[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] Re: SPAD and Aldor again

From: Gernot Hueber
Subject: [Axiom-developer] Re: SPAD and Aldor again
Date: Mon, 20 Nov 2006 09:49:43 +0100

Hi there, With great help from Gregory, I have started to implement a wrapper library to access GSL (GNU scientific library) functions from Aldor/Axiom. I use the Lisp approach with dynamic loading of a shared library. Yet some of the functions work fine - but there are far to many of them to cover them all for the moment. Furthermore, there are some unresolved problems (e.g. calling a Axiom/Aldor function from the library code). Gernot Gabriel Dos Reis writes:
C Y <address@hidden> writes: [...]
| I guess it depends on the details of how such things are handled.  You
| are proposing to have code at the SPAD level talk directly to things
| like external libraries?
My proposal is to formally specify a way for SPAD codes to talk to
external libraries.  I don't think such an interface must necessarily
be through Lisp first.

Pls tell me more.

| > All of that, including interfacing with nay reasonable language used
| > in the computational science community -- that list goes beyond
| > Fortran and C.
| | Indeed. C++, Java, Python, Haskell, CAML, ML, various proof
| languages... more I'm sure that aren't leaping to mind.  You know more
| about those things than I do Gaby, so perhaps the difficulties are less
| severe than I am imagining, but I was under the impression that
| translating from one language to another is highly non-trivial.
I'm not saying it is trivial.  However, if we must attempt only
trivial things there is little hope to make Axiom interesting.  We must be
ambitious and approximate the ideals.

agree, but things have to be practical, too.

| Particularly if the program is written in such a way as to assume
| communication only with other Java/C/ML/etc. programs.  Lisp's FFI
| systems are probably among the most general solutions to such problems,
| and even they have a host of issues.
Lisp interface may be a proof-of-concept, but my belief is that it
should not be the final answer.

-- Gaby

Axiom-developer mailing list

reply via email to

[Prev in Thread] Current Thread [Next in Thread]