Re: [idn] Using the last DNS header bit

["Donald E. Eastlake 3rd" <dee3@torque.pothole.com>]

From:  Dan Oscarsson <Dan.Oscarsson@trab.se>
Message-Id:  <200008021357.PAA17830@valinor.malmo.trab.se>
Date:  Wed, 2 Aug 2000 15:57:38 +0200 (MEST)
Reply-To:  Dan Oscarsson <Dan.Oscarsson@trab.se>
To:  idn@ops.ietf.org, ogud@tislabs.com

>...
>
>Using EDNS gives an unacceptable overhead as older servers will
>give an error response when using it. That is why the flag is needed
>in the normal header. One can question why DNSSEC were allowed to
>use up two of the three unused bits, DNSSEC could have used EDNS?

DNSSEC first RFC was 2065 from January 1997 (the RFC Index file
incorrectly fails to show, for the current base document, RFC 2535,
that it Obsoletes 2065).  EDNS0 is RFC 2671, August 1999, over 2 and a
half years later.

>Considering that I have no doubt that IDNs are by far much more
>important for most people on the planet, than DNSSEC.

I disagree that people would prefer internationalized information they
can not trust to approximately ASCII information they could trust.

>The alternatives to using the unused bit in the header are using one
>of the other bits only used in a response (not recommended by
>draft-ietf-dnsext-iana-dns-01.txt) or the response code field.
>
>   Dan

Donald