After spending some time trying to rewrite some examples
of the usefullness of Existential types without using them,
I am convinced that these things really do add some expressive
power.
I guess what I am trying to communicate by these questions, is
something along these lines: If existential types are so natural
(and reading Cardulli + Wagner makes me feel that they are) then
why arent they part of the types system from the get go? This made
me think that they dont really add anything to your expressive
power after all, and being a minimalist, I sought this angle. However,
it seems they do add expressive power, so around I go! BAck to
wonderign why they arent part of Haskell 98, or even mentioned
in the Haskell 2 stuff (on the 'future of haskell' on the haskell
page.
Another reason this stuff bothers me is that ok, I am learning
haskell 98, downloaded and printed the manual, bought ST's book,
and not doing to bad. I am pretty deep into a project to do a numerical
simualtion of an atom in a light trap (I am a physicist, not a CS
by the way! ) . However, when I look for some usefull packages, damnitt
to hell if the authors almost all seem to want to use every extension
they can get there hands on: Existential types. Multiparameter type
classes. etc. (My hats of to Jan S. for producing some nice stuff
in 'vanilla' Haskell 98 btw)
Ok, I guess I dont have a point, just complaining. I know I should
expect that since this is research language, probably shouldnt
be using it for real work, but I just have a thing about FP :)
Cheers!
+++++++++++++++++++++++++++++++++++++++++++++++++
Ron Legere -- http://www.its.caltech.edu/~legere
Caltech Quantum Optics
MC 12-33
Pasadena CA 91125
626-395-8343
FAX: 626-793-9506
+++++++++++++++++++++++++++++++++++++++++++++++++
