Re: Bind and Bvar

Dear David, dear all,

I agree with David's conclusion - grudgingly. But I would like to point
out that the OpenMath compatibility argument is somewhat problematic for
several reasons

  * strict content MathML is an implementation of _OpenMath Objects_ not
    of the XML endocing, so attributes do not even come in here.
  * OpenMath has a single <OMBVAR> element as opposed to multiple
    <bvar>s in strict content MatHML
  * having commonAtt on bvar gives you nothing that you could not do by
    putting them on the variables themselves. So putting the attributes
    on <bvar> gives another problem with canonical representation that
    we do not really want to have. 

But of course, I do not feel that we have much choice but arrive at
David's conclusion.


Michael




On 7.4.14 11:18, David Carlisle wrote:
> On 07/04/2014 06:09, Andrew Robbins wrote:
>> In the human-readable version of MathML3, bvar allows CommonAtt for
>> both Strict Content and Full Content, ... but in the DTD and RNC
>> versions, the bvar element does not allow any attributes.
>>
>> My question is which is wrong? the DTD/RNC? or the text? I would like
>> to be able to use id and class on bvar, but my documents will not
>> validate if I do.
>>
>> Regards, Andrew Robbins
>>
>
>
> Thank you for your message.
>
> The intended usage is as shown in
>
> http://www.w3.org/Math/draft-spec/chapter4-d.html#contm.bvar
>
> with the id on the ci child of bvar
>
> Which matches the text and the schema.
>
> However...
>
> You are correct to report though that the syntax table for bvar lists
> CommonAtt....
>
>
> Strict Content MathML is intended to be an encoding of OpenMath, and the
> OpenMath Schema allows id on the corresponding OMBVAR element.
>
> Also the MathML2 DTD allowed id on bvar.
>
>
> So I think the status is that the normative table is to be assumed
> correct and that CommonAtt are allowed.
>
> Almost all of the schema is mechanically constructed from the tables in
> the spec, but bvar, as the interaction with strict Content MathML is a
> bit tricky, was done by hand....
>
> This is a personal response but I'll fix the schema to add CommanAtt
> unless working group shouts.
>
>
>
> David
>
>

-- 
----------------------------------------------------------------------
 Prof. Dr. Michael Kohlhase,        Office: Research 1, Room 168
 Professor of Computer Science  Campus Ring 1, 
 Jacobs University Bremen           D-28759 Bremen, Germany
 tel/fax: +49 421 200-3140/-493140  skype: m.kohlhase   
 m.kohlhase@jacobs-university.de http://kwarc.info/kohlhase 
----------------------------------------------------------------------