Alex Ferguson <[EMAIL PROTECTED]> wrote in reply to Jeff Lewis: | > ... I can't think of a good rationale for it (aside from the "don't break | > old code" argument, which, if that's the only argument, doesn't seem | > strong enough to me). | | For Standard Haskell at least, it ought to be a pretty strong argument. I presume that the current discussion really pertains to Haskell 2 rather than Standard Haskell; assuming this, I would like to make a plea: PLEASE feel free to break old code!!! The dismal decisions that I have seen made in the C++ standardisation in the name of preserving compatibility with old code make me cringe (I have to use the results). Surely there is no need to reject possibilities for Haskell 2 just because they break old code. (I presume that Standard Haskell is a different story, but then tweaking the type system isn't up for that!)
