RE: [Axiom-math] TexMacs Windows Axiom<->Maxima incompatible (DLL?)

[Bill Page]

On January 25, 2006 2:50 PM Herb Martin wrote:

> ...
> it seems to say that tm_axiom is:
>
> http://wiki.axiom-developer.org/TeXmacs
> 
> "A new version of the Axiom interface program 
> tm_axiom that was written especially for Windows 
> is included in the latest Axiom for Windows 
> installation program. This will be used 
> automatically by the native windows version of 
> TeXmacs."
> 
> I took "interface program" to mean that which
> will display the output (i.e., what TexMacs
> presumably does) and apparently (now) it really
> means "the program that TexMacs talks to..."
>

Yes, that is correct. The information flow looks like
this:

   TeXmacs  <---> tm_axiom <---> AXIOMsys

tm_axiom is a two-way translator (interface) that
converts to/from TeXmacs compatible input/output
from/to Axiom. When you start an Axiom session in
TeXmacs, TeXmacs runs tm_axiom. tm_axiom in turn
runs AXIOMsys. These three processes are running
on your system when you interact with Axiom via
TeXmacs.

Could you suggest a revision to the text on the above
web page that would make this more clear?
 
> Or else there is NO connection between TexMacs
> and Axiom (except through manual transfer)?????
> 

The connection is automatic but slightly indirect.

> On an old message there was a statement that said
> that the drawing did NOT work on Windows -- is this
> still true?  NO drawing?
> 

NO drawing on Windows :(

The problem is that Axiom's graphics package depends on
the Xwindows environment which is not available on native
Windows.

Xwindows is available under Cygwin on Windows and in principle
the graphics and hyperdoc parts of Axiom could run in this
environment but right now we do not have an official build
for Cygwin. Further, the version of Axiom for Windows does
not have the same socket functionality as the version for
Linux (due to limitations in the underlying lisp - GCL).

But some progress has been made recently. Tim Daly was able
to compile (using Cygwin) and run a stand alone version of
hyperdoc (in native windows with a native windows Xserver).
This is still very experimental but in principle would also
allow Axiom graphics to run in native mode Windows (i.e.
without installing Cygwin) once the socket issue is resolved.

> This page indicates that TexMacs can use Axiom and
> a variety of other programs as "plug-ins": 
> http://www.texmacs.org/tmweb/plugins/cas.en.html#axiom
> 
> Likely I haven't done whatever is necessary to hook up
> TexMacs to Axiom -- or else I misunstand what they
> do together.
> 
> In general, I was expecting something like the Windows
> version of Maxima-TeXmacs.  It this completely wrong?

This is exactly what you should get if everything is
installed properly. On Windows everything works best if
you install Axiom first and WinTeXmacs second - that way
WinTeXmacs will "see" Axiom and setup the necessary
configuration.

But the Axiom plug-in for TeXmacs is not quite as
sophisticated as the most recent version of the plug-in
for Maxima. For example, it is not possible to use "2-d"
math input for Axiom, you are limited to entering
traditional Axiom commands.

> > 
> > Why do you say they are "apparently incompatible"?
> > WinTeXmacs should work with Axiom for Windows "right
> > out of the box".
> 
> I think it is (only) the TexMacs stand-alone
> load that is incompatible with the Maxima built-in
> TexMacs (likely one of the required DLLs is using
> the same name but a different version.)
>

Could you please describe this again very carefully, i.e.
step-by-step? I do not understand what you mean by:
"Maxima built-in TexMacs". Maxima is not built-in to
TeXmacs. It works that same way as Axiom and TeXmacs
except the most recent version of Maxima no longer needs
an intermediate interface program like tm_axiom.
 
Regards,
Bill Page.