Caterpillar

I need to come up with a way to reduce the initial scope of the algorithm that handles code updates as much as feasible, in a way that allows me to refine it iteratively. Reason being, that I don't want to waste any effort on making it better than it needs to be, before it can be replaced by something better.

I'd like to talk about my idea for approaching the code update algorithm. But before I can do that, I think I need to lay down some basics. First, the context that the update algorithm exists in: Its output, and how that output is going to be used. Then, some details about how code is represented in the Caterpillar compiler, that should allow the update algorithm to do what it's supposed to.

The purpose of the code update algorithm is to determine which pieces of code were replaced by other pieces of code. If you have some old code, a b c, and new code, a x c y, then you'd expect the algorithm to output "b was replaced by x" and "new code y added after c", or something along those lines. Tomorrow, I'm going to explain what the compiler does with that information.