WhiteTower: In the end it's just a system of accepting mouse clicks on certain images. The hard part is not the image map itself, but the fact that every possible board layout would need to be implemented, instead of every combination for a particular square. For this reason I think we need a board that can be drawn by joining parts.

Perhaps turning the board slightly to align the hexigons, and then making the "active" part of each square in the middle? Or something like that.