IMHO an important topic for a new mail system is also the way the
messages would be spread and where the messages are kept. These
as big influence on other things like tracing and authorization.
The current model is a mixture of push and pull. The messages are
pushed to the last message store and from there it is pull, at
least wihh IMAP and POP3 (pop3 could be more pull and useful if clients
wouldn't stupidly fetch the mail but provide a list to the user
and the user could select, anyway not the topic here).
The advantage of the push model is that the emails is carried as close
as possible to the recipient with every step, even if all MTAs in the
chain aren't connected/reachable at the same time (very important in the
early days of the Internet).
The push model also has a social component. Mailing lists are more
successful than online (web based) discussion boards, as you don't have
to jump to 100 places to follow e.g. discussions like this, but you
get it delivered straight to your mailbox. Newsletters are successful
because users want a notification and not visit a website 5-10 times a
day/week to catch one new information. So this is a big advantage to
both the sender and the receiver.
A disadvantage is that the intermediates and the mailstore usually has
to carry more load than the sender. If you have a 100 MB email sent to
10 people in the same mailstore it "explodes" to 1 GB of disk space.
Some tend to design a mailsystem that puts the burden of storing the
real content of the message to the sender (pull system). This has
the big disadvantage that the recipient and the the mailstore of the
sender have to be online both at the same time. This is more likely
the case for a mailstore located close (in terms of network topology)
to the receiver. Thus if one cannot reach the remote mailstore one will
not receive the message until the next login/poll and the problems would
be the same, as the message doesnt come closer to the recipient.
A big advantage would be that instead of large messages you first only
toss around small messages with notifications about a message waiting for
you somewhere. However without some string authentification scheme built
in it won't reducde the number of full messages transferred, because
as sales people if you get a message from someone unknown with a subject
of "contact" you'd have to retrieve the message anyway. So this will not
directly be a big difference for the spam problem. However it would be
probably easier to distribute information about valid mailstores and
one would not retrieve messages from unknown mailstores.
The "social acceptance" of such a system is smaller, as you always need
to act to get the message. When I opend this Mailbox today I had 123
messages waiting for me from this list and I could start reading the
ordered threads immediately. With a pull system I'd probably had to wait
some minutes to get all the messages to my mailbox and maybe some
servers wouldn't have been reachable or slow.
So until someone has a good strategy I'd vote for a new mailsystem
using push instead of pull.
\Maex
--
SpaceNet AG | Joseph-Dollinger-Bogen 14 | Fon: +49 (89) 32356-0
Research & Development | D-80807 Muenchen | Fax: +49 (89) 32356-299
"The security, stability and reliability of a computer system is reciprocally
proportional to the amount of vacuity between the ears of the admin"