The term "Push email" comes from a client signing on to the server and
issuing a "look for ..." instruction to the server. Also known as
idling or long-delay poll.
The logic of it is to have the client only issue new "look for ..."
instructions when those instructions change, and until the client
disconnects, the server should send "i have new ..." responses whenever
it figures out there is new mail. Following that, the client says
"gimme" and all are happy.
The only problem with this is NAT traversal where a busy firewall ages
the oldest idle connections. To combat this, developers make the client
issue "look for ..." rather frequently if such behavior is discovered.
Some developers however just play the safe route and always issue the
"look for ..." instructions periodically.
In effect, the logic of this isn't really changed from the original
design. It's just done a bit differently. Some of it is really just
the marketing aspect. Aunt Millie doesn't grok the protocol. She just
sees the "new feature!" printed boldly on a high priced M$ product box :)
The only truth in advertising, is that there is rarely truth in advertising.
-david
David Schlesinger wrote:
>Microsoft "push email" isn't push at all. If you read the
>specifications, it's just another method of polling a server to
>determine if and what segments of new content is ready for transfer.
I think this is true for the "Outlook Web Access" interface which, for
instance, Evolution (and Pocket Outlook on Windows Mobile 5!) use.
There's some sort of back-end interface which Outlook 2003 and
Entourage can take advantage of with an Outlook Server; I'm not sure
whether whatever they do there qualifies as "true push email" or not...
_______________________________________________
OpenMoko community mailing list
community@lists.openmoko.org
https://lists.openmoko.org/mailman/listinfo/community