Dear Terrence,

Terrence Brannon wrote:
> >We should never forget that programming languages are just tools, not
> >religions.  
> pshaw! 
[...]
> 
> oh yes _religion_ , not deep black magick. I'm sorry you're right. :)

:)

> I cant give up the interactivity of Lisp... it is so easy to experiment
> with.

Maybe with http://haskell.org/haskellwiki/GHC/As_a_library or with external
calls and dynamic linking it will be possible to do practical runtime native
compilation with Haskell in the future. Then some Haskell version of Slime for
Emacs could be fully supported. For the time being, I'm going to use the
interpreter ghci for interactive development and the compiler ghc, which is a
bit slow, for longer calculations and deployment.

> To me, Haskell suffers from two models of computation - monads and
> arrows. The predecessor of Haskell, Clean, offers uniqueness typing instead
> of these two models.

Yes, I heard about that. Unfortunately, Clean seems to have too few libraries
(it doesn't have a good foreign function interface) and a very small user
base. I'm still undecided about which of the two approaches to non-purity is
better.

> And I met you on the Maxima list Bruno. I dont think there is nearly as
> powerful a math package for Haskell as Maxima.

Yes, while there are some attempts to write computer algebra libraries in
Haskell (see
http://www.haskell.org/haskellwiki/Libraries_and_tools/Mathematics#Computer_Algebra
) they are all still in a research stage compared to Maxima. Oh well...

> On the other hand, I never savored having to have my program 
> type-check completely before running it.

I always used some type declarations as assertions and for efficiency in SBCL,
so I don't feel that much of a difference. With Haskell's type classes and
type inference, it's not all too much pain to have everything type-checked.

Best regards
  Bruno Daniel
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to