Re: Validation of MathML

"Pepping, Simon (ELS)" <S.Pepping@elsevier.nl> writes:

> > I would think it to your advantage if they can perform validation
> > before shipping to you.
> 
> I do not think an author will create an article according to this DTD. Their
> manuscripts will be converted to it by our suppliers. The manuscripts may
> contain MathML though, presentation or content.

One reason many math authors have been preparing their manuscripts in
LaTeX since the late 80's has been eliminating for them the
inefficient step of proof reading and correcting a technical
stenographer's rendition of the author's handwriting.

In the same vein it is rather expensive for all concerned to have a
markup contractor between author and publisher.  Suitable markup for
authors that admits free-standing validation should make it possible
to eliminate the intermediary.

DTD validation in my experience traps most human errors.  I imagine
that a successful (and noiseless) translation with a public tool at
the author's site into a publishing document type would be a sensible
finer level of validation.  (I also imagine that the editor would want
the source rather than the output.)

MathML islands in document types such as DocBook or TEI do not meet
the need because the islands are too inefficient for human authors.

It is a small step from writing LaTeX to writing in a suitable XML
document type -- never mind that it may not quite yet exist -- that
can be robustly handled by publishers upon direct receipt from the
author.  Such a document type will have math markup that is suitable
for human authors while being "strong" enough to admit automatic
translation to content MathML.

                                    -- Bill

Received on Monday, 3 March 2003 12:26:24 UTC