W3C home > Mailing lists > Public > www-math@w3.org > July 2006

Re: Math and MathML [forethought about rendering]

From: Patrick Ion <ion@ams.org>
Date: Sat, 15 Jul 2006 15:23:31 -0400
Message-Id: <0DA6817F-66C5-4F69-A802-F2CFD36787EE@ams.org>
Cc: <www-math@w3.org>
To: "<juanrgonzaleza@canonicalscience.com>" <juanrgonzaleza@canonicalscience.com>

Dear Juan,

>>> I understand that Lie said "a markup language that could be  
>>> presented
>>> using existing CSS properties."
>
> Lie did _not_ say, "a markup language that would be restricted to  
> using
> existing CSS properties."
>
You have quoted Hakon Wium Lie in a particular way, and as saying
that the development of markup languages often proceeds in ignorance
of or with cavalier disregard for the problems of rendering the markup
specified.   I can agree that may well be true.  The notorious case  
is the
specification of SGML long before there was anything like an  
implementation
of rendering much of it.  Then followed FOSI and DSSSL, which can
be claimed to have been kludges and were certainly very complicated.
It required millions of dollars provided by the US Defense Department to
get it all working.

However, MathML was not specified by a group that was not well
familiar with the problems of rendering markup: there were several from
TeX and LaTeX backgrounds, those from Mathematica who had set up
its rendered interface, people from the Maple and Scratchpad worlds who
also knew well what problems might be with rendering, those who made
equation editors for the most wide-spread word-processing program and
high-end tools for math document creation, someone familiar with the  
journal
production stream at the world's largest STM publisher, and generally  
people
who had seen a lot of different math publications, and more such  
examples.
It is calumnious to suggest that the WG did not think of problems of  
rendering.
We did and for several output forms too (in particular audio).  Of  
course,
it is not claimed that the resulting specification contains the last  
word
on any subject.

If the Hakon's comment is just restricted to the dictum that
 > Given that CSS existed when MathML was
 > created, I think the developers made a mistake
 > by not creating a markup
 > language that could be presented using
 > existing CSS properties.
then he is welcome to impute a mistake to the Math WG, which
presumably made several.  However, I think that
the CSS of the time was not able to arrange good
rendering of what is commonly asked for in math
documents (e.g., as much control as can be arranged
with simple TeX).  Also there was a lot of hope
in the air for a merging of CSS and XSL styling in
some way that would indeed allow the sorts of
presentational control that we had become used to
in printed mathematics.  In addition, the CSS WG had
a very big job on their hands and, though the Math WG
made representations to them over such matters as
baseline handling, no special allowances were made
for needs Math identified.  After all, math or
scientific documents are widely considered a niche
market of not very great economic impact.
[Perhaps Google Trends can support that!]

All that said, you can easily disagree with what resulted, as you  
apparently
do.  Please provide convincing examples of what you can do so much
better that all the MathML efforts should be scrapped.  But they'll  
have to be
good, work very easily exactly as you put them forward, and  
demonstrate clear
advantages to those whom you wish to convert.  I'd say you have a lot
of development work ahead of you to produce something of that quality,
but would welcome definite samples of some size to look at.

Meanwhile, it behooves the Math WG to collaborate as much as they can
with the CSS WG.  As Robert Miner points out, they share the same W3C
staff contact, which bodes well for that effort.

Patrick
Received on Saturday, 15 July 2006 19:23:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:58 GMT