Saving video gaming’s source code treasures before it’s too late

[Karlston]

straight to the source —

Saving video gaming’s source code treasures before it’s too late

Over 90 percent of pre-2000 gaming source code may already be gone.

Enlarge / The Days of Thunder NES prototype source code on these disks sat in a programmer's basement for 30 years before being uncovered upon his death.

VGHF

89 with 57 posters participating

 

When most people think of preserving video game history, they probably imagine a museum full of boxed consoles and cartridges, or maybe a massive, digital database of emulatable ROM files ripped from the original physical media. The Video Game History Foundation's latest project is looking past those kinds of basic archival projects, though, and toward collecting and preserving the source code behind many classic games.

 

"For a video game historian, an archaeological dig through source material is the next best thing to time travel,' VGHF's Frank Cifaldi said. "A really good source repo is the closest you're ever going to get to being a fly on the wall during a game's development."

 

Digging through a game's source code repository can help archivists discover previously unknown content and information. Back in 2017, for instance, a VGHF analysis found unused Disney character art and animation in the source code for 1993's Genesis Aladdin game ("Some of that was in a folder called trash," Cifaldi told Ars). More recently, the VGHF team discovered the source code for Nuclear Rush—a game prototype for the unreleased Sega Genesis virtual reality headset—and remade it to work on modern VR hardware.

 

But it's not just about finding and restoring "lost" content. Looking at source code can let a knowledgeable historian "understand the creative process a lot better than you would otherwise because it is the purest form of the game," Cifaldi said. "It is the raw building block that made it up."

 

By way of example, Cifaldi talked about his time looking at the source code for The Secret of Monkey Island and learning the language and tools behind LucasArts' famed SCUMM engine. "Now when I play a SCUMM game, I understand it a little more. I speak the language a little better," he said. "In knowing SCUMM, I just have this intuitive understanding of why creative decisions were or weren't made in its game, and I don't think I could possibly have that level of understanding without access to SCUMM language."

Before it’s too late...

Unfortunately, games like Monkey Island (where creator Ron Gilbert maintained and generously donated the source code) are the exceptions. For games made before 2000, Cifaldi estimates that the source code is completely gone for over 90 percent of "the games we care about."

 

"I'm sure a lot of source was saved at one time, then an office move happened, someone said, 'What's this old DAT we have in a closet? It's probably backed up somewhere, just toss it,'" Cifaldi said. "I think the situation is pretty bad. It's only going to get worse as the people who actually archived this stuff age out and leave the industry, if not this mortal coil."

 

Footage of a cut Sword Swallower enemy that was restored to 1993's Aladdin Genesis game from the source code.

 

When early game source code survives, Cifaldi said it's often by pure happenstance, "individuals who maybe brought their work home and happen to have that material." The source code for a previously unknown Days of Thunder prototype on the NES, for instance, were found by digging through "a pile of floppy disks in the basement of its deceased developer," as Cifaldi recalls. He imagines there are "hundreds more [game source code repositories] like that, just floppies rotting in someone's basement waiting for someone to recover."

 

That's why VGHF is putting out the call to developers who may have these important historical artifacts among their possessions. While the project is in its early days, Cifaldi said the foundation currently has about 100 titles with source code repos "in various states," including some that are just printouts on endless reams of paper.

Who owns history?

Even when the source code does still exist, preservationists often run into problems with protectionist companies trying to jealously guard their IP rights. "[Source code is] seen as a trade secret, and there's no statute of limitation on when people are comfortable with a trade secret," Cifaldi said.

 

Convincing companies to share those "secrets" for historical posterity—and to spend the time and effort looking for it in the first place—can be an uphill battle, Cifaldi said. But that's starting to change in some corners, in part because of the industry-wide trend toward remakes and remasters for well-remembered games.

 

In the case of a game like Grim Fandango Remastered, for instance, "the only reason that product exists is because employees had things in their houses and people know who to ask," Cifaldi said. Without the source code, doing that kind of remastering work is a much bigger lift—when Cifaldi worked at Digital Eclipse, none of the games had available source code, meaning the team had to create an entirely new engine and reverse-engineer the games from extant ROM files.

 

Companies are starting to cotton on to the value of this kind of preservation, though. Disney started to realize the value of source code preservation, for instance, when it used that raw material in 2019's collection of remastered Aladdin and Lion King games.

 

"In all cases, there is interest internally because a lot of the companies we've spoken to have recognized that it sparks fan interest if people are talking about these older products," Cifaldi said. "A lot of commercial film preservation, remastering movies for Blu-ray or whatever, a lot of that happens because the Library of Congress has the master film reel in its archives... The idea of video game source repositories is to me no different than that."

 

Commercial re-releases aside, though, Cifaldi said the source preservation project is seeking to provide the raw materials necessary for continued scholarship on classic games. He likened it to the documents and letters that serve as the basis for civil war histories or the archived production materials that illuminate the creation of Citizen Kane.

 

"We want to see more books on bookshelves, more documentaries out there, more in-depth study than we're currently seeing," Cifaldi said. "Source repositories like this, this is where history books come from."

The Gigaleak

It's hard to talk about game history and trade secrets these days without addressing the "Gigaleak" of Nintendo prototypes and unused designs that hit the Internet in July 2020. Cifaldi said that this kind of leaking of closely guarded company secrets is "not something that we're trying to encourage. We're not asking people to conduct commercial theft for the sake of preservation necessarily, but it's something we have to be cognizant of."

 

Cifaldi was particularly distressed that the leak included potentially embarrassing personal details, like naughty words used in source code snippets or emails between developers. "That's immature of the individual to do that, but it also reflects an immaturity in the public understanding of how games are made and that things can get a little messy."

 

That kind of stuff getting out "potentially could hinder efforts like ours" to preserve things legitimately, Cifaldi said. At the same time, "those leaks, the good side of it is that it is sparking imaginative conversation among fans and historians about the making of Nintendo's products. It is giving people a much clearer understanding of the game development process than they might have had in the past, which I'm always a fan of."

 

 

Saving video gaming’s source code treasures before it’s too late