <key>phantom</key> 
            <boolean>0</boolean> 
            <key>physical</key> 
            <boolean>0</boolean> 
            <key>position</key> 
            <array> 
                <real>1.03192138671875</real> 
                <real>-1.05799865722656</real> 
                <real>0</real> 
            </array> 
            <key>rotation</key> 
            <array> 
                <real>0</real> 
? 193?
                 <real>0</real> 
                <real>0</real> 
                <real>1</real> 
            </array> 
            <key>scale</key> 
            <array> 
                <real>0.5</real> 
                <real>0.5</real> 
                <real>0.5</real> 
            </array> 
            <key>material</key> 
            <integer>3</integer> 
            <key>shadows</key> 
            <boolean>0</boolean> 
            <key>textures</key> 
            <array> 
                <map> 
                    <key>colors</key> 
                    <array> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                        <real>1</real> 
                    </array> 
                    <key>scales</key> 
                    <real>1</real> 
                    <key>scalet</key> 
                    <real>1</real> 
                    <key>offsets</key> 
                    <real>0</real> 
                    <key>offsett</key> 
                    <real>0</real> 
                    <key>imagerot</key> 
                    <real>0</real> 
                    <key>bump</key> 
                    <integer>0</integer> 
                    <key>shiny</key> 
                    <integer>0</integer> 
                    <key>fullbright</key> 
                    <boolean>0</boolean> 
                    <key>media_flags</key> 
                    <integer>0</integer> 
                    <key>mapping</key> 
                    <integer>0</integer> 
                    <key>glow</key> 
                    <real>0</real> 
                    <key>imageid</key> 
                    <uuid>89556747-24cb-43ed-920b-47caed15465f</uuid> 
                </map> 
            </array> 
            <key>volume</key> 
            <map> 
                <key>path</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>16</integer> 
? 194?
                     <key>end</key> 
                    <real>1</real> 
                    <key>radius_offset</key> 
                    <real>0</real> 
                    <key>revolutions</key> 
                    <real>1</real> 
                    <key>scale_x</key> 
                    <real>1</real> 
                    <key>scale_y</key> 
                    <real>1</real> 
                    <key>shear_x</key> 
                    <real>0</real> 
                    <key>shear_y</key> 
                    <real>0</real> 
                    <key>skew</key> 
                    <real>0</real> 
                    <key>taper_x</key> 
                    <real>0</real> 
                    <key>taper_y</key> 
                    <real>0</real> 
                    <key>twist</key> 
                    <real>0</real> 
                    <key>twist_begin</key> 
                    <real>0</real> 
                </map> 
                <key>profile</key> 
                <map> 
                    <key>begin</key> 
                    <real>0</real> 
                    <key>curve</key> 
                    <integer>1</integer> 
                    <key>hole</key> 
                    <integer>0</integer> 
                    <key>end</key> 
                    <real>1</real> 
                    <key>hollow</key> 
                    <real>0</real> 
                </map> 
            </map> 
            <key>parentid</key> 
            <integer>120692409</integer> 
            <key>light</key> 
            <map> 
                <key>color</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                </array> 
                <key>intensity</key> 
                <real>0</real> 
                <key>radius</key> 
                <real>0</real> 
                <key>cutoff</key> 
                <real>0</real> 
                <key>falloff</key> 
                <real>0</real> 
? 195?
             </map> 
            <key>flex</key> 
            <map> 
                <key>simulate_lod</key> 
                <integer>0</integer> 
                <key>gravity</key> 
                <real>0</real> 
                <key>air_friction</key> 
                <real>0</real> 
                <key>wind_sensitivity</key> 
                <real>0</real> 
                <key>tension</key> 
                <real>0</real> 
                <key>user_force</key> 
                <array> 
                    <real>0</real> 
                    <real>0</real> 
                    <real>0</real> 
                </array> 
            </map> 
            <key>sculpt</key> 
            <map> 
                <key>texture</key> 
                <uuid>00000000-0000-0000-0000-000000000000</uuid> 
                <key>type</key> 
                <integer>0</integer> 
            </map> 
        </map> 
    </map> 
</llsd>