John L wrote:
I've done another draft of the infamous DNSBL draft, to turn it back into
a draft for an IRTF informational RFC rather than standards track, and
to update a few other details that either people pointed out or I noticed.
Take a look, particularly at the expanded section 7 on security
considerations, let me know if there's anything that would make this not
the consensus of the ASRG.
Two things (I'm still reading it, may be more to come):
1) It has no reference to the DNSBL BCP. It should - it's important to
ensure readers have access to as full a picture of DNSBLs as possible.
While the lack (or not submitting the BCP and spec in parallel) probably
won't cause political problems with IRTF deliberations (like it did with
the IETF), it should be there anyway. We shouldn't have to stand on as
much formalism as standards track last-call, we should consider them in
parallel here, and make sure they cross reference each other.
2) The main sticking point with IETF standards track (once you got thru
the religious wars and talked specifically about technical objections)
is apparently the use of A records for DNSBL returns, instead of a new
RRTYPE. I at least partially disagree with the level of importance they
place on that issue, point out that it doesn't cause problems in
practise, and I also don't think a new RRTYPE would _ever_ fly - I'll
elaborate if someone wishes.
Still, some sort of parenthetical remark like this:
It should have been a new RRTYPE, should not be considered
precedent for something else overloading DNS like this, but it is
far too late to consider adopting a new RRTYPE for the huge installed
base of DNSBL usage.
Would go a long way to make the less religious IETF people be happy with
the document, regardless of how it's adopted. One even offered to help
word it... ;-)
_______________________________________________
Asrg mailing list
Asrg(_at_)irtf(_dot_)org
https://www.irtf.org/mailman/listinfo/asrg