Michael Chaney

This. Is. So. Cool!!! The utter simplicity of the pocket canon program is why I remember it. Thank you for this. I'm still looking for the maze, but at this point I think I probably took the maze program that is in the Compute! book and made it into what I remember. It's been 40 years, time flies when you're having fun!

First, thanks for all the replies! I should have mentioned a few things: 1. I've searched extensively through online magazine archives. The program that I'm seeking was TI-specific (it used HCHAR and friends to draw the maze), although it was possibly in XBASIC. 2. It wasn't a game - it just generated the maze. It took about an hour to generate the whole screen on the TI. 3. It started at top-left and had a certain affinity for going left. The maze would start by going roughly across the top, down the right side, back across the bottom, and by then it usually had enough randomness to start being more of a maze. If I remember correctly, the algorithm just tried random directions 10 times or something like that, and then gave up and went in order checking each direction. That would cause such a pattern to emerge. 4. It seems like it was in a magazine where one wouldn't expect to find a TI program, like Byte. I might be misremembering, though. There was likewise a cool canon (music) generator one time that appeared in a magazine like that and was written specifically for the TI. I saw it, loved it, lost it, and spent years trying off and on to find it again. 5. I know *a lot* about mazes. This is me: https://github.com/mdchaney/jsmaze. That all started from the TI program nearly 40 years ago. 6. One other thing - whoever mentioned above about the 4 directions and all that - there are 16 variations of 4 walls being off/on, and it's easily stored in the low nybble and set in the character set. So 0 has no walls and 15 has all four. 14 would be the top wall open, as it's number NESW. Moving into a new cell means knocking down the wall in the current cell as well as the corresponding wall in the cell that is being entered. There's no recursion, but the only thing you really have to remember is how you entered a certain cell. That was likely done in an array. Thanks again for any help. I'm wanting to do a talk on maze generation, and you can imagine how awesome it would be to fire up a TI emulator (or, an actual TI-99/4a) and show where it all began.

Years ago, in the early 1980s, I recall typing in a simple maze generator from a magazine article. It wasn't the one in Compute's book of TI games. It used character definitions to create 16 different cell types, and so the entire maze was 28x24. I cannot find it. Does anybody remember this?