DOL-OS

One of the things I like best about IF is its austerity. Mainstream video games have had decades to develop the sensory delights they offer, from photorealistic graphics to pleasingly responsive interfaces to viscerally satisfying sound affects, but I often find the humblest piece of interactive fiction more impressive because it’s living and dying just by its text. Sure, there can be various bells and whistles on both the parser and choice-based side of things – integrated music, some attractive pixel art, that sort of thing – but to be honest I find that stuff rarely makes much difference to me when I’m reflecting on my response to a game: it’s all about the words, and how they’re used. So it’s hopefully a marker of DOL-OS’s aesthetic achievement that the main thing I keep coming back to is how pretty it is.

The game presents itself as a sort of found-object piece: the conceit is that it’s far in the future and you discover an old but still functional computer, so you decide to undertake some digital archaeology. It seems as though it dates from some time at least a few years on from our present, so strictly speaking it doesn’t make sense that the presentation relies on 80s-era markers – a yellow-green palette, graphics broken up by scan lines, chunky, pixelated icons – but this is what my brain, at least, thinks a computing artifact should look like, so I think it’s an inspired choice. And it commits hard to the conceit; every font, image, and glitched-out display is note-perfect. Similarly, I don’t know what kind of work was needed to torture Twine to create an interface that functions exactly like I remember the old Apple II ones working, but it’s similarly an impressive achievement – navigating the file system is immediately intuitive, and there are myriad extras all the way up to interactive implementations of hangman and sudoku. Truly, it’s a triumph – if, instead of a self-contained piece of IF, it was embedded in a AAA game like one of the later Fallouts, it would inspire excited PC Gamer blog pieces about this awesome Easter Egg everyone should check out.

(Now that I’ve typed that out, both in presentation and plot I realize that DOL-OS would perfectly fit as one of those terminals you occasionally run across in Fallout – I’m curious if there was any direct inspiration there?)

As for that plot, there’s a fair bit of it. DOL-OS proceeds in three distinct layers, with the game’s two puzzles gating progression between each act. The first act is a collage, with a variety of different documents painting a picture of a repressive, authoritarian society. The files on offer include news stories about public executions, official histories, annotated literary texts, official documents… it’s a real mélange, and while there are a few connections between the various small stories on offer, those function more as bonus insights; the point is just to experience the many different ways a society like this commits violence against its citizens. The writing here is often stilted, reflecting the way that fascist states manage to use language bluntly while still avoiding saying what they mean:

"We encourage still that anyone having had contact with The Gendarme to deliver to the nearest police station any information that might to help recover these documents or in relation to the young woman and her connections."

(I should note that the game was translated from French – it won last year’s French Comp – so some of these stylistic tics might be a result of that process. It still works well, though).

There’s room for a bit of humor, though – the story implied in the terse notation that one criminal was punished for “[stealing] a goose thrice (same goose)” is marvelous.

The second section is more focused; now the documents are following a young researcher who’s been brought on board a mysterious project, that involves both digitizing historical documents and developing an AI. This part of the game proceeds fairly linearly, as you read his diary and his involvement in the project gets deeper and deeper. The final section sees you engage with the research project directly and shifts from the document-review gameplay of the first two-thirds of the game to a more traditional choice-based interface, which effectively raises the stakes and indicates that the focus has moved from understanding what’s on this old computer to deciding what to do with it.

It’s a nicely-paced progression, and as a result I’ll reserve in-depth discussion of where the game winds up going for a spoiler section; suffice to say that that I think the plot works well, though perhaps takes a slightly more tropey and bloodless approach to an issue that could have been rendered with a bit more social realism. And while I’m being slightly critical, I’ll also say that the puzzles are rather desultory; there’s a guess-the-password bit that’s got some very blunt clueing, and a jigsaw-puzzle captcha where the main challenge was avoiding getting a headache from squinting at a bunch of yellowish smudges. They’re by no means bad, but at the same time I can’t help but think I’d have enjoyed the game more if they’d either been made more demanding, or dropped entirely so that progress just depended on reaching certain milestones in the document-review process. These are especially niggly niggles even by my standards, though – DOL-OS stands as a really impressive game that deservedly won the laurel in its Francophone version, and us English speakers are lucky to get another bite at the apple.

(Spoiler - click to show)Right, so the revelations: it turns out that the authoritarian era is well in the past by the time the researcher starts up his work, and in fact at first his job is just to digitize the records you find in the first chunk so they can be used as AI training data. The judicial system is overburdened in his time, you see, so the project is all about creating to a tool to speed up the slow business of deciding guilt and punishment; the previous project lead gets chewed up by the stress and ethical compromises, so the researcher gets thrust into the limelight, at which point it becomes clear that the bosses don’t care that the AI is bloodthirsty in the extreme. The story breaks off somewhat at this point, but when the third act kicks off and you’re able to engage the AI in direct conversation, it becomes clear that it was in fact deployed and wound up passing judgment on a whole lot of folks. This final tete a tete makes clear a lot of the stuff that’s established by implication in the first two sequences before building to a climactic choice of either consigning the AI to its doom on the failing terminal hardware, or copying it over in a fresh install.

This all works well enough, and I have to give the game kudos for creating a “save the AI y/n” moral dilemma where I was all in on letting the thing die, but I did feel like it could have played things with a bit more nuance. These kind of systems are being implemented in real life – most notably, a lot of jurisdictions have experimented with algorithms to make recommendations for who should be offered bail after being charged with crimes. You can see how this might be a good idea in theory, but in practice mostly they just wind up laundering racist decision-making via a Big Data disguise; there are well-known cases where first-time Black offenders aren’t recommended bail, while white career criminals get every benefit of the doubt, because that’s what the algorithms learned to do from the training data. Beyond these instances of straight bias, there’s also a ideological element of horror here; in the Anglo-American criminal system, at least, decisions of guilt are consigned to a jury of ordinary people, and taking a social judgment and turning it into a data-driven one is a radical shift, and I wish more hay was made of it. But DOL-OS mostly refrains from plumbing direct real-life analogues or self-consciously putting big ideas into play; the second section sticks to the well-worn Frankenstein-y scientists-create-monster-that-escapes-their-control plot beats, and the third section doesn’t create much nuance or ambiguity. All told that means I found DOL-OS an effective bit of sci-fi horror – and again, a gorgeous example of the form – but I was disappointed it didn’t try to do a bit more in the way of social comment.