I'm wondering if the code for attribute-based collision could be modified to be character-based instead.
I want to implement something like that to my game.
How easy would that be to do?
Screen characters are formed by a collection of 8 bytes, 256 bytes apart from each other. One simple way is to have a "character map" of 768 bytes (32x24) in memory to store the characters at each screen location, then the colour attributes would not matter.
I was gonna make the collision system in my game work by detecting the tile graphics themselves and their positions. (for example, a dude moves if there is 8 pixels of space next to him, and can't move any further if there is a brick tile right next to him without any space inbetween)I would need to see some code example to know how to make it.