I18N & XML Namespaces

This is a comment on

Namespaces in XML 1.1
http://www.w3.org/TR/2002/WD-xml-names11-20020403/


To comply with best practice on the use of non US-ASCII characters
in uri-refs, and for compatibility with the following XML specs:

XML Schema Part 2: Datatypes
section 3.2.17 anyURI
http://www.w3.org/TR/xmlschema-2/#anyURI

XML Linking Language
section 5.4 Locator Attribute (href)
http://www.w3.org/TR/xlink/#link-locators

XML 1.1
system identifiers
http://www.w3.org/XML/xml-V10-2e-errata#E26

It should be clarified that an XML namespace is identified by
a unicode string from the lexical space of the anyURI datatype
from XML schema datatypes.

viz:

[[[

The ·lexical space· of anyURI is finite-length character sequences
which, when the algorithm defined in Section 5.4 of [XML Linking
Language] is applied to them, result in strings which are legal URIs
according to [RFC 2396], as amended by [RFC 2732].

]]]

Since it is ugly for a core spec to depend on a non-core spec
it is probably worth clearly defining this type in the XML 1.1
integral specification, when discussing system identifiers, and
then referring to that definition from XML 1.1 Namespaces.

Moreover, given that XML 1.1 mandates an early uniform normalization
framework, alternative wording which specifically mentions Unicode
NFC could be used e.g. by copying the xink text and modifying it a
little.

(It would of course be helpful if charmod included this text, instead
of the non-normative reference to the IRI draft).

IMO, there should be no change to the current definition of
namespace equality viz:
[[[
(things) which identify namespaces are considered identical when
they are exactly the same character-for-character.
]]]

and the test case I posted to

http://lists.w3.org/Archives/Public/xml-names-editor/2002May/0001.html
and
http://lists.w3.org/Archives/Public/xml-names-editor/2002May/0002.html

should be considered legal.


It is necessary to gain the support of the I18N WG, particularly to
the notion of identity, and there should be a corresponding change
to the IRI draft to explicitly encourage such a notion of identity
when an IRI is being used as an identifier independent of any need
to dereference it (as in XML Namespaces and RDF).

I note that the RDF Core WG is minded to make a similar change to
RDF see:
http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Apr/0474.html
15: Issue: rdf-charmod-resources



Jeremy Carroll, HP

Received on Tuesday, 14 May 2002 08:38:43 UTC