RE: [Axiom-developer] Terms of Surrender (was: BAD tim)

[Bill Page]

On November 6, 2005 2:22 PM you Tim Daly wrote:
> 
> > > i'm advocating literate programming and have likely committed
> > > the rest of my axiom working life to proving (or disproving)
> > > the concept.
> > > 
> > 
> > Three cheers! On this we are in complete agreement. I agree
> > 100% with the emphasis on literate programming - so much so
> > that I have a hard time understanding why you feel that it
> > might be necessary to prove it.
> 
> how much of mathaction is developed as a literate program?
> -- qed
> 

Touché!

You are absolutely right. This is something I need to fix.

Much of mathaction itself is directly intended to support
documentation and a lot of the design of mathaction is
already described on the web site. For example:

http://wiki.axiom-developer.org/MathAction

But this is not a literate program. It needs to be merged
with the code itself which is available in the repository:

http://wiki.axiom-developer.org/MathActionRepository

MathAction currently amounts to about 200 lines of Python
code that is added to the LatexWiki code which in turn is a
patch for the ZWiki application in Zope. Excluding Zope, that
probably amounts to at most a few thousand lines of code.

Bob McElrath (the maintainer of LatexWiki) and I have been
discussing off the list the possibility of forking a version
of ZWiki specifically to support LaTeX and the mathaction
extensions for Axiom. One reason for doing this of course is
that LaTeX support is not the kind of main stream feature
that appeals to most wiki users, but to do it properly seems
to imply some fairly deep changes to ZWiki. I have also made
a number of changes on MathAction to ZWiki itself in order to
better support it's use with Axiom, e.g. the preview for
comments. Additionally Bob has expressed some misgivings about
the current copyright policies of the ZWiki developer visa vi
Bob's own view of free software.

We are both a little reluctant to do this because of the time
investments required to maintain an entire separate version,
but perhaps the time has come to seriously consider this. If
we go ahead, then I fully intend to document all of the code
from the MathAction extensions right down to the ZWiki level.

And like in the case of asdf in Lisp, we also need to think
carefully about how best to integrate the way applications
are normally distributed for Zope with the use of pamphlet
files to store Python application code. The match is not
perfect but that can no doubt be solved by the use of some
kind of make script.

Regards,
Bill Page.