RE: [tag] Proposal: Providing Decentralized Extensiblity of Enumerative-Attribute Values

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

Good point.  Because I offered a couple of alternatives for consideration, I
have helped muddy up the case of QName and anyURI being incompatible (some
QNames look like [relative] URIs).  (There is a special CURI syntax to
differentiate QNames from URIs when both are permitted in an attribute
syntax.  We don't need to go there.)

Let me suggest the following cleaner approach:

1. For an extensible enumerative-attribute value, the syntax shall be that
of a Namespace QName.  This is a recognized XML Schema base datatype.

2. When the QName is an UnprefixedName, the value of the attribute must be
one of the restricted values predefined in the specification.  No other
value in UnprefixedName format is allowed.

3. When the QName is a PrefixedName, the value is determined in accordance
with the namespace that is bound to the prefix and which defines the
interpretation of the local name of the expanded name.

4. Note: It is irrelevant in this definition whether or not the QName,
interpreted as a CURI, expands to a valid URI.  No interpretation as a URI
is required here, whether or not there are such uses of the same prefix
binding elsewhere.

 - Dennis

I would suggest that we do have a namespace by which the fixed (Unprefixed)
NCNames can be expressed using a PrefixedName.  That is to facilitate future
up-/down-level adjustment.

I would also suggest that all of our important namespaces be chosen to be
usable as prefix bindings in CURIs too.  This allows use as RDF and RDFa
terms, among others.  Again, this is idependent from the use in (1-4) above,
which is not based on use of URIs.

PS: I don't align with your use of compatibility in the observation that
normalizedString "is compatible with both."  I would say that all three are
mutually incompatible if a union of any combination of them is entertained.

-----Original Message-----
From: stephengreenubl@gmail.com [mailto:stephengreenubl@gmail.com] On Behalf
Of Stephen Green
http://lists.oasis-open.org/archives/tag/200911/msg00040.html
Sent: Saturday, November 14, 2009 10:40
To: dennis.hamilton@acm.org
Cc: TAG TC List
Subject: Re: [tag] Proposal: Providing Decentralized Extensiblity of
Enumerative-Attribute Values

I think we have another problem with using either
QName or anyURI in that both seem, from what
I've discovered so far, to be mutually incompatible.
[ ... ]