> Can anyone explain to me the rationale for allowing some attributes
> to be based off of xs:token (and thus allowing various punctuation)
> while others are based off of data.enumerated which is based off of
> data.name ? For example, in <biblScope>'s @type attribute,
> punctuation is allowed, but is not allowed in say that of a <editor
I think you may really just want an answer to the question "why is
punctuation allowed in type= of <biblScope>".
The answer is pretty easy: it isn't. In several (if not all) of the
pre-release versions of P5 there was an error in the schemas such
that attributes of type data.enumerated for which the associated
value list was "semi" (i.e., an enumerated attribute of the
"suggested values include" variety) were inadvertently declared as
xsd:token instead of xsd:Name. This has been fixed.
So now (in the development version of the Guidelines in Sourceforge,
and thus for the version to be released), for all enumerated
attributes the only punctuation characters allowed are those that are
allowed in XML Names: dot, hyphen, underscore.