Okay, so long as we're throwing ideas out, it always seemed to me
that we should have a multicast-friendly transport option for
push-mode messaging. We're carrying around a lot of wireless
gear, and a lot of my customers operate in such limited
throughput environments. Yet SMTP is ALWAYS unicast
connection-oriented. Okay, I admit this is the cheap end of the
multicast spectrum, but you should see my listserver chattering
away here opening a zillion connection when messages come in.
The waste of spectrum is a cryin' shame.
What if a list expander could associate a multicast address with
each mail list? That would certainly expedite matters. However,
some changes to the protocol would certainly be required. I like
e-mail in this role because it the only unattended push protocol.
If you've got a large file set to disseminate to a large group,
there would be no better and efficient way than a multicast mail
list scheduled to send in the wee hours. Thoughts?
I also strongly endorse the view from Jim Schaad that we need
security designed in from the start. The fragmentary state of
security implementation in existing e-mail is very disappointing.
I also like the notion of anonymity as a SECURITY service. It is
certainly part of what I mean by "privacy" in some contexts.
However, I am conscious of the fact that it has some potential
legal questions, and a serious yin/yang relationship with
anti-spam efforts. I have to say that I think anti-spam features
are also important, but not such a big deal as everybody makes
out. Despite the ever-increasing flood of spam, it is my
experience that most users don't even bother to use the tools we
give them today. Nevertheless, some better approaches both
maintain anonymity and help with to anti-spam efforts are
certainly possible.
Other concerns that I've seen mentioned and liked were in
Nathaniel Borenstein's list (though, not really my hot buttons):
-- Cleaner separation of header, envelope, and body
-- enhanced tracing mechanisms
-- internationalization, esp. of addresses
-- binary transport (though I think it's almost a
non-issue today)
Chris