-----Original Message-----
From: owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org
[mailto:owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org]On Behalf Of Mark
Lentczner
Sent: Friday, September 03, 2004 4:23 PM
To: IETF MARID WG
Subject: Re: TECH-OMMISSION: Time Phasing Rules For Records Not Included
In Core and Protocol (was RE: PRA Patent: License for Display in MUAs?)
On Sep 3, 2004, at 1:00 PM, Scott Kitterman wrote:
Recommend we add to protocol, page 24, a new paragraph 6.3:
6.3 Record Maintenance Time Requirements
...
I would recommend against such inclusions. I agree that the
information given is sound advice for system administrators. But, the
technical scheme is hardly dependent on following such advice, nor will
ignoring such advice lead to any major break down. Such language
belongs in a How-To document, not the draft.
I sort of agree, but the fundamental problem is that MUA level Sender-ID
validation is currently neither permitted not forbidden. I think the time
limits need to be written into the RFC somewhere if MUA validation isn't
prohibited. If their not, we will end up with disagreements about long
obsolete data should be maintained and whether or not someone's e-mail was
appropriately rejected.
I would expect that Jim's 28 days from the Caller ID spec was a surprise to
a fair number of people. It certainly was for me. The time frame for which
the record has to be valid is an important point. However we want to bound
the time domain, we need to do it. I don't propose we go back to the idea
of adding valid times to the records, that's complicated. I do think we
need to define limits however.
The best solution, I think, would be to have the MTA add headers that the
MUA can process at leisure and prohibit direct Sender-ID validation by MUAs.
Then the time bound is limited to the times associated with SMTP processing.
Those are generally well understood. I would have suggested that
originally, but was trying to avoid controversy.
Scott Kitterman