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

Re: Math on the web without MathML (CSS 2.1 rendering for HTML and XML)

From: Mark P. Line <mark@polymathix.com>
Date: Tue, 18 Jul 2006 11:41:34 -0500 (CDT)
Message-ID: <1294.>
To: www-math@w3.org

juanrgonzaleza@canonicalscience.com wrote:
> Mark P. Line wrote:
> juanrgonzaleza@canonicalscience.com wrote:
>>> Well, the most important feature of standards is that standards may
>>> work.
>>> Standard can be both improved and replaced. Improved when changes are
>>> minimal and changed when a new paradigm holds.
>> But not at the same time.
> Then why was ISO 12083 improved whereas MathML 2 proposed?
> Why is LaTeX being improved today whereas MathML also?

My point was that it doesn't make any sense to work at both improving and
replacing a particular standard at the same time.

But to answer your questions at face value:

ISO 12083 has an enormous scope compared to that of MathML. It's like
asking why is Java being improved today as well as the MasterCook file

LaTeX and MathML are best practice for different, possibly disjunct, types
of problems. They or their successors will continue to co-exist until we
AI guys figure out a way to do symbolic computation/manipulation,
interactive presentation and production typesetting from a single,
human-friendly input scheme (voice? handwriting? VR? telepathy? all of the
above?). Even then, they're likely to coexist in backends as conversion
targets for a long time after that.

> Reply also apply to MathML. Improving of MathML can be useful in small
> changes satisfying MathML users.
> Alternative to MathML is needed for users rejecting or not using MathML.

Okay, that makes sense. I didn't understand that you recognized the
existence of satisfied MathML users.

So you don't mean "replace" MathML in the sense of making it go away
entirely, which is what I thought you meant. You only mean "replace"
MathML by some other approach for certain tasks, which is indeed
compatible with simultaneously improving MathML for the tasks you think
it's good for. (I don't think of the latter as "replacement" but rather as

So, what tasks do you think MathML is good for?

With the understanding that you are not a native speaker of English and
that I only understand perfectly about 40% of what you write, I'd like to
ask if you could please be more clear about which of your suggestions are
intended as improvements on MathML and which are intended as guidelines
for choosing an alternative approach for certain tasks -- and especially
what those tasks are.

I'd like to understand the issues better, but I'm a complete agnostic when
it comes to the rendering wars.

>> I don't think it's reasonable to expect MathML to become obsolete in any
>> plannable timeframe, and it happens to be best practice for what I need
>> it for.
> Then simply use it. Standards are for that. But many other people dislike
> MathML and can use another stuff. What is the problem?

As I said above, I hadn't understood that you recognize the existence of
satisfied MathML users. If you do, then there may not be a problem.

> So far as I can see you are using content MathML.

Yes, because it's embedded in CellML and SBML and because I also need to
be able to discretize and simulate artificial society and other
agent-based models from systems of diffeq's.

> Neil recognized that content MathML was not so well-thought.

Had the W3C asked *me* to design an XML-conformant math content standard,
I'm sure the result would be far superior to content MathML for my
purposes. But they didn't, so it's not. But I'll live. Even if I did
choose to come up with my own language, I'd still be translating MathML
into a subset of it. The train has already left the station.

> Some people has strongly rejected content MathML and OpenMath.

That's okay. I've strongly rejected perl, C++, Windows, MS-Word, .NET and
J2EE, in spite of what the majority would have me do. I'm used to being
weird. If I weren't weird, I'd be normal. And then where would I be?

> I think that many people think that content MathML is not the correct
> way, therefore they would like alternatives.

That's fine. If a different standard takes the place of MathML, especially
in its embedded role in computational biology standards, I'll be happy to
use that instead.

-- Mark

Mark P. Line
San Antonio, TX
Received on Tuesday, 18 July 2006 16:41:40 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:38 UTC