On Wed, Jan 23, 2008 at 03:10:11PM -0500, Jonathan S. Shapiro wrote: > On Wed, 2008-01-23 at 19:52 +0000, Sam Mason wrote: > > Haskell has a nice feature where any sequence of punctuation characters > > is treated as an infix operator. > > Umm. The term "nice" is a value word. It is certainly preferable to > going without oxygen for an extended period of time, but not by much. > > This particular "feature" of Haskell and ML renders (human) program > analysis near as dammit impossible when it is actually used. Programs > using this feature extensively are completely immune to inspection. If > you don't believe me, pick ANY input file in the Isabelle source base > that you don't already know, and attempt to explain what it does without > reference to other files.
OK, I was thinking about the language being used for different things. I'd not thought about the whole validation/inspection issue that's going to be a major part of BitC. Is this a problem because symbols are more opaque than names? > > The custom operator stuff is great because it means the standard library > > isn't special and can be more readily extended. > > It merely means that the parsing of certain tokens can be extended. > Extension of the standard library is neither enhanced nor reduced by > this feature. humm I think I mixed things together in that sentence that shouldn't be. Maybe it's better described as the boundary between the standard library and user code is less obviously defined. For BitC this is probably a negative point. > > The dynamic specification of infix associativity is probably pretty > > tricky to implement, though it does improve expressiveness. > > Actually, it's *very* easy to implement. For some reason I thought it would be easy to implement, but thought I must be missing something so wrote it as tricky. Ah well. Sam _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
