W3C home > Mailing lists > Public > www-math@w3.org > April 2003

Re: More mathvariants

From: Robert Miner <RobertM@dessci.com>
Date: Tue, 29 Apr 2003 09:58:19 -0500
Message-Id: <200304291458.h3TEwJi07016@wisdom.geomtech.com>
To: S.Pepping@elsevier.nl
CC: www-math@w3.org

Hi Simon,

> The font specification attributes in MathML1 allowed one to specify many
> more fonts than the mathvariant attribute in MathML2. For example, it is now
> not possible to specify italic or bold variants of double struck characters.
> (And if it were possible there would not be corresponding Unicode points.)
> Is this based on the STIX research? Typesetters do have such variations
> available, and I think they use them from time to time. Is there a way to
> specify such variants? mglyph would not be an ideal alternative, because it
> hard codes a specific font.

It was partly on account of STIX, but more directly because of
requests from the W3C style activity.

The design of the mathvariant attribute basically corresponds to the
math alphanumeric characters that STIX requested from Unicode in plane
1.  The idea is that these families of symbols frequently carry
semantic meaning in equations, e.g. fraktur fonts for Lie algebras,
etc. so the specification of these broad categories of fonts was
properly part of the markup.  By contrast, just exactly which fraktur
font a typesetter used was part of the styling, and should thus be
kept separately from the markup.

The W3C style people were very much oriented to the Web browser
environment, of course, and therefore tended to assume the existence
of something like CSS where one could use mathvarient="fraktur" as a
selector for a rule that picked a particular font.  Of course, MathML
is coming to be used much more broadly than that, particularly in
publishing environments such as yours, where there is no CSS.  That is
why the Math WG dug in its heels and refuse to remove the MathML 1
style attributes, and merely deprecated them.  But we more or less
lost the larger battle, and compromised on the mathvariant mechanism,
can be used to specify a font, but only indirectly through CSS, etc.

This discussion appears with a little more detail in section 3.2.2 of
the spec, but you proably already know that.


Dr. Robert Miner                                RobertM@dessci.com
MathML 2.0 Specification Co-editor                    651-223-2883
Design Science, Inc.   "How Science Communicates"   www.dessci.com
Received on Tuesday, 29 April 2003 10:58:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:54 GMT