Based on constructive working group discourse, the co-chairs of MARID
observe the following:
1) There is consensus within the MARID working group for the use of SPF
syntax over other encoding schemes. While this consensus is not
unanimous or overwhelming, it is rough consensus. The working group's
rough consensus on this issue derives from several considerations;
among them is the belief that MARID's output should focus on the
short-term needs of MTA authentication (consistent with the charter of
the group), and that open-ended extensibility for email policy within a
MARID record may lead to problems of interoperability. The co-chairs
also note that there were no practical or obvious examples of
extensions to a MARID record that could not be represented by the SPF
syntax.
2) The consensus for the use of SPF syntax does not preclude continuing
work on the PRD and SUBMITTER concepts. The co-chairs note that there
has been very constructive discourse on these concepts and that the
working group should continue to refine these ideas for use with
MARID's output.
3) Despite the working group's consensus for the use of SPF syntax, the
co-chairs find that there is no consensus for declaring the SPF
specification finished. The co-chairs observe that there are still
many unanswered questions regarding extensibility in SPF, specifically
how and if SPF modifiers should refer to other policy frameworks and
the need to more clearly define default behavior and safe-guard against
side-effects caused by unknown SPF modifiers and mechanisms. The
co-chairs note that such semantics of extensibility are not specific to
any type of syntax or encoding.
4) Finally, the co-chairs observe that the MARID working group has a
very strong consensus, though not unanimous, on the reuse of TXT
records for initial deployment and on a reuse of TXT syntax in the long
term. The consensus is centered around the belief of most participants
that a MARID record in TXT form will generally be small enough as to
not require DNS over TCP. The co-chairs find that the working group
has not come to consensus on the use of a record prefix vs TXT records
at the zone apex and has not reached consensus on how to address
domains with wildcard MX records.
It should be noted that none of the above findings preclude the future
re-chartering of MARID to define a new DNS record type and/or a new
encoding for that record.
To meet the tight deadline this working group as set for itself to have
a candidate for a proposed standard by the end of August, 2004, we
propose the following schedule of activities:
- Due 2004-07-02: Decide if CSV is complimentary, parts to be
incorporated, or dropped.
- Due 2004-07-08: Decide how MARID output will work with already
deployed SPF records (v=spf2?).
- Due 2004-07-31: Refine SPF syntax and extensibility semantics.
- Due 2004-07-31: Fold in PRD and SUBMITTER.
-andy & mtr