Crosscut

After the latest improvements to the debugger, I've been able to make significant progress on the game side. There's basic user input; there's game state, updated in an ongoing simulation that considers input; and there are simple graphics to display all that. What's still missing to complete the game, is some kind of challenge, and win/lose conditions.

This is definitely the most advanced thing I ever wrote in any version of Caterpillar (although the language itself was more sophisticated in earlier prototypes). Which further strengthens the hypothesis, that a debugger is both practical to implement, and sufficient to tame a very basic language that would otherwise be too confusing.

Keeping the language so basic, was a decision made to keep the current prototype manageable. But I'm now considering to deviate from that plan. Working with Caterpillar was an interesting (if tedious) puzzle at first, but it's starting to grate on me. I think putting some work into the language itself might be a good investment right now.