Michael Camacho wrote: > I've noticed that (using 'foo' ;)) if I enter a sequence of moves, and then > work backwards and re-enter one of the previous moves (as if I was entering > a variation), then that move will appear as a variation, meaning that we now > have two variations that start with the same move. Ah yes, that behaves different in gGo and glGo. gGo will enter the variation if you play the move again, glGo will create the move as new variation. Need to fix glGo, as the gGo behaviour makes more sense. The prototype is using the glGo code, so the bad behaviour occurs. Good point. Another thing I am pondering about is how to make the tree display prettier. The black/white circles are somewhat ugly. I like the CGoban2 display very much with small stone graphics. I suppose graphics slow down the tree significantly, but probably not so overly much it becomes unusable. I didn't bother about looking pretty so far, as I concentrated on functionality. On the other hand, a pretty display can be done later after the tree is in gGo/glGo, as that code will be Java/C++ specific so cannot be reused from the prototype anyways. Peter