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

Re: A couple of questions on presentation mathml

From: David Carlisle <davidc@nag.co.uk>
Date: Mon, 27 Mar 2006 15:46:29 +0100
Message-Id: <200603271446.k2REkTbV010583@edinburgh.nag.co.uk>
To: www-math@w3.org
CC: Richard Kaye <R.W.Kaye@bham.ac.uk>

> In particular I was looking at the phrase "all of the children except 
> the first must be mtd elements".  This is I suppose correct 
> but I think misleading if in fact all children including 
> the first must be mtd's.   What is the correct status of this point?

As a personal response I believe that the dtd (and test suite examples)
are correct (all the children should be mtd) and that the text is wrong
to imply that the first child can be anything. I believe that you are
the first person to report this error, so thanks for that. (There is
certainly an error somewhere as the text and dtd are inconsistent).

> 2.  Why can't an mglyph be an image (given via a URL)
> rather than just a point in some font?

As you may know we're hoping to recharter to work on MathML3 and this
would be an interesting extension to look at. It interacts a bit with
the general (harder)  question about how mixed namespace documents
should work, for example in mozilla based browsers you can already
include an image inline in mathml just by using <xhtml:img, that is
because mozilla essentially implements a combined xhtml+mathml(+ now
svg) language they can be mixed in ways that are not exactly authorized
by the specs. It's not clear if the ideal solution is to allow such
mixing which has big advantages within a combined system, but makes
things difficult if such a "compound" fragment of mathml with embedded
html markup is moved to a non browser setting such as a computer algebra
system, and even within browsers, makes it difficult if the browser
implements compound documnts using plug in components (such as
mathplayer in IE). If you allow html-in-mathml-in-svg-in-html then the
components would have to do a lot of talking to each other to make
things work out. The alternative of adding an image element (or
equivalently extending mglyph) in the mathml namespace is simpler but
some would see that as duplication of functionality (if html img were
also to be allowed). 

> but I don't think I know how to use the mglyph feature *at all*, as it
> currently stands! 

True, mglyph hasn't turned out to be as useful as one might have
hoped. It's hard to predict these things in advance. There were hopes of 
browsers being able to access web fonts of one sort or another but as
far as I know that never really happened and mostly font usage relies on
fonts installed on the local system. Even for "general" mathematical
characters this is a problem and results in mozilla (for example) having to
pop up dialog boxes warning end users about possibly missing fonts.
For accessing non-standard fonts via mglyph the problem's even worse.


This e-mail has been scanned for all viruses by Star. The
service is powered by MessageLabs. For more information on a proactive
anti-virus service working around the clock, around the globe, visit:
Received on Monday, 27 March 2006 14:48:25 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:37 UTC