What to keep in the registry database (Was: Tonus (was: Re: Casefolding Sigma (was: Re: IDNAbis Preprocessing Draft))

Vaggelis Segredakis segred at ics.forth.gr
Fri Feb 1 17:51:24 CET 2008


Thank you Stephane for this -on the point- clarification. We have our domain
names in the registry database in the *original* as you describe it form,
without the final sigma though (they are stored with the small sigma). I
feel we *have not* violated any RFCs, we were very cautious not to do that.

I am not worried about a database conversion if the final sigma was to be
stopped from being registered but rather the user experience of the IDNs
which would become poorer without normal language representation.

Vaggelis



-----Original Message-----
From: Stephane Bortzmeyer [mailto:bortzmeyer at nic.fr] 
Sent: Friday, February 01, 2008 11:41 AM
To: Patrik Fältström
Cc: Vaggelis Segredakis; idna-update at alvestrand.no
Subject: What to keep in the registry database (Was: Tonus (was: Re:
Casefolding Sigma (was: Re: IDNAbis Preprocessing Draft))

On Thu, Jan 31, 2008 at 12:26:49PM +0100,
 Patrik Fältström <patrik at frobbit.se> wrote 
 a message of 81 lines which said:

> because of that U+03C2 can not be stored in the core of a database
> for matchings.

A warning: the thread mentioned the terml "database" in two different
meanings, the DNS database (the zone file for those who use a RFC1035
text zone file) and the registry database. The second one is richer
and carries more information (and is not limited by the RFC).

For instance, while we do not have IDN in ".fr", we studied a possible
implementation and one of the decisions was to keep in the database
the *original* Unicode label, before nameprep, because nameprep is
lossy (it loses some information as the one mentioned above) and it
seemed to us safer to keep the original information, in case of, for
instance, disputes ("I registered straße, not strasse!")

So, when our greek colleague says that he registered names with a
final sigma, it does not mean he violated RFC 3490 and 3491.



More information about the Idna-update mailing list