RE: [Axiom-developer] Openmath, Singular

[Bill Page]

On October 3, 2005 2:03 PM Michael Brickenstein wrote:

> I am very sorry for not expressing me more clearly. I hope
> nobody feels offended. I appreciate the work of all axiom 
> Developers. It was only some weakness of English language
> (in German you say "Dein"  for Singular and "Euer" in Plural
> as translation for "your").> ...

Nema problema. One would expect that the developers of English
would have corrected such bugs by now! ;)

> 
> For connecting Singular to MathAction directly via Python:
> This will not be possible. Singular is not a library, so the
> only thing possible for me was writing some embedding not an
> extension, which is a subtle difference, but means that Python
> runs inside Singular and not Singular inside Python.

Ok. Similarly, Axiom does not "run inside Python" but is simply
called as in external process communicating by stdin and stdout.
If your extensions of Singular need to call Python that is not
a problem.

> 
> I am looking forward, actually openmath itself is not so a
> difficult thing to get run, it only needs some motivation to
> implement it. There is no motivation, if nobody uses it.
> 

I agree. I think the goals of OpenMath are noble, even though
perhaps overly ambitious and (arguably) a little naïve.

> This is a point, where I attack. Additionally to axiom I am in 
> contact to Sage, and a CAS for representation theory which is
> in planning. I  also know from cocoa, that Openmath is their big
> new feature (and some other systems, for which I don't know how
> important it is for them). Of course I will further increase my
> contacts in the community. Together we should form a critical
> mass.
> 

That is a good plan.

One of the original design goals of MathAction was to be able
to support a large number of different mathematical software
with the same user interface. OpenMath goes beyond that by
providing the means (in principle) to exchange mathematical
results between different programs. I would really like to
see such a thing work, even if not in 100% of cases.

I think Axiom provides a good test environment for OpenMath
precisely because of it's very general approach to static
typing of mathematical objects. I don't know of any other
CAS which does this so consistently.

Regards,
Bill Page.