Re: [Axiom-developer] Re: [ANN] new version of axiom mode for emacs.

[Bill Page]

Quoting Ralf Hemmecke <address@hidden>:

> On 05/23/2007 06:59 PM, Bill Page wrote:
> ... 
> > I think there should be a sub-directory in
> > the Axiom source distribution that contains User Interface
> > code, e.g. also the customized 'tm_axiom.c' program and
> > the associated Windows version for TeXmacs. 
>
> Oh, Bill, your tm_axiom.c cannot go into the archive. It's not
> a .pamphlet file. ;-)

You are right, I am looking for some motivation here. :-)

> But honestly, tm_axiom.c maybe better live with the TeXmacs
> people, although we then probably lose a pamphlet style of it... 
> and I would really like to see that file documented. 

I disagree. I think we should do what Andrey Grozin first
suggested and actually implemented for Maxima - that is
to eliminate the interface program altogether. The protocol
required by TeXmacs is actually pretty simple (unfortunately
it not as well suited as say something based on XML) but
still easy to produce and parse in spite of the odd control
characters choosen as delimiters. In the case of Maxima
what was done was to provide a small set of customizable
delimiters that are added to the usual text-based user
interface. This is enough to allow Maxima to interact
directly with TeXmacs when started with the right options. 

At one point Andrey stated his intentions on this email
list to produce such an extension for Axiom so that it
too could interact directly with TeXmacs. 

Of course it is tempting to try to address the more
general question of how to define a proper application
program interface (API) that would permit almost any
program to interact directly with Axiom. This was done
for example by MapleSoft for recent versions of Maple. 
The Maple API library is published and is in fact used
for their (infamous) new Java-based GUI. Maple users
are also encouraged to use this interface (in Java, C,
or whatever language binding they like) to write their
own customized interfaces for special purposes. 
In spite of the difficulties of calling Lisp programs as
an external library, something like this could also be
provided for Axiom. (That was one possible aim the
original AxiomUI Google Summer of Code project that
unfortunately did not materialize.)

But for now I would be quite happy just to follow Maxima's
lead and get rid the 'tm_axiom' program by making just
simple extensions of the Axiom text-oriented user interface
so that it is more friendly to external programs that wish
to send commands and parse it's output. 

> ... 

Regards,
Bill Page.