If I run through the entire thing each time with an algorithm that finds the big block, some something as simple as if it finds a # it looks for another # catacorner and goes from there (relitivly simple), operating from top to bottom (as an example). If I had a long short one on the bottom say 4x2 on the last row and I started putting blocks one top of it and it then looks like this:
# # <- both newely added
# # # #
# # # #

then the algorithm would make it two blocks one 2x3 on the left side and a 2x2 on the right. Where it should rightfully be a 4x2 with 2 individual blocks on top. Thats why I can't get past the idea that its gonna need to remember the big blocks positions from the previous loop.

What I'm thinking about is only checking the newly placed blocks cause nothing else will chenge from the previous screen. Somewhere along the lines that if it lands on a big block of its # then it checks to see if it has neighboring #s of its type to make the big square grow or if it makes a new block that has to be two by two...
I'm gonna be busy on homework projects for a while from school. But I'll try to still discuss theory about this with anyone who is willing to listen to me blaber.

NOMAD