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

Re: data-* attributes

From: David Carlisle <davidc@nag.co.uk>
Date: Mon, 3 Nov 2014 08:19:17 +0000
Message-ID: <54573A85.40807@nag.co.uk>
To: <www-math@w3.org>
On 31/10/2014 10:13, Peter Krautzberger wrote:
> Hi www-math,
>
> I've recently noticed that the W3C validator
> <http://validator.w3.org/check> considers  HTML5 documents with
> data-* attributes on MathML elements invalid.
>
> I'm wondering if this is a bug.

Not in that one (as it's using the xhtml1/mathml2 xml dtd but

the html5/validator.nu based validator

http://validator.w3.org/nu/

also says it's invalid which I think is unfortunate. It's not really a
bug in the validator though as (mostly) it just says what we tell it to
say (in the relax NG schema for mathml3 in that case).
>
> It also made me wonder in how far the MathML spec could adopt data-*
> in general. The spec suggests
> <http://www.w3.org/Math/draft-spec/mathml.html#chapter2_interf.unspecified>
> namespacing for unspecified data. When working in both XML and HTML5
> contexts this seems less than ideal.
>

My personal view is that we don't need to add data- in a "pure
xml/mathml" context) but that for mathml-in-html we should
push for as close integration of html/svg/mathml as possible.
At the DOM level a series of bug reports (and fixes:-) on html have
largely achieved that with many features such as .innerHTML etc being
pushed from HTMLElement to Element interface in the DOM so apply to
MathML and SVG.

At the markup level I think there is more we could do. One reason why we
never updated the old HTML+MathML+SVG dtd from MathMl2 to MathML3 is
that HTML moved away from grammar based syntax (data-* being a good
example of something that can't even be approximated in any of
dtd/RelaxNG or XSD schema).

But I think we should have a mathml-in-html note that explicitly
authorises data- and event based attributes such as onclick
onmouseover etc when used in an HTML context.

MathJax could easily ignore data- not sure how easy it would be to
make the mouse events work for rendering back ends other than native
mathml...

Which brings us back to Browser support, as Frédéric just reminded us in
another thread we perhaps need to push for better native support in more 
browsers before further standardising that support....

David
(personal response only, none of the above is a working group position)
Received on Monday, 3 November 2014 08:19:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:27:47 UTC