The beloved simulation game Dwarf Fortress has an issue. Well, a few. First: no one outside its over a decade old community of fans really understands what it is, how to play it, or how to even wrap their minds around its huge, abstract ambitions.
The second is that the brothers who develop it, Tarn and Zach Adams, hold the only keys to the code of this ever-expanding, community-driven world they estimate will only truly be complete after many more years of development. And while they might be the benevolent gods of the Dwarf Fortress world, they're still just two mortals. They will die one day, putting this seemingly endless experiment at particular risk of extinction. To ensure it lives on, though, the creators are turning to fans by promising to open up the source code (ideally long before the Grim Reaper comes for them), and by also gifting that code to the dutiful archivists at the Museum of Modern Art in New York City.
"It's not just Dwarf Fortress, either—everything digital dies out if you don't take specific steps to preserve it," Tarn told me over email. He said that figuring out the logistics of preservation is also, "especially important if you want to keep making additions."
That's particularly important to Dwarf Fortress, because the whole conceit of the increasingly detailed simulation is that it has no finite end, really. In active alpha since 2006 and currently at version 0.43.05, Tarn estimates they're about 43 percent there when it comes to making a complete game. That means they're on track to release version 1.0 in the twenty-year deadline they set for themselves, completing their massive 2,600 development item list. But that might not even truly be the end of it. They've got some pie-in-the-sky ideas for after they get out of alpha, "But if 1.0 is really 20 years away, I'll be fifty-nine [years old] and who knows?"
While digital preservation is important in all areas of virtual culture, the historical cost of not preserving Dwarf Fortress could be particularly devastating. If the source code was lost, it'd be an unimaginably arduous process for modders to reverse engineer it from the binary, and it would take an even longer time to clone the game. And, according to Tarn, "There's a lot of detail in the simulation, and quite a bit of it is unique to the game, as far as we know."
Despite its conceptual hugeness, the Dwarf Fortress source code is actually fairly small. The duo's current arrangement with the museum stipulates that, if they died, someone on behalf of the developers would send the source, project files, and digital design docs (which are just a bunch of text files) to whoever's running MoMA's design department at the time.
But the museum is pushing to get their hands on even more than that for historical record. "They even asked for any paper notes we'd discarded, but we haven't taken them up on that," Tarn admits. "It's almost hard to take that seriously, looking at our scribbled garbage."
More than even the MoMA, though, the brothers are putting their faith in the community to keep the dream alive long after they're gone. "I don't think we'd actually need to do anything, even release the source, for Dwarf Fortress' community to carry on," he says. "It would just be better if people could make their own changes."
Correction: This article originally stated that Tarn Adams thought Dwarf Fortress would take hundreds of years to complete. Adams was exaggerating and jokingly referencing the ninety-ninety rule. The article has been updated to clarify Adams thinks it will only take 20 years to complete development.