RE: Last call on draft-gahrns-imap-referrals-01

["Mike Gahrns (Exchange)" <[e-mail]>]

Hi Harald,

I agree with your concern regarding avoiding areas that would need to
deal with synchronization between multiple servers, and will definitely
avoid that.

However, I am was thinking about a less ambitious approach.  As in the
SELECT of a mailbox case, the referral would be just a hint to the
client that it needs to make a connection to another server to access
that mailbox.  For APPEND I envisioned something like:

e.g.
C: A001 APPEND mailboxfoo data...
S: A001 NO [REFERRAL IMAP://Server2/mailboxfoo] Remote mailbox

The client would then need to do a connection to the server2 to append
the data.  This saves the client from needing to select the folder first
before doing the append.

COPY would be more difficult for the client.  Essentially a tagged NO
with a referral would be a hint to the client that it needs to make a
connection to another server, and APPEND the data itself, rather than
relying on the server to do it.

These suggestions have come about as our client team is getting some
implementation experience under their belts with this draft.  I would
like to try to incorporate their  ideas/experience if possible. 

What would your thoughts be, if  a subsequent draft, took the less
ambitious approach outlined above?  

P.S.  I will be travelling the rest of this week, and may be quite
unresponsive on e-mail.  Apologies for any delay in getting back to you.

thx,
-mikega
> ----------
> From:
> [e-mail][SMTP:[e-mail]]
> Sent: 	Tuesday, March 25, 1997 2:26 AM
> To: 	Mike Gahrns (Exchange)
> Cc: 	'IESG'; 'Keith Moore (IETF Area Director)'
> Subject: 	Re: Last call on draft-gahrns-imap-referrals-01 
> 
> Mike,
> it seems to me that adding COPY and APPEND to the redirects leads
> you into the area where you have to synchronize between multiple
> IMAP servers (espec. COPY).
> I'm not sure this is a good idea on the first try; would it be better
> to issue this Proposed Standard now, and add extensions later?
> 
>          Harald A
> 
>