This is a postmortem for Assembly.
Rereading, it doesn’t seem too spoilery… but you might enjoy playing the game first anyways.
Timeline
-
Summer 2005 (approximately): I discover IF, pore through a bunch of craft/design posts, judge in one IFComp, then mostly forget about IF for a decade or two.
-
Summer 2021: I started working on a Z-machine as a pandemic project, so I was hanging around the forum by this point. My computer tells me that earliest files for the game date back to then. I think I was mostly playing around with the furniture assembly/disassembly code still; there was certainly no real game there yet. (Also, the original working title was Furniture Assembly Simulator; extremely glad I pared that down.)
-
Summer 2022: The overall shame of the game got figured out, and I made an honest go at getting it together for that year’s IFComp. Despite a couple weeks of mid-September crunch, it was pretty clear that I wasn’t going to have anything worth playing at the deadline. I’m very glad I didn’t try and push through at the time… though it killed my momentum on the project, and it saw basically no changes for several months.
-
Summer 2023: I’d decided to try and polish things up for that year’s comp. It took much longer than I thought still, and some important elements of the design — the wardrobe, the overall puzzle chart — didn’t gel until pretty late. I ended up with just a few weeks for testing, which of course was not nearly enough. Thanks to testers @Doug_Egan, @Scrooge200, and @Keltena for battling through some pretty rough bugs and bad writing… the game is hugely better for it.
Concept
Assembly started with the concept, naturally. I was interested in the way that rituals give a significance to ordinary (and conveniently parser-friendly) actions — didn’t @zarf write something about this? — and the furniture-assembly thing, despite being unforgivably silly, had a bunch of thematic connections that felt solid enough to hold things together. Plus you get a setting for free.
@VictorGijsbers’ review points out another parallel: between following instructions in an instruction booklet and typing out commands from a walkthrough. This was also an important part of the concept for me: “type out commands from the instruction booklet” was there from the very beginning.
However, a problem is: typing commands out from a walkthrough is not inherently very fun. It’s probably not a smart decision to choose an inherently unfun mechanic for your video game, no matter how charming you think the idea is, but I do not always make the smart decisions. So I tried hard to compensate in other ways:
- Keep the instructions short. Any ritual the player is likely to complete more than once has only a couple steps.
- Be pretty flexible about input, disambiguation, etc. “Put legs on tabletop” is not really what the instructions say, but the intent is clear, so it’s allowed.
- Write rich descriptions. I spent a lot of time trying to come up with interesting things to say when assembling the DÖLMEN… that text serves a few different purposes now, but I originally put it in just so the experience wasn’t so dry.
I think this mostly worked out, though there’s certainly more to do.
The Comp
Assembly seems to be the most rated and reviewed parser game for the comp. I don’t have any explanation for this. (Many better games were played less often.)
One particularly bonkers thing to me is that about one in three people who rated the game also reviewed it. And the reviews are overall extremely thoughtful! I mostly didn’t respond to reviews during the comp, but I do want to take the chance now to thank everyone who found something to say about my little game.
And specifically, I’d like to thank folks who wrote critical reviews. Back when I first encountered IF, I was struck by the depth and quality of the theory/craft discussions rattling around the community. Many things have changed since then… but the thoughtful and serious approach folks many folks in this community take to the craft has not, and it’s been a remarkable experience to have my own game carefully taken apart and examined.
Retrospective
This is the first fiction I’ve written in well over twenty years. As far as I can recall, the last thing I wrote for fun was also a fantasy story… but it began with the line
“Aaaaaaaaaaaaaaaaaaaah,” Devlin yelled as he fell into the dragon pit.
and did not improve. Which is to say: I had a low bar to clear with respect to the quality of the prose. I still struggle with descriptive writing, but I’m happy with some of the work in there, and I’m pleasantly surprised that many reviewers were too.
The pacing of the story is not ideal — the ending’s quite rushed, for example — though some of the individual setpieces seem to work. I was very worried about the early down-the-rabbit-hole moment, which is abrupt and dense in a way that I could not seem to avoid, but it seems I didn’t lose too many players there.
I’m not sure what to think about the puzzles. Some players liked them a lot, while others had a hard time. I have a few theories about this, though some of it probably comes down to taste. I think I’m too attached to the overall puzzle design to make big changes there, though I do think some of them could be signposted better or provide better feedback.
@kaemi’s review notes that Assembly doesn’t actually do as much with its concept as it could, or perhaps promised to do. (Like, thematically — I think it squeezes most of the juice of the puzzle idea.) I think this is right and I have regrets about it. If anything convinces me to do a proper rework of Assembly, aside from the usual post-comp tweaks and bugfixes, it’ll be finding a way to do these ideas more justice.
Other things I feel like mentioning
I didn’t realize it going in, but for a short game with a novel / systematic mechanic, a lot of the trad text adventure tropes are also there: a maze, a rope, a darkness puzzle, etc. Not sure how that happened.
The themed web interpreter was pretty easy to do – it’s just a hacked-up single-column Parchment template – and I recommend it to other authors. Parchment provides a bunch of variables that make the CSS pretty easy to tweak, and the payoff is a potentially nicer and more distinctive reading experience.
