Re: [Texmacs-dev] plugins

[Pablo Angulo]

Andrea, I'm not sure I understand your motivations, but here are some
comments:
> I think that the possibility to have a common open source interface to
> different mathematical packages is an awesome feature of Texmacs,
> something really unique, but a stronger effort is needed to make it
> work reliably, taking also into account software upgrades, etc.
This is the task of the package system. The only step required is to
package plugins separately and state their dependencies. Usually this is
not necessary, because the same texmacs plugin works with different
versions of the software.
> In principle it would be nice to have some degree of unification, for
> instance one should be able to copy and paste mathematical expressions
> between different mathematical packages.
That would be a huge effort, may be impossible. Do you know about SAGE?
They've done this work for many free libraries, which are now called
using the python language that SAGE relies on. But trying to do that for
the software you describe is a different thing. First, that software is
non-free, so you can't distribute it along with your software, so you
cannot rely it will be there. Worse, you don't even know which version
of the software will be there, and the languages they use all evolve.
Sometimes they use really different aproaches and they have different
capabilities, so they cannot be mapped in a way as simple as the one you
describe with the definite integral of a function of one variable.

If you're talking about something more restricted, may be it could work,
but then it wouldn't really be that useful. You said you want to check a
computation, but that way you introduce a new source of errors, namely
the texmacs translator, working on a specific version of texmacs, and a
specific version of the software being used. This is difficult to debug.

May be a project like openmath could help, but even then you're talking
about a considerable effort.

Regards,
Pablo Angulo