Eric Dean wrote:
so, I'll argue for system that will allow automated systems to
talk to each
other by e-mail without any human interaction.
In fact, we'll have to support both. Newer software can automate this
process while we'll have to rely on plain old SMTP for legacy use.
why differentiate? Or at least, why differentiate on the basis of target or
protocol? You can handle both automated and manual responses in the same
message without working hard or increasing complexity in the network.
By placing sufficient information for automatic handling in the headers, and
human oriented information in the body, you can satisfy the demands of both
automatic handling in human handling in one message.
example: in camram, I include an instance of a stamp with every message. In
this message, I will generate two stamps one to asrg and the second one to
eric(_at_)purespeed(_dot_)com(_dot_) They look something like to stand generated for an earlier
posting:
X-Camram: stamp; 0:030514:asrg(_at_)ietf(_dot_)org:c407927fbaad1fa8
In the body of the message is the human oriented material i.e. content for the
mailing list. The receiving filter sees the stamp, verifies that it is
"correct" (i.e. 20 bit string of zeros collision) and that it is addressed for
the recipient (one type of double spending) with the date that isn't too old
(second type of double spending). On receipt of a valid stamp, the filter knows
to pass the message on to my inbox. The same model can apply to other forms of
automated challenge/response.
part of the reason I take this approach is because I realized a while ago that
e-mail is a two-level protocol. You have the lower-level transport protocol
(SMTP) which is content independent. Then you have the higher level protocol as
represented by the message format which is instructions given by one e-mail
client to another. By modifying the higher level protocol by adding stamps or
other action triggers, you can improved e-mail system functionality without
destroying the existing infrastructure.
and yes I will be working on a working draft describing a stamp based protocol
as soon as I can clear the decks a little bit.
it would help if I could resist answering e-mail... ;-)
---eric
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg