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

Sorry for letting this drop;
I've made changes to sections 2.2.1
and 3.3.4.1 to reflect this interpretation;
namely that mathbackground sets the background,
not the default, for math and mstyle.

You can see the changes at
http://monet.nag.co.uk/~dpc/draft-spec/Overview-d.html

Thanks for your comments.
bruce


On 04/21/2010 11:31 PM, Karl Tomlinson wrote:
> Taking Bruce's comments out of order, hopefully not too out of context.
>
> Bruce Miller writes:
>
>> In other words, I think it would be more
>> natural and useful for math@mathbackground
>> to set the background for the contents, not the
>> default.
>
> I agree.
>
> I guess what is not clear is how this should be specified.
>
>> I believe the issue stems from there being two ways
>> to think of mathbackground (&  mathcolor) being
>> allowed on mstyle.
>>
>> (1) is that mstyle is a presentation element, and
>> so it accepts the mathbackground attribute. The
>> effect is described in 3.1.10, namely it fills the
>> bounding box of the element (and it's children).
>>
>> But
>> (2) is that the mathbackground attribute is accepted
>> on mstyle since the mathbackground attribute is
>> accepted on _other_ presentation elements (see 3.3.4.1).
>> In such cases, the effect is to change the default
>> mathbackground.
>>
>> Those effects are subtly different, as you noticed.
>> .. and in fact, I think (2), is kinda screwy
>> [mathcolor seems to bypass the screwiness
>> by having it's default be inherited, anyway].
>>
>> I would argue that (1) ought to take precedence
>> since it's just so much more direct.
>
> Perhaps only (1) should apply, but I think for that to be
> the case it needs to be added to mstyle's exception list.
>
> (1) is actually new to MathML3.
> MathML2 already had (2).
> Both (1) and (2) can actually coexist, but we don't like that
> behavior.
>
> Now that we have (1), we like that behavior better and so want
> to deprecate (2).
>

Received on Thursday, 27 May 2010 13:56:58 UTC