RE: [tag] Re: Checking Enumerative QNames for the TAML-defined simple values

["Dennis E. Hamilton" <dennis.hamilton@acm.org>]

Additional PS:

The [xml-names] specification for XML Namespaces says nothing about the use
of QNames as the value of attributes (apart from the restriction of ID and
IDREF type to NCNames as a consequence for namespace validity).  

Using prefixes and QNames in the values of attributes is an extended use and
it has to be established by any individual application that introduces it.
Fortunately, there is enough call for this that the XML Schema supports it
with a base type, as you demonstrated when you ginned up the union case.

I believe the derivation of a type that has QName for prefixed cases and
NCName for unprefixed, enumerated-constant cases is completely within
bounds.  And it solves the problem you raised originally which is to provide
tight syntax checking on the permitted NCName values.

Although XML Schema does not show a connection between QName and NCName, you
will see in the XML Namespaces specification that QName is built up from
NCName via PrefixedName and UnprefixedName.  Lexically, a QName is either an
NCName or an NCName:NCName.  We want to use only the ":" flavor as a QName
because the bare NCName is not intended to have any namespace or be checked
for Namespace well-formedness.

 - Dennis
 

-----Original Message-----
From: Dennis E. Hamilton [mailto:dennis.hamilton@acm.org] 
http://lists.oasis-open.org/archives/tag/200911/msg00070.html
Sent: Sunday, November 15, 2009 12:36
To: 'stephen.green@documentengineeringservices.com'
Cc: 'TAG TC List'
Subject: RE: [tag] Re: Checking Enumerative QNames for the TAML-defined
simple values

Oh, I'm sorry.  It was never my intention that the unqualified NCName be
treated as a (semantic) QName.  

The string that has the simple format of an unprefixed NCName is a constant
member of an enumeration that is defined for the specific attribute.  It has
no namespace, not even a default one, and it should not be understood as
having one.

Only the PrefixedName form is intended to be understood as a QName.  If XML
Schema had a built-in PrefixedName type, I would have used it. 

If you don't put in the ":" requirement, the union is useless because NCName
and UnprefixedName have the same syntax and so the restriction on NCName is
meaningless.  That is why "deprecated" got through without qualification.

It really is intended that only prefixed names be treated as QNames by the
schema assessment.  Furthermore, pattern facets for QName restrictions are
explicitly allowed by the XML Schema specification.

 - Dennis

PS: I suggest you try it and then see how well your examples turn out.

-----Original Message-----
From: stephengreenubl@gmail.com [mailto:stephengreenubl@gmail.com] On Behalf
Of Stephen Green
http://lists.oasis-open.org/archives/tag/200911/msg00069.html
Sent: Sunday, November 15, 2009 11:31
To: dennis.hamilton@acm.org
Cc: TAG TC List
Subject: Re: [tag] Re: Checking Enumerative QNames for the TAML-defined
simple values

[ ... ]