Currently I do this 40 times :
Create a BCBrush (small square terrain image) and draw on a large bitmap according to a hexagon path...
Should I do the opposite ?
Create one transparent BCBrush, draw on each square terrain image according to an inverted path of a hexagon, then draw each bitmap on the main bitmap ?
I also realize that in this case I could use the same Path values for each terrain image instead of calculating/storing/using a path for each tile's location on the large bitmap. However the performance gain would be close to non-existant while drawing.
I think I have better ideas at work than at home. Is this good for my project ?