ned+dkim(_at_)mauve(_dot_)mrochek(_dot_)com wrote:
Barry wrote (though not speaking as chair):
I'm sympathetic to the idea that we'd like not to spend a lot of time
on an aspect that's pretty much never going to show up... on a "corner
case".
That's the main idea I was trying to convey, but it appears I utterly
failed. Oh well.
Some day maybe I'll learn that geeks will always prefer to beat edge
cases into the ground, even when there are more important issues to deal
with.
Then you've completely missed the point of several people's comments, my own
included. The problem in general is that one person's unimportant edge case may
well be an essential feature for some large group we are unaware exists. Or it
may not. There's just no way to know, and unless someone here has access to an
oracular fountain of some sort, no way to find out.
In the specific case of multivalued From: fields, we're not talking about an
obscure syntactic fillip like, say, the ability to stuff comments in random
places throughout a structured field. Such capbilities provide no consistently
meaningful additional semantics and are therefore not something it makes sense
to spend a lot of effort to support. But this is different: We're talking about
the mechanism the protocol provides to associate multiple authors with a
message. Given that messages can be and sometime are written by more than one
person the potential utility of this semantic is quite clear.
I would really love it if we could get past the meta-discussion of "is
the multiple From: case important?" to the proposals that have been made
to address the issue. These include:
1. Perform SSP checks on the domains of all From addresses in the
message, with the exception of addresses having valid Author
Signatures. If any of the checks result in a Non-Compliant (formerly
Suspicious) result, then the message is considered Non-Compliant.
or
2. In the case of multiple From: addresses in the message, and the
domain part of one of the addresses matches the domain part of the
Sender address, then perform an SSP check on that address unless it has
a valid Author Signature. If the Sender header field does not match the
domain of one of the from address or is missing [violating 2822], revert
to alternative #1.
There are some other variations, but I think these are the two main
proposals.
-Jim
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html