Greetings. I'm not on the [EMAIL PROTECTED] mailing list, but
I'm the primary author of nameprep. Feel free to toss nameprep
questions directly to me. (The main reasons I'm not not on
[EMAIL PROTECTED] are that I'm not a code developer and also
I'm a BSD bigot...)
>Harald Alvestrand writes:
>
> > has anyone taken a look (or want to) on the internet-draft
> > draft-ietf-idn-nameprep-00.txt, to think about whether it is
> > easy/sensible/hard/nonsensical to implement it?
>
>It is quite sensible; the draft can be implemented in ca. two weeks of
>coding.
I know someone who did it in significantly less than that using the
IBM ICU freeware in Java.
> It can be done with good performance with rather small tables:
>2 KB for the prohibit table, 7 KB for the lowercasing table, and a
>similar amount for the composition tables needed for UTR15.
>
>But the draft leaves two questions open:
>
>- Is Han folding (canon-2.3) completely removed from the proposal,
> or is it only postponed until a later draft? I'm noting that a
> different draft says "There are no well-established rules for
> such folding, and some of the proposed folding would be
> locale-specific." - therefore Han folding appears not to be suitable
> for inclusion in such a canonicalization engine.
It is completely out. The main proponent of Han-folding has an
Internet draft that essentially says it is impossible to do
predictably. No one has disagreed with his draft.
>- What is the encoding in which internationalized domain names are
> exchanged between DNS servers? UTF-8 or UTF-5?
There are no standards now. The leading proposals (where "leading" in
the IETF can mean "the least despised") are UTF-8 marked with EDNS0,
and RACE encoding. Don't expect any clarity on this for many months.
--Paul Hoffman, Director
--Internet Mail Consortium
-
Linux-UTF8: i18n of Linux on all levels
Archive: http://mail.nl.linux.org/lists/
