HERE BE (TEMPLATE) DRAGONS
The two best things I've done in the past three years:
- Iterate like a mofo
- Evaluate like a mofo
I've been doing a fantastic job of banging out minigames, and conducting my own efficiency analyses to find the most obvious pain points in each project, and identify how to mitigate them. You'd be amazed how much you can find wrong when you actually go looking. I will typically keep a hit list as I go through a project, noting potential spots for improvement (along with the occasional /* this is jank, plz fix */ comment, directly in the code).
To that end, I've taken to making a yearly 4x template, so I can dive into a project with minimal effort. I make incremental improvements per-project, but I save all the big refactors for the yearly templates. I just finished my 2025 template, and have made improvements to how shaders are organized and handled per-actor, made the transition between desktop to web deployment as seamless as possible, and also made switching between mouse and keyboard/gamepad interfaces painless. I also put in some quality of life improvements for making and maintaining actor lists, and finally got around to rolling in the various improvements to the MIDI interface from Vertical Hold.
The idea was to get from "new project" to "I have something on-screen" in a matter of seconds, as opposed to minutes or hours. In the case of my "big" framework, there are so many possible program states that it's potentially mystifying if nothing shows up on screen. The 4x stuff is lean and encourages hacking. I have to fight the urge to roll everything into it. It doesn't even feature an established event pipe.