Re: Exploring new vocabularies for HTML

> Omitting <mn>, <mo>, and <mi> for ASCII seems easy and obvious.

It's not at all obvious that breaking the existing  mathematical toolset
which has standardised on MathML is a good thing. Nor is it obvious that
in the 21st century building in a feature that only works for ASCII is a
good thing either.

> Are there any other error cases that we could use to imply markup? For 
> example, how about making the second and subsequent elements of an <mfrac> 
> always be wrapped in an <mrow>?

A fraction with three children is just a structural error. Why have an
arbitrary rule that distinguises the first child, rather than the last?
This just encourages people to produce broken markup. You want to be
able to step past the broken expression, flag it somehow and render the
rest of the document, just silently guessing a probably incorrect
mathematical structure makes no sense to me at all.

> On a related note, is there ever a time where two or more <mn> or <mi> 
> elements are siblings without an intervening <mo> element?

I assume you mean as children of mrow? they are often in that
relationship as children of other things, such a mfrac for exaple which
tyically has two sibling mi with no intervening mo.

It's not an error in mathml to have to adjacent mi as children of an
mrow although normally one would suspect that there ought to be an 
invisible times or function application in between. However it has
defined rendering behaviour, and actully you find it quite commonly in
conversions from TeX where often the original markup has no indication
of the operator, and a convertor just copies the TeX structure rather
than add an <mo>invisible times</mo> without any evidence that that is
the intended meaning.


The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs. 

Received on Sunday, 30 March 2008 22:13:03 UTC