Re: Minor problems found in MathML 3 specification

On 19/03/2012 21:35, Andrew Miller wrote:
> Hi all,
>
> I've found a few (primarily editorial) issues with Chapter 4 of the
> MathML 3 specification;

Thanks for these.


> I can't find the link for a bug tracker for specification issues, so
> I presume this list is the correct place to send things like that?
yes this is the right place.
>
> 1. Rewrite: partialdiffdegree This rule refers to the expression as
> 'expression-in-x1-xk' in the match rule, but as 'A' in what it is
> supposed to be transformed into.

Yes this seems to be the case. I haven't confirmed with the rest of the
Working group, but it looks like A is in error I agree.
>
> 2. Section 4.2.2.1 Strict uses of|<ci> This section neglects to
> specify how ci elements with type 'complex' are transformed into
> strict form, given that there is no complex in the mathmltypes CD.

Again, not checked with the WG, but it's a slightly odd choice to have
polar and cartesian complex representations considered as different
types at the OM level, it makes sense for literals (ie cn) but not
really for identifiers, other to record a preferred form for expressing
any value to which the identifier might correspond (which was, as I
recall, why we decided to keep them as distinct types in the mathmltypes
cd, to match the annotations allowed on cn in mathmml2). I think that ci
type="complex" should be considered an alias for
type="complex-cartesian". It would be least disruptive (ie no changes to
the strict schema) if we continued to view both forms as "strict", but
of course an alternative would be to view ci type="complex" as not
strict and explicitly rewrite it to use complex-cartesian.
>
> Best wishes, Andrew

thanks for these comments. Once there is an update I'll reply again, and
update
http://www.w3.org/Math/Documents/mathml3-errata.html

David

Received on Monday, 19 March 2012 22:16:53 UTC