Ambiguous description of annotation-xml/@encoding and @name

Dear all,

  I spotted some ambiguities and legacies in
http://www.w3.org/TR/MathML3/chapter5.html.

For MathML annotations it is said that the MIME types are
application/mathml-{content,presentation}+xml.  However, all examples continue
to use the legacy syntax of MathML 2, such as encoding="MathML-Content".

Then, about the annotation keys, the descriptions of
"alternate-representation" and "contentequiv" overlap.  For example, our
JOMDoc renderer (http://jomdoc.omdoc.org) outputs PMML annotated with, mostly,
OpenMath as parallel markup.  Now, what annotation key should we use?  Citing
myself from a mail to project-jomdoc@jacobs-university.de
(http://lists.jacobs-university.de/mailman/listinfo/project-jomdoc):

> @cd/@name: points to a symbol, usually from the built-in "mathmlkeys" CD,
>  that describes the semantic relation of the annotation to the formula it
>  annotates. The two predefined values for @name are
>  "alternate-representation" and "contentequiv".  Now I'm not so sure which
>  one to use; the MathML 3 spec is ambiguous here.
> 
> "alternate-representation" includes the meaning "to provide an equivalent
> representation in another markup language", "not alter[ing] the meaning of
>  the annotated expression".  This is certainly the case for our annotations
>  of PMML in OpenMath.
> 
> "contentequiv" annotations are used "to disambiguate the meaning of a
> presentation MathML expression", for "clarifying its precise meaning". 
>  This is also what we are doing.

Do you have any recommendations?

Cheers, and thanks,

Christoph

-- 
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701