Hello Kaveh, Sunday, August 6, 2006, 5:40:26 PM, you wrote:
I think we need a subset of haskell as a new language (or as a developing pattern) to work with and teach and learn more easily as you have mentioned.
it called Helium :) but in general problem is what Haskell's way to deal with problems is to find most general and abstract solution, such as enclosing monads just for doing I/O. it's great for programmers interesting in raising their programming languages knowledge (i usually study several new languages each year, so Haskell with all its extensions partially replaces my need in new interesting languages ;) ) but can be nightmare for 99% of programmers. after all, they need to learn many many other things besides of language tricks interesting for souls like me
I had read a text about mathematics which was something like this : "New mathematic theories does not populize because of their fabiolus logical theorems, but because of death of elder mathematicians and forging a new folk of them that were rised by new theories.".
that's true for any science (recall for example Freud's story) and moreover for any ideology (Moses drived Hebrews over the desert for 40 years just in order that Jews that born in slavery was died)
That is usefull to have in mind. (And because of that, maybe there is no force to compele someone to disturbe his mind on a peacfull friday afternoon! ;) ) Anyway the point is developing more efficient and easily; instead of serving marketing features of languages.
that is entirely different question. while Haskell is rare bird, _i_ can't use it in commercial environment. as one manager said "i can easily find 10 C++ programmers. but where i will find 3 Haskell developers?" ;)
And there must be - and will be - someone (or some will) to makes things better. "very high level of FP" in Haskell is good but there is no lower level of them somewhere else. So almost the only choice for a useful FP environment is Haskell.
if i correctly understood that you mean, there is Ocaml and FP features in many languages. most important things, imho, are anonymous closures and 1st-class functions. even current C# should support this -- Best regards, Bulat mailto:Bulat.Ziganshin@gmail.com