Thanks for this. The working group has recently decided to prohibit
un-prefixed CURIEs, thereby solving the entire problem. A new draft
reflecting these changes will be available shortly.
Norman Walsh wrote:
A casual reading of the CURIE spec raised the following technical
questions in my mind:
"When a CURIE is used in an XML grammar, and the prefix on the CURIE
is omitted, then the prefix MUST be interpreted as the current default
XML namespace."
Current practice with respect to unprefixed names where a QName is
allowed are inconsistent on this point. In XML Schema, they are
sometimes taken to be in the current default XML namespace. In XSLT,
they are always in no-namespace. Was it the conscious intent of the
CURIE specification to remove this flexibility from specifications
that choose to adopt CURIEs? What is the rationale for this
restriction?
"When a CURIE is used in a non-XML grammar, the grammar MUST provide a
mechanism for defining the default prefix."
The default prefix? Do you not mean the default namespace?
Be seeing you,
norm
--
Shane P. McCarron Phone: +1 763 786-8160 x120
Managing Director Fax: +1 763 786-8180
ApTest Minnesota Inet: [EMAIL PROTECTED]
