After learning lots about Axe and programming in general, I glanced over the source code for Minesweeper, and I was appalled by how inefficient and unnecessarily complicated it was in many places.  As a result, I've decided to attempt completely rewriting the program, and while I can't promise that I'll have time to finish this, I've decided to start.  So far, I've rewritten my menu system (although most of the graphics should be the same), finally created an algorithm that implements wrap-around (after promising this feature for practically forever ), created a much more optimized stack-based algorithm for the expanding bubble thing, and shifted code around in general.

However, I have several questions:
I was considering adding an option that lets the user use nearly all of their RAM so that they can make massive levels if they want.  Should I do this, or just stick with using the amount of data in L1 instead?  If I do allow this to be a feature, should I leave it as a permanent feature, or should I let the user turn this option on or off?

Besides toggling question marks on and off, toggling wraparound, and toggling showing the positions of the mines when you die, should I add any more options?

What else should I add to make Minesweeper even better?

After learning lots about Axe and programming in general, I glanced over the source code for Minesweeper, and I was appalled by how inefficient and unnecessarily complicated it was in many places.  As a result, I've decided to attempt completely rewriting the program, and while I can't promise that I'll have time to finish this, I've decided to start.  So far, I've rewritten my menu system (although most of the graphics should be the same), finally created an algorithm that implements wrap-around (after promising this feature for practically forever ), created a much more optimized stack-based algorithm for the expanding bubble thing, and shifted code around in general.

Same here with Simul and all that stuff Axe's just so fast that you don't notice if it's really, really unoptimized.

For your giant maps, I suggest creating an appvar buffer only if the map won't fit in L1. You can check for this at runtime. Store your map offset to a var and use that var everywhere you would normally use L1. That's what I did for cUbEs.

THey would be compressed upon quitting, right? Otherwise that might be a bit slow if they're decompressed everytime they're drawn. X.x

By the way should I create a sub-forum for this?

Hmm I see. I hope you aren't planning to quit calc stuff anytime soon. it would suck to see you go so early.