RE: [Axiom-developer] Re: hyperdoc

[Bill Page]

On Friday, January 14, 2005 10:19 PM Gabriel Dos Reis wrote:
> 
> root <address@hidden> writes:
> 
> | XML is a horror show in practice. XML is standard, of 
> | course, and the standards are on the web so you can just
> | point at them. Assuming, of course, that web pages stay
> | around. And the parsers are large, slow, and full of buggy
> | corner cases. XML added a factor of 5 to the file size
> | and a factor of 10 to the processing time.
> 
> hear, hear, hear.

Nay, Nay!  :)

> 
> | XML added nothing of value. Other tools and formats do
> | it better, easier, and faster, and even more portably. 
> | 
> | I'm sure the issue can be debated and I was once a major 
> | XML advocate so I'm very familiar with the supposed "pros".
> | No longer. 
> | 
> | In short, XML is gonna be a hard sell.
> 
> well, people with extensive experience with XML and fair
> assessment are rather the exception, instead of the norm;
> so it might end up being an easy sell  ;-)
> 

Well, I do use XML and the XSL style language extensively and
I have expressed my positive attitude toward XML based solutions
some time ago in this same email list.

The major error that you are making here is simply in the use
of the name XML. XML is a meta standard that carries almost no
semantic content. The actual syntax of XML is very simple - much
simpler than LaTeX for example. But comparing XML and LaTeX is
completely wrong. The describe things at a two very different
conceptual levels. It would be like comparing nested lists in
parenthesis notation to a C program.

MathML and OpenMath are *applications* of XML, in exactly the
same way that HTML (more specifically XHTML 1.0) is an application
of XML. In fact XML is already very widely used on the web.
I completely disagree that parsing XML is inherently more
complicated that parse lisp list structures since that is exactly
what millions of browsers around the world are doing right now.
Of course there are issues with particular XML libraries etc. but
that is besides the point.

Now, saying that XML is a good thing is certainly not the same
as saying that MathML is necessarily a good thing. I agree that
there are good and bad aspects of MathML and I do have some of
the same reservations about OpenMath as expressed by other people
here - specifically the near impossibility of trying to achieve
the overall design goals. At best I think OpemMath is an interesting
experiment, but I do think that MathML is hear to stay just as
the XML-based scalable vector graphics (SVG) is likely to remain -
because these are the only web-compatible formats for a very
important class of things that advanced web users want to display.

So I am in 100% agreement with Bob McElrath that support of MathML
in the LatexWiki/MathAction web environment should be an immediate
goal. And using SVG for graphics display would also be real nice
but I think available appications for this are still a little rare.
These applications of MathML and SVG are internal and completely
transparent to the user. All they see is better quality display
of mathematics and graphics in their browser.

The issue of whether to resurrect the OpenMath work that was done
by NAG in Axiom is still an open question in my mind. I think it
is very much still a research topic and it would depend on whether
anyone chooses to takeup Axiom for research in this subject.
Compared to some other existing implementations, I think that
for research purposes Axiom's model of mathematics is potential
much more interesting because of the generality that dealing
explicitly with types implies. I recall that there were some
enquiries about 1 year ago on this email list about the availability
of the OpenMath code for Axiom, perhaps now we are in a better
position to offer Axiom as an interesting research platform in
this regard. But personally I would place this rather low down
in the list of priorities at least until someone steps forward
agressively to make use of it.

Regards,
Bill Page.