David Hopwood wrote:
> > 2) Define a stringprep profile for IDNs explicitly. > > > > One of the hard part of this approach is deciding on case- > > conversion. Life is easier if everything is lowercased but as > > shown, this breaks local-parts of email addresses. > > It breaks e-mail addresses only if a local part encoded as a label > in an RR is required to be normalised. Why would it need to be > normalised? I guess I mis-phrased the option. The value of a profile would be to normalize IDNs closely similar to nameprep (with different exception lists) so that Mark's particular concerns would be addressed. I think we are roughly in agreement on the point, judging from your response to my poorly worded response. The local part is a special label that gets special treatment. We should not apply global rules to an FQDN unless it is an IHN; for IDNs each label has to be treated specifically. > The same name equivalence relation should be used for all names in > queries, and names in RRs should preserve case (just as it is now for > ASCII). Can you expand on that? -- Eric A. Hall http://www.ehsco.com/ Internet Core Protocols http://www.oreilly.com/catalog/coreprot/
