Re: mathcolor/mathbackground on <math/> and <semantics/> elements?

I completely concur with David.

The first version of Firefox's handling of MathML didn't handle semantics,
and this was a big problem for a lot of MathML that was generated.  Other
implementations also skipped it because <semantics> was listed in chapter 4
-- content MathML.  We have tried to make it much clearer that the semantics
elements should be considered part of both presentation and content MathML,
and that a renderer that deals with only one side of MathML should handle
semantics to at least the extent it makes sense for that side of MathML.

To be a bit perverse, if <semantics> accepted mathcolor/background, then
someone could complain that it shouldn't because it is part of content
MathML, and the section you quoted doesn't say it should be part of content
MathML.

    Neil


2010/4/21 David Carlisle <davidc@nag.co.uk>

> On 21/04/2010 11:27, Frédéric WANG wrote:
>
> Thanks for the comments, the following is a personal reply, we'll get
> group agreement before actually making any changes.
>
>
>  Hello,
>>
>> The spec says that mathcolor/mathbackground are accepted on all MathML
>> presentation elements:
>> http://www.w3.org/TR/MathML3/chapter2.html#interf.toplevel.atts
>>
>> <math/> and <semantics/> are not listed among MathML presentation
>> elements, so I think mathcolor/mathbackground should not be allowed on
>> them, and this is confirmed by the RelaxNG schema.
>> http://www.w3.org/TR/MathML3/chapter3.html#presm.summary
>>
>> However, the two sentences below could be interpreted as saying that
>> <math/> accepts mathcolor/mathbackground attributes:
>>
>> "The math element accepts any of the attributes that can be set on
>> Section 3.3.4 Style Change <mstyle>"
>> "mstyle elements accept the attributes listed in Section 3.1.10
>> Mathematics style attributes common to presentation elements. "
>>
>>
> I think this is fairly explicit that math should really work (as far as its
> children are concerned) like an mstyle. If the schema doesn't have that
> attribute that's an error, sorry about that.
>
>
>  and this one for <semantics/>:
>>
>> "semantics should be considered part of both presentation MathML and
>> content MathML"
>>
>
> That's more of a general hint that for example "just do presentation
> mathml" systems just  should at least accept semantics to the point of
> rendering a presentation attribution or a first child if that's
> presentation. (and not raise an error on semantics as an unknown element) I
> don't think that this informal note should have any effect on the schema.
>
> 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.
> ________________________________________________________________________
>
>