Jim Fenton wrote:
That version is quite a bit better. I'd suggest that the block "Verify
Signature" be "Verify Signature(s)" instead. The diagram is a bit of a
tradeoff, but I agree it needs to be.
Tempting, but that would have a cascading effect, because then, for example,
the
diagram would have to deal with different sources of signature and different
key
stores. Messy.
In the notes underneath, you talk about "validating", but I don't think
I have seen that term defined anywhere. From the context, it seems to
be authentication + assessment, but this should be more explicit. I
also prefer the word "validation" to "validating".
Oops. The word should be "verifying" since it only was meant to refer to
verifying a signature. THe gerund form of the word is intended to echo the
"signing" term used for the the other ADMD being mentioned. Parallel
construction, and all that.
Last paragraph, s/are not defined in this document/are at the discretion
of the validator./
1. The current language is formally correct. It declares an explicit boundary,
with this issue outside of it.
2. Anything that is not within a specification is always at the discretion of
whoever is doing the processing. Perhas counter-intuitively there is actually
more information in saying explicitly that the issue is outside the
specification than in saying the processor may do whatever it wants.
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html