I don't really feel like I learned anything, but it's probably because I worked super intensely during one day and then dropped the project... no time for introspection / retrospection. Most of the code I wrote without access to GM, then I copypasted the script into objects / scripts once I got home, so there was no time to tweak and polish. It let me achieve a lot of stuff in a short time, so maybe I'm test playing too much and focusing too hard on game feel when I can just test every line of code before proceeding?
I implemented exactly 2 changes to my orginal ideas:
- Animals originally had a value based on the number of letters in their name. I had to cancel this because some puzzles needed 1s or 2s and I couldn't come up with any animals with names 1 or 2 letters long.
- I hadn't originally planned the "unknown" operators, but I realized hiding information from the player and forcing them to experiment a bit essentially would fit pretty well with how they'd need to tackle some of the more obscure operators (multiplication and division doesn't bind tighter than plus/minus in 20w0 Keeper, the square-and-add / cube-and-add / backwards square root operators aren't super self-explanatory) so it would be a natural progression in difficulty. Since I was starting to run out of ways to make the maths harder around lv.10, I decided to make 10 more levels with more and more unknown operators.
I already knew in beforehand that you get the best ideas when you test play a game, seeing what of your original ideas works and what doesn't, but it's always good with a refresher.