Luis Bruno wrote:
I've been catching up with the archives and I don't comprehend
this point. Why not support the use of TXT as a fallback while
the new record isn't widely supported?
The biggest problem is that if you publish a new record format, a
significant fraction of the people on the internet will not be able to
see it.
If I'm behind a Microsoft firewall (and I am, at the moment), I simply
cannot see a new record type that you've published. The firewall won't
pass a port 53 UDP packet, and the DnsQuery API can't cope with a new
record type. (I'm not defending the design; I'm merely reporting the
facts as they are.)
Presumably, if you're publishing a MARID record, you're doing so because
you want people to see and interpret your record. Doing so in a way
that makes it invisible to a significant fraction of the internet isn't
in your interest.
I don't believe that it's in *our* interest (our == MARID WG) to define
a standard that encourages people to publish records in a form that
can't be used by a significant fraction of the internet. A new RR type
does exactly that.
Note that the above argument is completely independent of whether a
Microsoft server can *serve* a new record type: it's been demonstrated
that it barely can (if it received the definition from a zone transfer,
and hasn't been rebooted since then).
-- jimbo