wayne wrote:
> I think it is time to split tracks and pursue both the RFC2821 and the
RFC2822 identities separately. I think we can move much quicker than
the schedule on the RFC2821, but the it will take longer. Right now,
we appear to both be doing a rush job and foot dragging, creating the
worst of both.
As long as the MARID record is reasonably extendable and appears to
allow for RFC2822 validation, I see no reason not to go full steam
ahead with RFC2821 validation.
Can we do a split track?
There is one interesting question - do we let the sender specify which
identity he is using or not? We can either let the sender specify the
identity, and then split the work into two sections (2821 and 2822), OR
like Andy's proposal called for, we can simply record the data, and then
split the work into two sections focusing on the algorithms that will
validate it. Chances are that the 2821 algorithm can be codified and
tested before San Diego, with the 2822 whenever we get to it.
It would also be useful if Microsoft can provide some data on Caller ID
and how they are doing header checking, including possible code and
problems.
Yakov
-------
Yakov Shafranovich / asrg <at> shaftek.org
SolidMatrix Technologies, Inc. / research <at> solidmatrix.com
"And this too shall come to pass"
-------