RE: [Axiom-mail] FW: [Maxima]

[Bill Page]

Bob,

Thanks for your comments about how to properly implement
the idea of "external algebra systems". Yes, I suppose
a lot of this could be made just "pluggable" but on the
other hand really the duplicated coding is quite short
and simple. At least based on the current sample of two,
it is dealing with the particulars of the target system
that takes up most of the code and I don't see how that
could really change. Fortunately Python's regexp features
are quite powerful and easy to use so in any case it is
only really 10 or 20 lines of specific code.

I would really support your idea to do this all in a
better way if the goal was to immediately support all
20 or so of the RosettaStone systems (such as we are
considering in Tim's "Doyen" project where actually the
number might be closer to 200 if we included all of the
Quantian tools). But that is really a different (: and
hopefully better funded/resourced than the current work :)
Right now I can only see this progressing in small
increments, so it is hard to find the time to "really
do it right"...

So, if/when you get the time and motivation to consolidate
this in the manner you suggest, then I would certainly
appreciate it.

So, back to the Maxima increment:

1) How does Maxima do LaTeX? What flavour TeX/LaTeX?
2) Can we run a command line interface the same way
we do in Axiom and Reduce that processes code in chunks
like this:

   > setup environment
   > read file1
   > read file2
   > read file3
   > ...
   > quit nicely

3) What kind of syntactical delimiters can we depend on
in the output file? For example both Axiom and Reduce
produce appropriate user input prompts which can be
used to separate and match the output chunks to input
chunks.

4) If someone can run Maxima in a "filter" mode like
this

  maxima <inputfile >outputfile

where the inputfile is structured like in 2) and
the chunks would be typical of what one might want
to put in between \begin{maxima} ... \end{maxima},
then I can take a look at the problem of parsing
the output.

5) Of course I *could* do all this again on my own, but
for my goals and purposes there are quickly diminishing
returns so I would prefer if someone else stepped up to
do Maxima and then someone stepped forward to do Yucas
etc. etc.

In fact I would say that with Axiom and Reduce, and the
Reduce interface to gnuplot, I have what I currently
need to do the things that I want personally to do with
this system - except I would really like to include Maple!
Because of damned licensing restrictions of non-free code,
though, maybe I will do that on my own personal "daughter"
system. Really I too just want to get back to simply
*using* this stuff for physics ... :(

Finally, about Reduce and licensing. Yes, it is "non-free".
Pity. But it is "almost free"... and I hope that we can
continue to put pressure on the current developers and
those marketing it, to fully release it as open source.
I have also repeatedly made the same argument to the Maple
developers. So far they do not "see the light" but they
have continued to move in the right direction.

Cheers,
Bill Page.

> -----Original Message-----
> From: Bob McElrath [mailto:address@hidden 
> Sent: Thursday, August 26, 2004 1:25 PM
> To: Bill Page
> Cc: address@hidden
> Subject: Re: [Axiom-mail] FW: [Maxima]
> 
> 
> Bill Page address@hidden wrote:
> > Someone anonymously tried the following command on MathAction.
> > Is there any interest in a Maxima interface on MathAction? If
> > so, then I would be willing to work with someone how knows
> > Maxima to accomplish this. It should be quite easy.
> 
> Yes yes yes!
> 
> Maxima over Reduce, I say, as Maxima is free.  This kind of 
> integration and comparison will help to improve all.
> 
> I'm sorry I am extremely busy with physics projects right now,
> but I still want to consolidate your work on external algebra
> systems.
> 
> Basically you've been copying my ReplaceInlineLatex.py and
> latexWrapper.py for each new algebra system.  I would rather
> have one ReplaceInline... which does regex matching and HTML
> stuff (there is much code duplication currently) and dispatches
> to some kind of class which just gets handed the code and
> returns html.
> 
> Also someday I plan to plug this all into the StructuredText
> class, which will give it more rigor.  (Since zwiki and latexwiki 
> are a pile of regexps, it's not too hard to make all this fail)
> I can identify \begin{...}...\end{...} as a block in a
> StructuredText module, and hand it off to a plugin to render...
> See my StructuredTextLatex.py which is a start in this direction.
> 
> But as you have been doing all the work on this, what I wan't
> doesn't really matter, now does it!
> 
> Thanks for your work!
> 
> Cheers,
> Bob McElrath [Univ. of California at Davis, Department of Physics]
>     
>     It is unpatriotic to question the Kleptocracy.
>