W3C home > Mailing lists > Public > www-math@w3.org > February 2005

Re: Semantic information for math representations of physics

From: William F Hammond <hammond@csc.albany.edu>
Date: Thu, 10 Feb 2005 13:03:41 -0500
To: www-math@w3.org
Message-ID: <i7fz04b8ci.fsf@hilbert.math.albany.edu>

JB Collins <joebmath@yahoo.com> writes in part:

> Regarding the TeX/LaTeX issue: 

TeX is a programming language.

LaTeX is a structured typesetting language with a layer of
abstraction.

When used as Lamport intended, it can be rather close to semantics.
The problem is that many do not begin by reading Lamport but by
copying the examples of others.  Without reading Lamport one can fail
to understand LaTeX's structure.

> I think that new standards rarely supercede old ones
> by completely bypassing them. The value of standards
> is in creating efficient infrastructures. Superceding
> an old standard means a whole infrastructure must be
> modified or replaced, and that's hard to do without
> grafting the new standard to the old in some way, if
> only temporarily.  Content/ semantic representation is
> a higher level of abstraction then typesetting, and
> TeX/LaTex is principally typesetting oriented. I don't
> think that re-defining TeX/LaTeX so as to directly

Let's say "LaTeX" here rather than "TeX/LaTeX".

> incorporate semantic representations is necessarily
> the final answer. But I wouldn't reject it outright,
> either. A pristine standard without users is not
> useful. Getting users to abandon their familiar tools
> is very difficult. For example, physicists still use
> Fortran - they don't want to take a few weeks off to
> learn a better language. (Last I heard, the best
> compilers for fast code were still Fortran compilers).

The Fortran/LaTeX analogy is good on the question of breaking old
habits but weak in the matter of needing to get something new out of
something old that was not so designed.

> That being said, it seems to me that OMDoc having a
> LaTeX output format may be an adequate answer to this
> concern.

That would not address the problem of meeting the users where they
are.

Another approach would be to ask how much needs to be added to gellmu
article in order to have OMDoc output.  I've not looked at this
question.

                                    -- Bill
Received on Thursday, 10 February 2005 18:03:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:57 GMT