At 8:35 AM -0800 3/28/06, Jim Fenton wrote:
Paul Hoffman wrote:
At 6:57 AM -0800 3/28/06, Michael Thomas wrote:
So I'll ask: is there anybody out there who thinks that the current
set of mechanisms in -allman-01 is insufficient and/or wrong for dealing
with this problem?
Yes (insufficient). The current proposal for allowing signers to only
have to compute the hash once for large message bodies that will be
sent out numerous times (such as in a mailing list) seems like an
improvement.
IMO (In My Observations), most mailing lists don't need this feature
because they send the same message, body and headers, to all recipients
so the signature is the same even under the allman-01 mechanism. I get
a few messages with customized To: addresses, but most of them have
customized body content as well.
Paul, do any of the mailing lists you administer send the same body with
different headers to different subscribers?
No, but that's not the point.
The term "mailing list" can mean discussion lists like the one we are
on, and it can also mean one-way lists, like the ones that many
commercial organizations use to talk to their customers. The latter
type often used customized bodies, customized headers, or both. We
don't get as much of that now as we should because wanted commercial
email often looks like spam. If the DKIM protocol can support one-way
mailing lists better than -allman-01, it should.
--Paul Hoffman
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html