On 1/6/07, Serge Knystautas <[EMAIL PROTECTED]> wrote:
On 1/6/07, robert burrell donkin <[EMAIL PROTECTED]> wrote:
> sorting isn't good enough for me either: emails need to be tagged with
> meta-data on the server. meta-data can then used to present a folder
> based view to a client or for any other purposes. i think that a rules
> engine plus a domain specific language would work better than SLIEVE.
Robert,
I'm right there with you. I don't know if GMail has it perfect, but I
find tagging and better searching much more powerful than the
foldering system I have from 10 years of emails in IMAP.
+1
gmail's limited: i can't write scripts or automate and the number of
tags is far too few.
for example, agents should be able to automatically tag every mailing
list based on the start headers so that i don't have to.
ATM i have over one hundred client side rules for filtering. given a
decent meta-data solution, i would expect to very quickly move to many
hundreds of rules.
Do you see any way to backfit tagging concepts into IMAP, or do we
need a separate protocol (perhaps http/restful as you say)?
one way that this could be done by adding new headers to the email.
this approach would be storage and protocol agnostic. could probably
use same code for client- and server-side filtering. may not be as
elegant as a pure server-side solution but i haven't analysed the
mechanism in detail. opinions?
i'm not sure that IMAP could easily exploit this fully. IMAP is a
*tough* protocol: big, complex and difficult to implement securely
(kudos to the james IMAP crew). the meta-data capacity is limited,
non-extensible and hardwired into the protocol. adequate performance
would be difficult to achieve. but yes, i think it would be possible
to add new features to IMAP to support meta-data. standardization may
be difficult, time consuming and potentially politically difficult.
not something i'd be motivated to take on...
IMAP has a lot of disadvantages as a protocol (as least for what i
have in mind). given that fitting full meta-data support would mean
changes to the protocol, i think that this is an opportunity to push a
new protocol based on REST which is designed for the internet: an
email protocol for the new millenium (and no, i didn't coin that
phrase ;-) there's interest in the DAV community and the WG are
willing to help implementors discover the right ways to implement.
i don't believe that a fully functional internet scale meta-data aware
server can be created by working within the current IMAP standard but
the protocol connecting the client and server is only one component.
a meta-data aware server would be able to deliver a IMAP
implementation with advanced features such as server-side virtual
folders but these would probably require management from a console on
the server rather than integration into the client.
- robert
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
