Re: [idn] nameprep forbidden characters

From: Mark Davis (
Date: Tue Sep 19 2000 - 12:56:37 EDT

It is the IETF that is responsible for domain names. The W3C i18n committee is
contributing to the discussion, but has made no mention of forbidding

The current discussion in the IETF is to allow Hebrew points in the names, but
fold them in comparison: e.g. comparison would ignore them. The same cannot be
done in general with accents, since they are quite significant in some

Mark wrote:

> Edwin F. Hart asked:
> > Is there a need for a "fuzzy" comparison where names with and without
> > points in Hebrew? Is there a similar need for other scripts such as
> > Arabic?
> Mark Davis replied
> > UCA (#10) already handles that. You will get a "fuzzy" compare if you
> > mask off less important weights, and you will get a much
> > better ordering
> > than binary compare as well.
> But then, why does the W3 Consortium want to *forbid* some Unicode
> characters from host names, rather than specifying rules for a loose
> comparison, maybe based on UTR#10?
> This way, people would have the possibility to give an orthographically
> correct name to their server (e.g., "www.Coƶperate!.ut"), but also be able
> to use common orthographic variants, if they cannot type all the "dots"
> (e.g. "www.cooperate.ut").
> _ Marco

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:13 EDT