RE: [office-comment] MathML is now "foreign" / illogical conformance (ODF 1.2)

Rob hi

> I thought MathML was only specified via a DTD. 

Correct, though v3 is defined using RELAX NG, and has good backward
compatibility with earlier versions. (However, it is probably too soon
to be thinking about MathML v3 for ODF).

> Is there an official
> RNG
> schema for them?

No, but it is of course possible to construct a RNG schema that is
technically equivalent to the DTD.

>  Or are you suggesting we just use NVDL?

Hadn't considered that ... interesting idea.

>  I think
> Michael
> looked into that, but nothing was proposed for ODF 1.2.
> 
> In any case, there is a difference between what the schema allows and 
> what the standard allows.

Sure (though as I understand it those differences are generally
irrelevant for the purposes of ODF conformance).

>  It is perfectly legitimate for the schema to specify something as 
> being an positive integer and then having the text of the standard 
> further restrict it to a prime number. They can't contradict each 
> other, but the specification can define additional constraints that 
> are not easily expressed by the schema definition language itself.

That is true. But generally ODF take the approach that where the grammar
*can* help enforce the correct markup, it does.  That is rather the
point of having a schema in the first place.

The comment in the schema says:

<!-- To avoid inclusion of the complete MathML schema, anything -->
<!-- is allowed within a math:math top-level element            -->

-- just sounds like laziness to me!

- Alex.