W3C home > Mailing lists > Public > www-math@w3.org > October 2002

Re: [www-math] new XSL stylesheets

From: David Carlisle <davidc@nag.co.uk>
Date: Fri, 4 Oct 2002 14:14:31 +0100
Message-Id: <200210041314.OAA30125@penguin.nag.co.uk>
To: msclrhd@hotmail.com
CC: www-math@w3.org


You've been busy!

I have had a quick look over the files although not run them yet, from
the looks of it I'm sure there will be improvements that I can port back to
the original stylesheet, thanks.

I note that you've used msxsl:node-set freely so it is microsoft specific
so it would need some modifications to work cross browser.

Since MathPlayer became freely available and Presentation MathML made it
into the default builds of netscape I have thought people would be less
interested in the stylesheet transforming MathML into HTML+CSS+javascript.
However you have clearly shown that you _are_ still interested in this
approach, I'd be interested if you could expand on when you think it is
still useful. (I had wondered about making my life simpler by dropping
the pmathmlcss stuff and assuming the presence of a mathml renderer....)

Some comments on some of your points.

> I have split the implementation into a presentation and 
> content folder that contain several xsl files, each of which dealing with an 
> individual element in the MathML 2.0 specification. The reason for this is 
> to simplify the modification of the way that an element is presented. 

I had hoped that the "single file" approach would also make it easy for
people to modify individual templates, all you need is a stylesheet with
one new template that imports the original stylesheet, the xsl import precedence
will ensure that the new template is used. Having the file in so many
parts (while good for maintenance) is quite a cost in extra network
traffic when you view the file. (For similar reasons the entity files in
the DTD are all incorporated into one file in the "flattened"
xhtml+mathml dtd that is available from the site.

> [2] ...There was a bug in the way that MSXML produces 
> generate-id() values for the tests I performed that resulted in duplicate 
> values being given for some generate-id() values when a document contained a 
> large number of MathML fragments.

Hmm I've not seen that, can you send me (off list) a test case, thanks.

> [5] I am transforming the Content markup into an XSLT variable, then 
> processing that element.

Yes what I wanted to do as well, but unfortunately this is mainly needed
in mozilla (as mathplayer and techexplorer have native content
rendering for IE) however mozilla does not support any extension
functions so it isn't possible to convert the variable back to a node
set, that's why the mathml.xsl stylesheet has the less elegant aproach
of writing out a processing instruction to invoke a second stylesheet.

> [6] I am transforming the mfenced element into the equivalent mo/mrow 
> construction.

As a matter of interest why do you need to do that, surely any
presentation mathml renderer will render the mfrenced anyway?

> [7] I have added rules to condense spacing of adjacent mo elements so they 
> display better.

This looks a bit suspicous, why are you getting adjacent mo anyway?
It may be due to

<xsl:template mode="c2p" match="mml:curl">

I think that should be mi rather than mo (my error, not yours:-)
the markup for a prefix function ought to be something like

> [8] I have added specific rendering changes to the content markup to get IE6 
> to correctly display the various mathematical operators (this is a personal 
> extension and need not be included.)

still I'd be interested if you could highlight these changes, the
structure of the files are so different it's hard to find them
just by looking over the code. It may well be that some of the
presentation choices in my original stylesheet are not good and should
be changed.

Thanks again for your interest and interesting re-design of the


This message has been checked for all known viruses by Star Internet
delivered through the MessageLabs Virus Scanning Service. For further
information visit http://www.star.net.uk/stats.asp or alternatively call
Star Internet for details on the Virus Scanning Service.
Received on Friday, 4 October 2002 09:14:38 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:32 UTC