It is proving more difficult (actually more monotonous) than I had anticipated but the results are well worth it. On the fly calcs of a tile's neighborhood as it is being painted and the creation of lookup tables for the possible combinations is slow and prodding but I am about 85% there. Maybe another week or two. My biggest problem is just to get it down and stop tweaking it. I will have a really good cut of it and then think of a better way to do it and then spend the next night changing what was working just because it is cleaner. I have actually been spending a lot of time looking into hills & mountains. At first I was just going to have precomputed heights but VertexNormal has a really cool demo app which I have been studying and will try to emulate. His growth algorithm is amazing. I need to ping him to get some more info on how the heck he does it. My hat's off to him.
