On Wed, 2003-10-08 at 02:37, brogoff@speakeasy.net wrote:
> While it's nice to have stopgap solutions to these "problems", IMO the right
> solution will be possible when we have some kind of overloading in the language.
> Rather than having lots of squirrelly notations for hashtable access (and string
> access, and BigArray access, and...) we should just be able to index like arrays
> and be done with it. Lack of user defined overloading has always been a weak
> point in the entire ML family of languages, IMO.
> 

Felix has overloading. My feelings: lack of overloading
has two advantages.

(1) definiteness

Its definite what you're refering to: the lookup rules
are simple. Unlike overloading systems such as the lookup
and overload resolution in C++.

(2) prevents newbie abuse

Well, I have seen so much *horrendous* C++ garbage
where people thought overloading was clever.

I have also seen bad consequences where more expert
people constructed a badly designed mess -- such as the
C++ iostream facility. They got confused, and tried
to 'overload' cout << x for x being a 'character'. Only,
when iostreams got templated (character type became a type parameter)
it no longer made so much sense .. 

Yes, I miss overloading in Ocaml. But not all that much :-)


-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners