W3C home > Mailing lists > Public > www-math@w3.org > February 2014

Re: Update to The MathML3 DTD

From: David Carlisle <davidc@nag.co.uk>
Date: Mon, 03 Feb 2014 22:26:35 +0000
Message-ID: <52F0179B.6050305@nag.co.uk>
To: William F Hammond <hammond@csc.albany.edu>
CC: "'www-math@w3.org'" <www-math@w3.org>
On 03/02/2014 19:51, William F Hammond wrote:
> David Carlisle writes:
>
>> The MathML3 DTD has been updated at
>>
>> http://www.w3.org/Math/DTD/
>
> Great!
>
> Looking forward to:
>
>      xhtml-mathml3-f.dtd
>         Combined XHTML + MathML DTD. Not Yet Available


Perhaps, although since the current incarnation of the HTML WG doesn't 
provide one I am not sure the Math WG should be responsible for 
maintaining an HTML schema, but see below.

>
> Two ideas:
>
> 1.  How hard would it be also to assemble xhtml-mathml3-f.rng ?

There is one in the validator.nu sources. It uses the MathML3 Relax NG 
plus a Relax Schema for HTML.  The schema used though doesn't everything 
that could be checked as it is simplified and more involved checks are 
not necessary in that context as the surrounding validator java code 
handles more things than could be expressed in the schema anyway.

>
>      When I run "trang" on the 20061127 version of
>      xhtml-strict-math11-f.dtd, I get the message 'error: no
>      namespace declaration for prefix "xsi"', but it makes
>      the .rng.  Then with "jing" on a trivial xhtml file, I
>      get:
>
>         xhtml-strict-math11-f.rng:10988:18: error: conflicting ID-types
>         for attribute "xref" of element "arccsc" from namespace
>         "http://www.w3.org/1999/xhtml"
>


which is related to the change of type for xref and ID that started this 
thread. trang always has problems with them as the interpretation of 
IDness in DTD and Relax is so different. Try just changing ID and xref 
to CDATA before running trang.
>
> 2.  Have y'all thought about making an sgml dtd for a
>      profiled version of html5 with mathml3 (for use
>      validating text/html serializations in shop without
>      hitting on http://html5.validator.nu/

You can download the validator.nu jar file and use it locally without 
hitting the web client (or its copy hosted at W3c)

>
>      I could pass along privately a similar quick hack for
>      mathml2 with not much for html beyond 4.01, but I suspect
>      David also has one of these.
>
>      In this connection I'd like to see someone with young
>      eyes pick up the ball and bring openSP up-to-speed with
>      the astral planes.  (Meanwhile one can write astral
>      things in html5 using cdata entity references, and they
>      will pass current [v. 1.5.2] onsgmls validation but will
>      not show up correctly in a parse.)  [ As a
>      mathematician, I imagine that this would be a suitable
>      summer-length honors project for a student in computer
>      science.  :-) ]


I haven't used opensp for decades but I thought there was a patch back 
then to work with the higher planes. perhaps my memory is failing.
>
>                                      -- Bill
>
>




David
Received on Monday, 3 February 2014 22:27:06 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:46 UTC