On 27/Feb/12 21:30, Patrik Fältström wrote:
Yes, a pain. But...we just must be able to get new RRTypes
deployed. And not give up.
I agree it is a praiseworthy task to pursue the diffusion of unknown
RRTypes. However, it cannot be piggybacked on client protocols, that
have their own difficulties and version life cycles. When it will be
possible to deploy unknown types, protocols that nowadays reuse TXT
will convert to their own type, taking advantage of binary formats and
unambiguous querying.
There have to be transition periods, during which TXT records are
accepted for backward compatibility. However, when a new protocol's
version implies using a different record type, backward compatibility
should result from supporting the old version. To allow both TXT and
a specific RRType --especially one that uses multiple bytes to encode
a 4-bit version number-- to serve exactly the same protocol versions
is an /unfounded excuse/ that does nothing to encourage the diffusion
of that RRType. For SPF, it just doubled the number of queries.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf