RE: [Axiom-developer] Axiom Lisp API for OpenMath

[Page, Bill]

On Thursday, October 13, 2005 11:46 AM Camm Maguire wrote:
> 
> Do I take it that this entails making lisp functions with the
> following names in some appropriate package with the functionality
> of the C code listed adjacent to them?:
> 
> setup_type const om_setup[] = {
>       /* LISP Name */                 /* Unary */             
>       /* Binary */            /* Nary */
>       {"om-openFileDev",              wrong_no_3a,            
> wrong_no_3b,          om_openFileDev},
>       {"om-openStringDev",            too_few_2,              
> om_openStringDev,     wrong_no_2},
>       {"om-closeDev",                 om_closeDev,            
> too_many_1,           wrong_no_1},
> ...

Yes, sort of. What is required I think, are lisp bindings to
the existing OpenMath version 1 library (freely available and
written in C). We have a prototype for these bindings written
for ccl lisp :(of course not directly usable in GCL).

Some of these functions look a little peculiar and specific
to Axiom's internal representation, for example there is a
function that returns a pointer to an Axiom string which is
used as a global buffer to accumulate OpenMath output. But
if this worked once (albeit with ccl lisp), it seems likely
that it can be made to work again.

As an alternative there is a freely available XML/OpenMath
library written in Aldor. But interfacing the current Axiom
OpenMath functionality to this Aldor library seems likely to
be more work then just getting the OpenMath version 1 library
to work again, even if the end result might be more maintainable
in the long run. Version 2 of the OpenMath specification is
already available, but according to Mike Dewar there are
currently no plans to make a version 2 of the OpenMath library.
Having the library available in Aldor would be an advantage.

Regards,
Bill Page.