On August 16, 2005 at 18:13, Douglas Otis wrote:
When efforts to capture the RCPT TO prove expensive, this is only
expensive for the sender, not the recipient. Shifting burdens onto
the sender is an excellent trade-off. If each message created from
the To or Bcc headers captured the RCPT TO separately and are then
signed separately, this would not expose the Bcc list which would be
stripped. It would require that the signing process be located where
messages are disseminated.
Nice justification for requiring senders to do the "split" instead
of receivers.
I was thinking of general bandwidth aspects. It will increase network
bandwidth, for senders and receivers, but for a given message, a given
receiver bandwidth will be <= sender bandwidth. I do not have enough
information to determine if bandwidth considerations will be a problem
or not (much will depend on existing behavior and the frequency of
multiple envelope recipients to the recipient-side MTA).
--ewh
_______________________________________________
ietf-dkim mailing list
http://dkim.org