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

Re: Problem with MathPlayer and CSS

From: Robert Miner <RobertM@dessci.com>
Date: Wed, 8 Jan 2003 10:17:52 -0600
Message-Id: <200301081617.KAA26411@wisdom.geomtech.com>
To: davidc@nag.co.uk, Bernhard.Keil@soft4science.com
CC: www-math@w3.org


Hi.

MathPlayer reads the CSS environment at the root node of a MathML, and
uses that information to initialize the fontsize, fontfamily,
fontstyle, fontweight, color, background and displaystyle (in the
language of MathML 1).

You can set the CSS environment at the root any way you want -- via
class or style attribute, through inheritance from a parent element,
etc.  MathPlayer merely consults the runtime CSS data structure, and
uses whatever it sees.

MathPlayer does not read the CSS environment at nodes below the <math>
node. 

There is a philosophical decision to be made as to whether one should
style MathML via CSS even in an environment that supports it.  There
is a practical argument and a theoretical one:

The practical argument is that Netscape 7/Mozilla and Amaya are
currently the only implementation where MathML can be styled via CSS.
By contrast, virtually all MathML implementations implement MathML's
own "style" machinery.

The theoretical argument is that it is dangerous to use CSS to
internally style MathML.  "Style" directives in equations, such as
setting a variable representing a Lie algebra in a fraktur font, are
not style information at all, but rather carry semantic meaning.
Thus, it is better to encode that directly in MathML.  If it is
accomplished via CSS, the meaning of an equation can be inadvertently
changed as a side-effect from replacing a stylesheet, etc.  

For these reasons, my own personal view is that equations should only
be styled via CSS at the root level, and internally, MathML's own
mechanism for controlling display properties should be used.  Of
course, I realize many will disagree.  

--Robert

------------------------------------------------------------------
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 Wednesday, 8 January 2003 11:18:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:52 GMT