----- Original Message -----
From: "David Maxwell" <david(_at_)crlf(_dot_)net>
So, for general protocol design, I'd agree that an 'SMTP replacement'
should probably require fewer round trips than current SMTP. Perhaps
something like:
True! The number of roundtrips is a matter of statistics and mathematics. It
would not really be creative to take the textbook as a divine commandment.
DCOM, CORBA, Web Services and COM (in Component Services etc) address the
resource consumption by avoiding roundtrips, but a mail server will have
many more break points, i.e. "Not OK" so the textbooks are not 100%
accurate. The important thing is to convince others that SMTP has too many
rountrips. A more slimmed approach is needed.
I have nothing against looking at replacements for SMTP - but I would
suggest that it should occur on a sub-list. It's a longer term project
than tweaking the existing protocol, and I'd prefer that it not bog down
discussions of changes that could be made more rapidly.
I agree. But I think we are too few to discuss this issue and the number of
possible implementations are too many. Possibly the best way is to implement
solutions and test them. My students will implement a Web Service based
solution with no connection to SMTP. I hope to encourage some of them to do
a final thesis on the subject push or pull, hopefully with some
implementations cooperating with SMTP mail servers. Since some of my
colleagues believe in SMTP I think no one need worry that the students will
get only the twisted ideas. ;-)
I would be happy passing any papers on pull to the students... Papers on
push are not cencured but they are abundant so the students will find them
anyway...
/DK
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg