Keith Moore wrote:
I certainly agree that DKIM appears to have lower barriers to
deployment than some of its predecessors (e.g. S/MIME), and I also
think that there's more of a perceived need for something like DKIM
than there was for its predecessors...if S/MIME were being promoted as
a new thing today, it might be more successful.
I'm not sure that DKIM lowers the barriers enough to enable the
"network effect", but I think it's a step in the right direction if it
can lower the deployment barriers AND be made to provide the right
functionality. (I don't think it does the latter yet)
Ok, tell us what the right functionality is.
but I think goal #4 is unrealistic or misstated. DKIM should be
relatively non-hostile to legacy MUAs and MTAs (as compared to
multipart/security based solutions) but MUAs and also some
MTAs will need to be upgraded to significantly benefit from
DKIM.
Yes, of course some MTA's will need to be upgraded, what
I meant is that you can pick and choose them selectively
which makes the chore easier. As for MUA's, I agree that
the maximal benefit will arrive when they can take advantage
of DKIM, but I don't think it's strictly necessary -- I can
today devise a filter in almost all MUA's which blats something
onto the screen so I can tell whether the message verified or
not. Maybe you consider this "modification", but I meant it in
the sense of having to wait for a design/release/uptake cycle,
as well as the sense of "no flag day required".
Mike
_______________________________________________
ietf-dkim mailing list
http://dkim.org