> It would also change the meaning of existing valid programs such as: > > x = 1, > y() This is the strongest argument against the idea that I've seen so far. It could be solved by *not* treating , as an operator, and by keeping the open bracket rule.