W3C home > Mailing lists > Public > www-math@w3.org > May 2004

Re: The MathML documentation is broken

From: Robert Miner <RobertM@dessci.com>
Date: Fri, 21 May 2004 06:30:19 -0500
Message-Id: <200405211130.i4LBUJq22583@wisdom.geomtech.com>
To: Thomas.Fischbacher@Physik.Uni-Muenchen.DE
Cc: www-math@w3.org


Thomas,

> Please notice that the displayed text claims that the XML source uses
> 
>          <mo accent='false'>&OverBar;</mo>
> 
> while indeed, it _does_ use
> 
>           <mo accent="false">&#x000AF;</mo>
> 

Note that &OverBar; is defined to be &#x000AF;.  XML requires that
conforming implementations treat these the same thing, so your
observation is essentially a bug report about Mozilla. Of course, to
some degree that is beside the point.

> Now, wouldn't it be much simpler, better, nicer, easier, and more 
> reliable to auto-generate the "Source is ..." part from the XML math code? 

In fact, no.  The handling of entity names is rather problematic in
theory (for example, there is no good way to specify them using
schemas) and in practice, the situation is even worse, since
implementations vary widely in how well they actually deal with entity
names, as your experience with Mozilla underscores.

Secondly, you assumption that the source was not autogenerated is not
correct.  The entire test suite is automatically generated off of a
database of XML records.

The point is that we actually made a conscious choice to expand the
entity names to numeric references in the source, while leaving them
as names in the quoted code displayed to the reader.  The idea was to
avoid interoperability problems with entities between renderers, while
presenting the reader something more easily understandable.  The
rationale was that the issues with the entity names are in some sense
generic XML problems, and therefore independent of the MathML
implementation per se.

Moreover, there is a section of the test suite devoted specifically to
entity names, numerical references, and raw UTF-8 characters.  If an
implmentation had a problem with characters, then the thinking was it
was enough to catch them there, without hobbling the rest of the tests
in the suite.

Of course, one can argue the other side as well, as you are doing.
But my suspicion is that no matter how we do it, there will be certain
groups of users that won't like it.

--Robert

------------------------------------------------------------------
Dr. Robert Miner                                RobertM@dessci.com
W3C Math Interest Group Co-Chair                      651-223-2883
Design Science, Inc.   "How Science Communicates"   www.dessci.com
------------------------------------------------------------------
Received on Friday, 21 May 2004 07:30:53 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 20 February 2010 06:12:56 GMT