Saving Virtual Worlds From Extinction

Coriolanus

Thought some here might find this interesting.

http://arstechnica.com/gaming/news/2010/06/the-art-of-archiving-virtual-worlds.ars

Sometime this August, librarians at the University of Illinois will finish archiving over a dozen famous computer games, then step back to consider where to go next with their project. These programs go back over four decades, and include a 1993 version of Doom, various editions of Warcraft, and even MIT's Spacewar! circa 1962.
We wondered, given the gaming nature of most of the software being preserved, why the venture is calling itself the Preserving Virtual WorldsJerome McDonough, Assistant Professor of Library and Information Science at the University of Illinois, to ask him about the name. project. So we called up the project's coordinator,
Ars Technica: Why "worlds?"
Jerome McDonough: Mainly because we couldn't come up with any word that described all the types of different things that we were interested in. We're looking at everything from Spacewar! on the PDP 1 at MIT in 1962 through Second Life today. If we look across all these games, there's no commonality in terms of hardware or operating system. There's really not that much commonality in terms of thematic content of the games.
Pretty much the only similarity you can point to is that in each of them you have an effort to develop some sort of virtual environment or virtual world. They're all in some way a simulation. In Spacewar! you are dealing with the first space combat simulation game. In Adventure, the classic text game, you've got a virtual cave system that you are negotiating. When are you are dealing with Star Raiders on the Atari 2600 you've got a galactic space battle with a galactic map. All of them are defining some kind of virtual space.
Ars: Why do librarians and archivists want to preserve games?
JM: The really simple, one-sentence answer is because games are important. In the United States we're looking at about 80,000 people who are directly employed by the gaming industry and maybe another 240,000 people involved in related, tangential industries that rely on gaming companies for their existence. So just as a monetary phenomenon, games are important. You probably saw the sales for Modern Warfare? We're talking a single game that realized over a billion dollars in sales. Sort of shows on a monetary level the importance that games have taken within our economy.
This has certainly made librarians take note of games, but also they've become important culturally. There's a long history of wanting to say "popular culture is lower culture and therefore we should not be preserving it." For all of us in our project, we're rejecting that point of view. Popular culture is the most important culture we need to preserve. It shows what people were actually interested in and what they were doing.
In the United States, we have two thirds of American households as active gamers. This is what people are doing with a significant amount of their time, and if people in the future are going to understand what this society is like, they need to understand gaming. And how people gamed.
You also can't understand some other parts of our cultural world unless you preserve some of the game world. There's a lot of sort of interpenetration of media. On the importance of preserving a game like Doom—well, if you're preserving something like The Simpsons, you're not going to be able to understand the Doom references that they made visually in a few Simpson episodes unless you've got a copy of Doom sitting around.

Ars: How will these games be preserved?
JM: Really good question! And I can't say we're completely certain of all the answer at this point. But we're looking at whether some of the models that have been developed within the digital preservation community for dealing with more static forms of data can be applied to software. And what we're finding is that they can be. But if you're going to apply those models, you really have to be very proactive in the materials that you collect.
Take, for example, Star Raiders on the Atari 2600. If you're going to preserve this, you've got a couple of problems. The first is that it is on a cartridge that is designed to work on a particular system that is no longer manufactured. And as long as you've got a hardware dependency there, you're really not going to be able to preserve this material very long. What we have been looking at is how feasible is it for things that fundamentally all have some level of hardware dependency there—even Doom has dependencies on DLLs with an operating system, and on particular chipsets and architectures for playing. How do you take that and turn it into something that isn't as dependent on a particular physical piece of hardware.
And to do that, you need information about that platform. You need technical specifications that allow you to basically reproduce a virtualization that may enable you to run the software in its original form in the future. So what we're trying to do is preserve not only the games, but preserve the knowledge that you would need to create a virtualization platform to play the game.
Ars: Where will these games be preserved?
JM: For the purposes of our project we're looking at how our existing repositories at some of our universities handle this type of material. So at the University of Illinois we have a central institutional repository called IDEALS. It's based on the DSpace software that was developed by Hewlett Packard and MIT. At Stanford they have a different repository system, the Stanford Digital Repository that's running off the Artesia TEAMS software. So we're going to be ingesting the games into those two different repositories to see how they handle the rather large body of metadata that we're having to develop to track all these bits and pieces that we need to preserve any given game.
Longer term, part of our project is actually going to be trying to provide the Library of Congress with guidelines on what a national collection policy for games ought to be.
Ars: Any chance people will eventually be able to go and play these old games online?
JM: Sad to say, probably not, at least not through us. One of the other issues that we have been looking at in this project are issues of copyright. What can we do legally under existing law, including the Digital Millenium Copyright Act. What actions can we take? What can we not take? Everything we're doing is within the bounds of the law, but some of what we're doing may involve negotiations with companies in which we say, "We don't want to ruin the potential market for your game, we just want to preserve it. So will you allow us to make preservation copies of this material if we dark archive it and don't make it publicly available?"
A lot of publishers that might otherwise have problems with us going and storing a copy of the game and making it available are a little more agreeable if we say that we're just concerned about long-term preservation. We'll agree to dark archive this material for the term of copyright as long as you give us permission to preserve it. So we're trying to keep this stuff around in the long haul. Will we be able to give public access to all this material? No. But some of it we can because some of it has already gone out of copyright.
So early games like Spacewar! and Adventure that were actually published prior to changes in copyright law in 1976 and were published without a copyright notice were basically entered into the public domain at that point. There are other games, like Mystery House, that the creators actually made freely available over the Internet. So some of the materials we can make available publicly, other of the materials we can't.
One of our biggest issues for copyright has been Second Life, because everyone who creates something on Second Life retains copyright over their creation, which means that if I want to make a copy of a Second Life island, I have to contact every rightsholder on the island and ask them for permission to copy their particular stuff.
Ars: That sounds complicated.
JM: It's been painful. We had to develop our own software first to identify all the rightsholders on the island and then develop an automated system to contact them, tell them what we're doing, and ask their permission to archive and enable them to go to a website we'd developed where we say "We think these objects are yours. Indicate the ones that you are willing to give us permission to copy."
Ars: What's your timeline on getting this going?
JM: Our project actually ends in August, so what we're hoping to have accomplished by then is get all our initial case set described, packaged up and put into the repositories at Stanford and the University of Illinois. Also provide the Library of Congress with guidance on how we think these materials should be packaged up as well as advice on what a national collection plan for games ought to look like.

We're already talking about doing some further research in this area. One of the things we've become worried about is just how successful running these games under emulation is. It was actually Doom that got us worrying about this. We were trying to look at the very original DOS shareware version of Doom and discovered that a lot of the existing DOS emulation software had problems running Doom perfectly. And probably the biggest problem for the emulators was sound.
Because this was the DOS world and because it was before the great harmonization of sound that happened with DirectX on Windows, most of the DOS emulators don't have perfect handling of all of the different types of sound cards that existed in the pre-Windows world. Some of the DOS emulators we couldn't get sound out of at all. Some of them we got something, but it wasn't the original music. We had something like six different emulators we were trying on four different platforms. And of those various combinations, I think only two of them allowed us to get the sound out of Doom perfectly.
Ars: Where do you want to go from there?
JM: From there we want to start looking at how effectively we can preserve these things using emulation software. One of the basic tenets of digital preservation is you want to leave the original bitstream intact. For those cases where we've got a binary, executable form of the game like Mystery House, if I'm going to provide access I basically have to run an emulator of some kind.
How good a job do emulators do? We've seen enough in our existing research to know that emulators are not always perfect. So we're looking at what properties of games do people think are significant, and how effectively do emulators preserve those significant properties for people who are looking at them.