"The very nature of digital [history] is that it's both inherently easy to save and inherently easy to utterly destroy forever."

Jason Scott knows what he's talking about when it comes to the preservation of digital software. At the Internet Archive, he's collected thousands of classic games, pieces of software, and bits of digital ephemera. His sole goal is making those things widely available through the magic of browser-based emulation.

Compared to other types of archaeology, this kind of preservation is still relatively easy for now. While the magnetic and optical disks and ROM cartridges that hold classic games and software will eventually be rendered unusable by time, it’s currently pretty simple to copy their digital bits to a form that can be preserved and emulated well into the future.

But paradoxically, an Atari 2600 cartridge that’s nearly 40 years old is much easier to preserve at this point than many games released in the last decade. Thanks to changes in the way games are being distributed, protected, and played in the Internet era, large parts of what will become tomorrow's video game history could be lost forever. If we're not careful, that is.

Throwing away the layers

In an industry with a nearly constant focus on the latest and greatest, it may seem silly to want to preserve old, outdated versions of today's games for posterity. But the current nostalgic and research interest in games from the ’70s and ’80s shows just how important it is to try to save a complete record of today’s titles, as ephemeral as they sometimes are.

"I totally get that people look at this and say all of this game history stuff is navel-gazing bullshit... an irrelevant, wasteful, trivial topic,” Scott told Ars. “[But] mankind is poorer when you don't know your history, all of your history, and the culture is poorer for it. It doesn't matter if it's games or civil wars or highways or government machinations. If you don't have that historical context, you make poorer decisions, you make the same mistakes again and again, and you end up with an eternal present. You don't understand where things are and where they're going, because you're constantly in the now."

And in today's game industry, being "constantly in the now" often means throwing out masses of current history without a thought. "I use FarmVille often as my go-to example of this because, like it or not, that game is historically significant and will be studied," gaming historian and Lost Levels creator Frank Cifaldi told Ars. "Keeping an offline game safe is pretty easy, but what do you do for FarmVille, a game that is constantly updated, to the point where Zynga manipulates it server-side? Do you try to take one daily snapshot of it? Is the FarmVille you can play right now actually FarmVille? What about the FarmVille that existed a month ago? What about the very first build of it? Is it even possible to preserve enough playable code to say that the entire experience that was FarmVille is safe?"

These days, it's not just Facebook games that are having their internal history slowly peeled away. "I was just talking to someone about Diablo, [and] he was saying how he kept Diablo II going, how the original version is his favorite to play," Jon-Paul Dyson, director of the International Center for the History of Electronic Games, told Ars. "With Diablo III the game is very different [now] than the first version with all the updates that have been made to it. From a preservation point of view, are you collecting every version that's been created?"

"An analogy here is maybe to a piece of architecture," Dyson continued. "When you're seeking to preserve a historic house, there may be layers, it may have been lived in by many different people. Mount Vernon had been lived in by George Washington's descendants, so they made a decision to restore it to George Washington’s time and erase this later history. Do you make the same kind of decision with games?"

The death of “accidental ambient archiving”

Such historical restoration may not be easy with many modern titles. When updates are automatically pushed out and applied over the Internet every time you log on to a console or PC to play, those historical layers are erased en masse without a thought. Where patches may have once gone out via FTP sites, where they could be archived and studied, now the process is hidden. That's more convenient for the players, but it's the equivalent of a constant digital purge from a historian's point of view.

"For that convenience [of automatic updating], we lose a lot of what you might call accidental ambient archiving," Scott said. That's the kind of archiving that happens when players store copies of old cartridges or discs in their attic, where they're eventually recovered as a static record of the digital times. In the near future, it's going to be difficult to find people who held on to purely digital games on their hard drives and much harder still to find unaltered launch-day versions that haven't been overwritten by these automatic patches.

As an example of why this kind of early version archiving might be necessary, Scott came up with a theoretical case of a game featuring an enemy group logo that looks similar to that of a real-life terrorist group. After an outcry, the developer issues an automatic patch to replace the original logo with something more benign. That would be a PR win in the present, but a major loss from a future historian's perspective. "Now we have a revisioning of what happened in the world, and [it becomes] extremely difficult to point to this as an example of anything from cultural imperialism to the nature of politics," Scott said.

That problem will only become more common as games continue their seemingly unstoppable transition from static physical objects to ongoing services provided through a centralized server. "Eventually it's going to be… Skyrim 10 as an environment and you pay $14 a month for everything, and then Skyrim slowly changes over its lifetime,” Scott said. “How do you capture the earlier states? Preserving any idea of that is going to be very tough."

The original developers and publishers themselves may save copies of those “obsolete” versions of games, but most companies don't have the resources or the wherewithal to keep track of this kind of internal history in an exhaustive way. "There's no downside to them in destroying work product on the way to the next project," Scott said. He's heard stories at conferences of developers who, "when they finished the game, they just wiped the hard drive of the work machine and set it up for the next project. That's all that intermediate work, gone."

Convincing companies that this stuff is worth saving for posterity or sharing with an outside historian can be difficult. "Often when working with companies, it comes down to finding a passionate individual in a company who will work with you," Dag Spicer of the Computer History Museum said. He recalled efforts to track down early versions and source code for Mac Paint inside Apple. After he hit a dead end with the “official” channels, Spicer finally broke through when a friend of Steve Jobs put him in touch with the late Apple founder and CEO. "[Jobs] sent a one line e-mail saying it was a good idea, and it was done the next day," Spicer recalled. "Having an internal advocate is key."

Outside of that, Spicer said we may be reliant on lucky breaks to save these pieces of computer history from a "digital dark age." Researchers have already stumbled on old computers that have original versions of Windows that were just never updated and historically valuable prototypes that went home with engineers when companies went under. In a few decades, historians may end up searching for someone with an ancient iPhone who never bothered to click the "update" button on that launch-day version of FarmVille.