In a comment on my recent post on text generation, Dryman wrote
I was wondering if you’d been following some of the recent games criticism discussing where procedural content has ultimately failed to be very interesting or engaging (in games such as Spore or No Man’s Sky), and might have some general thoughts about how procedurally generated text content can potentially be made to resonate more strongly than was the case for the largely graphical reskins in those games. Much of the discussion has focused on what Kate Compton has called the problem of “procedural oatmeal” – i.e. it is very easy to pour 10,000 bowls of plain oatmeal, with each oat being in a different position and each bowl essentially unique, but it is very hard to make these differences *matter* to an audience, and be perceived as truly different in any memorable or thought-provoking way.
(“Some of the recent games criticism” also includes this article from Mike Cook on changing how we talk about procedural generation.)
In response to which, I partly want to point to Bruno Dias’ recent Roguelike Celebration talk, which addresses exactly this point: that procedural generation is not a way to get large amounts of content from small amounts of input, and that procedural generation requires an appropriate design approach.
But to speak for myself: I think the key question in oatmeal-avoidance is whether the generation is connected to anything mechanical. I might be able to generate haiku, or funny food names, or imaginary constellations, or names of funny English-sounding towns, but all that generation is purely decorative unless it is tightly correlated with gameplay — and the player will soon realize that it is decorative and start looking past it.
But what kind of mechanical connection is required? It would be a lot to ask that every procedurally generated variant needed to correspond to different gameplay affordances — that procgen is only interesting if having 10K output states means 10K different verbs or verb clusters or play strategies that could be unlocked.
That’s a very demanding design problem. It’s not completely inconceivable to have elements that work together to generate fresh mechanics: Dominion‘s longevity comes from the fact that the cards play off one another in really interesting ways and make each other useful (or not useful) quite quickly: Peddler might be pointless in a game without +Action and +Buy cards, but in a game with a good action/buy engine and, say, Bishop, it might become a massive source of victory points. There are entire blogs devoted to pointing out productive Dominion card combos and strategies. But that’s a challenge in mechanics design, and if every application of procgen text had to be that mechanically rich, we wouldn’t need it very often.
But there are other applications as well, happily. In The Mary Jane of Tomorrow I wanted pretty much everything the robot said to serve as a reminder of her current training state: it acts as both status information and as a perceivable consequence of what the player has done with her. Show her a book about cowgirls, and she’ll start speaking in a funny accent. Teach her about botany, and suddenly she’ll get really specific about the types of apple she can bake in an apple pie.
Procgen methods are great for this kind of low-level, layered consequence: representing how the player has changed the world state persistently in a small way, rather than massively in a way that forked the narrative.
At the same time, ideally, you get some results that aren’t exactly what the player might have predicted or anticipated. The robot making up haiku with botanical names and a cowgirl accent is juicier, richer, and more appealing than a status display on the side of her neck that says
Or to look at this another way: I spend a lot of time, in my projects, coming up with little jokes and surprises and colorful images to reward the player for trying a particular thing. Even when I’m writing for Fallen London, I really like to come up with callbacks: one of the most fun things about writing The Frequently Deceased was inventing ludicrously unsuitable ways for small children to interact with the player’s inventory items. (You can pick up a very, very large collection of child-inappropriate possessions in Fallen London.) Procedural methods let me create some of those experiences dynamically.
And then there’s the specific task of making dialogue feel highly responsive to the context in which it appears. A lot of my own interest in procgen text comes out of the desire to combine authored, narrative centric dialogue with ways of expressing relationships, social moves, and emotions. So I’m less casting about for what the model should be, and more trying to figure out how to write the generator to produce the level of richness I want.
Still, at its best, I think procedural generation, textual or otherwise, can be used to reinforce that sense of agency and presence in the world, rather than to make it feel as though nothing matters. But it isn’t a substitute for designing content. It’s a way of designing content — one that is often at least as labor-intensive as other ways, and that also demands a strong capacity for abstraction and the ability to characterize one’s aesthetic goals.