how is it much different than a regular collision map
Take a look at the collisionmap in the example. Zoom in a few times. You'll see that it's one pixel per tile. So one tile on the normal map equals one pixel on the collisionmap. If you take a closer look you see there are several different colors used (Red, Black, Yellow, Blue). The red
pixels are the tiles where you can
come. The pixels with any other color than red are the tiles where you cannot
come. If you use the action Move RPG Sprite, you'll notice that the first argument is a variable. This variable will be set to the color of the tile he will be going to. So if the variable equals PA_RGB(31,0,0) , which is red, the sprite is able to walk in that direction. If you use different colors and he bumps against a, say, yellow
tile it knows he can't go there, but you can also say:
If variable is equal to PA_RGB(31,31,0)
So you could put yellow
pixels on a door's house. That'll mean that If variable is equal to PA_RGB(31,31,0) you go into the house.
you could easily adapt this into a platformer engine right?
Not really, I mean this engine works in tiles, a platformer doesn't use tiles at all.