Hello David,

Thank you for the information, it is good that the entity name problem will be solved
in the future.

I still fail to see why if it is so simple to expand each page before serving it,
it is so difficult for a browser to do the reverse and just read the named
entities and have itself do the expansion.

It is very similar to saying that all web-pages should expand web-addresses to the IP
numbers of all the links before sending out the page.  This is doable, but not done.
The browser handles dealing with IP address numbers.

Firefox, MathPlayer, and to a limited extent Amaya all handle MathML used
entity names.

Even Section F.5 "Operator dictionary entries" of the MathML 2.0 standard
uses entity names instead of the expanded numerical values.

While I am sure that browsers that don't automatically expand
named entities are technically correct, perhaps they should do the
automatic expansion anyway.

A author who makes a MathML web-page that validates as correct and
displays well on several browsers, but fails to display well using a particular
browser could be understood  as to thinking that perhaps the particular
browser does not handle MathML well.

I WANT the entity names to show when the source code is seen, so that others
can understand what is being done. 

For example:
What does this number represent?
∑ 
Who knows, but
∑
clearly means something with summing things (summation)

Named entities are important and should be understood by the
browser.

I use this page for finding entity names
http://www.robinlionheart.com/stds/html4/entities-mathml

    Joe


-------- Original Message --------
Subject: Re: MathML browser test page
From: David Carlisle <davidc@nag.co.uk>
Date: Fri, July 17, 2009 1:23 am
To: joe.java@eyeasme.com
Cc: www-math@w3.org


> The problem is that it difficult to write any MATHML without
> using the named characters (expanded entity references).

yes but (for any xml not just mathml) it's probably worth expanding
the page before serving it, either statically or as part of the server
configuration. Basically any XML parser will do this, so if you use say
rxp and instead of serving foo.html you serve the output of
rxp foo.xml
then all entity references are expanded.


> It is not like the the Unicode symbols get renamed that often.

That depends how often you think is "often", every version of mathml
(1.0, 1.1, 2, 2 2nd edition, 3, has had to re-arrange these names
to fit with additions to unicode as names that were defined to resolve
to private use characters or dubious use of asian punctuation characters
could finally be defined to use mathematical characters as these were
added to Unicode. I think now, finally, with unicode 5.1 we are in a
position to have a set that is fixed going forward, and split off into
its own standard

http://www.w3.org/TR/xml-entity-names/

so that MathML and other XML languages (and HTML5) can all use the same
names.

Speaking of html5, if things go acording to the current plan, the mathml
entity names will be added to the existing html entity names and
incorporated into HTML5 systems so they can be used without reference to
a DTD (for files sent as text/html). Files sent as XHTML would be parsed
by an XML parser and so would still need a DTD reference (which would as
now need to be a local subset if you wanted to ensure that it was always
read)

David

________________________________________________________________________
The Numerical Algorithms Group Ltd is a company registered in England
and Wales with company number 1249803. The registered office is:
Wilkinson House, Jordan Hill Road, Oxford OX2 8DR, United Kingdom.

This e-mail has been scanned for all viruses by Star. The service is
powered by MessageLabs.
________________________________________________________________________