>
>While it was well designed, I noticed a few inefficiencies. I offer them
>not as criticism of Combat, but as little hints on building better 6502
>code:
I noticed first and foremost that the code examples you provide have
better variable/label names than any of my files. What file are you
using, or is it your own creation, or a hybrid of yours and someone
else's, or what?
>Last, there's definitely code in there to flash an LED wired to bit 4 of
>port B when the game is over. Presumably Atari ditched the LED to save
>money.
I still think that it sets a data direction register of some kind.
There is a reason you can't control the tanks while the game is over,
and it's a different reason than the reason why you can't control them
when your opponent is exploding. When Player 0 explodes, Player 1's
forward velocity remains at whatever joystick position it had before,
but when the game ends, both joysticks are frozen to a centered position.
I'll try removing it and recompiling.
>Overall, Combat *IS* an example of really good programming. There's very
>little wasted code, and what's there does a lot for being only 2K.
>
>There should be plenty of time left during vblank to do AI for a one
>player mode. Does the computer get to cheat during invisible tank mode
>and know where the player is, though? ;)
Perhaps it could pursue the last known position, or an extrapolation
thereof.
Another source of time could be the overscan area. Think about it.
--
To unsubscribe, send the word UNSUBSCRIBE in the body of a message to
stella-request@xxxxxxxxxxx
