John Levine wrote:
I was referring to the swamp of prefixed TXT records.
The SPF part of this zoo clearly needs a major cleanup,
the wannabe-"experimental" phase ended in April 2008.
adding a prefix isn't the hard part, managing a
potentially unlimited sinkhole full of them is.
ACK, we agree that this zoo will never get a manager
wrt to TXT records. IOW folks creating such wildcard
records have to manage this for their specific goals.
If they missed "not recommended" + "discouraged" it is
their problem.
SPF, being only experimental, isn't an interesting
precedent.
The SPF folks submitted the draft for PS and IETF Last
Call, the IESG decided experimental without Last Call.
At that time there were about a dozen implementations,
and about a million of domains with an SPF policy. As
far as "running code" and deployment go that was clear
enough.
The technical point is that this is *no* precedent for
subtyping TXT, SPF got its own DNS record type. You
are nevertheless free to say that ADSP records have to
start with dkim=, and readers are then free to figure
out that this is compatible with RFC 1464 among others.
And unsurprisingly RFC 4406 managed to be incompatible
with RFC 1464 among others. But we know that RFC 4406
records don't start with dkim=, and we can ignore it.
A reserved RR type doesn't make it any less
experimental.
And I'd bet that there were at least ten more popular
IETF protocols than SPF in the history of the Internet.
Frank
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html