--On July 6, 2005 11:05:33 AM -0700 Paul Hoffman <[EMAIL PROTECTED]> wrote: > > Spelling out the abbreviations as "Unicode Normalization Form C" and "Unicode > Normalization Form KC" is fine; referencing them is *not*. A reference to the > Unicode Standard inherently points to a particular version, and the tables > used > for NFC and NFKC change from version to version.
XML already has normative references to Unicode, so we can't exactly avoid those without dropping XML. Of course, the correct choice of normalization rules for atom:id is not the ones from the XML spec, but the IRI rules from RFC 3987. We really could have two sets of "standard" normalization rules in one document, one for XML and one for atom:id URIs, so I think it is worth pointing to RFC 3987 for indirect references to NFC/NFKC. Without clarification, this is a legitimate chance for confusion. wunder -- Walter Underwood Principal Architect, Verity
