Hotwash

Well, the 2024 Autumn Lisp Game Jam is over... And hey, not too bad a showing this time!

So what went right?

First the concept. I was originally thinking about doing a visual novel or adventure game of some sort. However, just before the jam started, I got the idea of platformer with a cat. That broke stuff. Very simple concept, but it would have enough complexity to be interesting without overwhelming me. I figured I could produce something playable in a week with that, no matter what.

I wanted to have a really polished looking game this time. So, animation, sound, scoring, losing (yes, that's a feature...), particle effects and whatever else I could cram in. I wanted it to feel like a real thing. So, I tried to get some kind of mechanic working as quickly as possible, so I would have something to actually polish -- instead going all high concept and never delivering because of unresolvable bugs at the last minute (looking at you, Ghost Stars. Sigh).

And how did I get the mechanics working so quickly this time? Canned Heat. I don't feel you can really call it a "game engine" in the Unity or Godot senses of the word, but... for the first time, I wasn't having to work out fundamental architectural questions. The basic game issues were taken care of, which added significant velocity.

Then there was the punch list. Just a simple list, an org-mode file listing the features I wanted. I did not work on a change/bugfix/feature until I had added it to the punchlist (mostly). This helped tremendously in keeping my focus as time ticked down. It's probably the only reason the credits page has fireworks...

Lastly, I got a lot of mileage out of just implementing things and not worrying about what else they would break, or how scalable they'd be or anything like that. Trying to inject design rigor is death to momentum. I'm not trying to build a sewage treatment plant or an airplane here... this is at best a prototype for a game, so anything goes. And often just doing it was easier than I thought it would be and caused less problems than maybe it should have.

What went wrong...

I avoided a lot of the traps I'd stumbled into before, so that was good.

Probably the biggest problem I had was continuing to work on something past the point it was unproductive. I lost at least half a day on a NaN issue that was later cleared up in minutes with the benefit of a night's sleep. I have to keep in mind that if I stall more than a few minutes on something, it's time to switch context and do something else! There was so much else to do, too... Oh well.

All in all, though, pretty happy with the results. Now that I've proved my architecture, next time I'm hoping for something more ambitious!