Re: OT: Languages and type systems (was RE: Settlers ...)

[Adrián Pérez <copylight@...>]

El 20/11/2004, a las 18:40, Andrei de A. Formiga escribió:
>    I don't find it surprising that people think
> statically-typed languages are so unconvenient,
> because they are exposed only to C, C++, Java and the
> like. [...] The magic is called "type inference" and is a
> technique known since the 70s.

Yeah, "dynamicity" started with Lisp a long, long time ago...
(but Lisp is not compiled to machine code).

>    Interested parties should look at OCaml [...]

It's a very interesting option, we use it regularly at the
University and there are lots of people that prefer it over
C, C++, Java and Pascal, for example. It's a very nice
OO functional language.

Moreover, if you want all (and more of) the introspective
features of C# but compiled to machine code, without
abandoning C heritage, you surely want to take a ride
with Objective-C (a beast that mixes C and Smalltalk
and compiles to machine code). It has two major
drawbacks I can live with:

  - No garbage collection directly in the language. It's
   pretty straightforward to add support for the Boehm's
   garbage collector (used i.e. by Mono and GCJ).

  - No direct support for exceptions in the language.
    Well, Apple folks patched GCC 3.3.3 so it has ObjC
    has exceptions, but it's _not_ standard yet. Other
    way of adding exceptions is via setjmp()/longjmp()
    tricks.

-ap

Attachment: PGP.sig
Description: Mensaje firmado digitalmente