On 3/24/07, Vivian McPhail <[EMAIL PROTECTED]> wrote:
I agree with Sven, but... What I want to push is a 'mathematically sound' numeric prelude. A proper numerical prelude should have bona fide mathematical obects like groups, rings, and fields underlying common numerical classes. It would be edifying to the student who discovered that the particular data type he is using is an inhabitant of a known class and can thus take advantage of known properties, presupplied as class methods. Reasoning and communication about programs, data types, and functions would be enhanced.
One problem with that is that the instances are often times not mathematically sound - Int and Double certainly aren't. Anyway, I agree with the general gist of this idea. However, I think rather than the Prelude, we should just be an interface to the internal stuff, a 'Builtin' module. Actually, there could still be a prelude - it would contain all the common imports (to maintain our decent scores in golf competitions :) ). I also like the idea that the addition of a module line would remove the implicit prelude import - the source code has become 'proper'. Before the module line, the programmer is probably in quick-hack/script mode. Numeric prelude-ness would be good. This brings up an issue - would the numeric classes need to be included/imported in Builtin? To avoid this, I think a bit of evillness (depending on your perspective, i suppose) is necessary - builtin functions would be stuff like AddInts and AddDoubles, not (+). The numeric stuff that provides the classes would also provide instances to wrap the builtin stuff. Hmm, just realized that PreludeBuiltin already exists... Ahwell, I think the renaming is a good idea - classes should not be involved in the builtins.
_______________________________________________ Haskell-Cafe mailing list [email protected] http://www.haskell.org/mailman/listinfo/haskell-cafe
