On Mon, 22 Aug 2005, Aaron W. LaFramboise wrote:
I still think it would be very useful to have support for a state where results are returned very quickly (near constant time with respect to number of messages or size of mailbox) in the case that 'nothing has changed since last time.' Even with a notification mechanism, wholescale synchronization must happen every now and then, such as whenever the mail client is started.
This is already being addressed by other portions of the IMAP protocol: the \Marked and \UnMarked mailbox flags, the STATUS command, and the proposed CONDSTORE extension.
It's important to recognize that there is no disagreement over the desirability of such capabilities. It's a question of what is feasible to accomplish, given a number of technical and non-technical constraints (some of which are not always obvious).
For users like myself who have many mail folders that may not have changed since the last access, it would be nice if this initial synchronization could happen quickly, rather than the 3-4 minutes it takes me. (In my case, this time is solely spent reading MBX files; there is no appreciable network latency or bandwidth limitations.)
The mbx format, while a substantial improvement over traditional Unix format, has likely reached the limits of its scalability in today's world. You are correct that a new format is needed, one which has open time that is not dependent upon the size of the file (as traditional Unix format is) or the number of messages (as mbx format it).
I suppose there is also the classical biff/comsat mechanism. I guess my next step is to see what it takes to get Mozilla to accept new mail notifications from external sources (other than polling). It seems like it would be easy enough to have procmail, or whatever, generate a notification of some sort when it delivers mail, before passing off to dmail.
This is indeed the right direction. Your message has encouraged me to dust off my work from three years ago and call it to the attention of the IETF Lemonade Working Group which is working on this very issue (but is still in discussion stage on it).
Thank you for getting the ball rolling again, and please feel free to send reminders to nudge it along periodically.
-- Mark -- http://staff.washington.edu/mrc Science does not emerge from voting, party politics, or public debate. Si vis pacem, para bellum. _______________________________________________ Imap-uw mailing list Imap-uw@u.washington.edu https://mailman1.u.washington.edu/mailman/listinfo/imap-uw
