Re: [idn] iDNS re-chartering proposal, take 2

[Dan Oscarsson <Dan.Oscarsson@trab.se>]

>>For example, UDNS requires a coordinated release for at least two
>>purposes: so that name lengths can be synchronized, and so that
delegation
>>entities can
>>coordinate delegations they make as UTF-8 and ACE encoded
representations
>>of the same UCS string.

It does not. Name lengths is specified and will be supported by all
supporting UDNS.
Though some name lengths will not be possible to have in ACE, but ACE is
just
for backward compatibility. Never expect everything new to work in all
old
systems.

And there is no need for coordinated delegations. UDNS only uses one
representaion
of a label. The ACE format is always created from the UTF-8 native form.

There is no double delegation.

>
>uDNS a) is not a complete specification, and b) imposes operational
>requirements that are not viable in the existing DNS infrastructure.
>
Easy to say, but I have not seen anything to prove it.

UDNS is  a complete specification, but it does leave specification of
ACE and
how DNSSEC sorting is to be done to other RFCs. IDNA is a very
incomplete
specification - it does not solve the problem of having non-ASCII
in DNS, it just defines how to encode non-ASCII in "host names" into
an ACE.

And I do not know what operational requirements you think of.
I can start using UDNS today in my DNS server and it will work
fine.

    Dan