On 11/11/2010 9:45 PM, David F. Skoll wrote:
On Thu, 11 Nov 2010 21:35:11 -0500
Jason Bertoch<ja...@i6ix.com> wrote:
After many complaints from the DNS community over SPF "hijacking" the
TXT record, a new SPF record type was eventually accepted.
The proper fix would have been to make SPF lookups for "example.com"
request the TXT record for "_spf.example.com" I guess it's way
too late for that now. :(
A whole new record type for a technology of dubious utility seems
silly to me.
Sure, the new record type was not necessary, but that's what the
anti-SPF community demanded at the time. However, I'm not as convinced
as I used to be that SPF is a technology of dubious utility. With mass
deployment of SMTP-Auth and modern webmails supporting external
POP3/IMAP connections, the issue of "SPF breaks forwarding" is nearly a
moot point. Moreover, I suspect that, as we convert to IPv6, current
methods of using negative IP reputation data will be unsustainable.
Barring an SMTP replacement, a requirement to publish known sending
servers may become a necessity, if for no other reason than to keep
blacklists relevant. Of course, the +all and ~all will need to be
dropped from the specification before SPF is of any use at this level.
--
/Jason
