W3C home > Mailing lists > Public > spec-prod@w3.org > October to December 2006

Re: using MathML in a W3C recommendation

From: David Carlisle <davidc@nag.co.uk>
Date: Fri, 13 Oct 2006 17:42:26 +0100
Message-Id: <200610131642.k9DGgQhU027628@edinburgh.nag.co.uk>
To: connolly@w3.org
Cc: spec-prod@w3.org

> It's not clear that W3C process allows this; I see...

for the mathml spec itself we use html+images for the normative version,
and link to a separate xhtml+mathml version.

I think that's reasonable for mathml itself, you shouldn't need a
working mathml setup to read the mathml spec, but it would be nice if
8 years after the initial mathml spec, the w3c allowed other specs to
use normative xhtml+mathml.

That said, I don't think it's unreasonable to supply a html-only
version, especially if the mathematical layout requirements are not to

You can generate images, but an alterntive is to use to use something
like the pmathmlcss stylesheet available from 

which is advertised as a client-side solution for rendering but can
equally be done as a batch process. It takes xhtml+mathml document and
downgrades all the mathml to xhtml+css+javascript for the stretchy

the javascript part is a bit suspect really but works on older browsers
(ie browsers current at the time I wrote the thing) There are some newer
attempts to do mathml via css but they require good support for css2 and
3 features such as generated content and seleting on attribute values,
so it's not sure they are really usable as a general fallback for
browsers without mathml support.

Received on Friday, 13 October 2006 16:42:37 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:42:17 UTC