Re: [imap5] Designing a new replacement protocol for IMAP

[Bron Gondwana <[e-mail]>]

On Wed, Feb 15, 2012 at 07:25:52PM +0100, Jan Kundrát wrote:
> On 02/15/12 19:10, Bron Gondwana wrote:
> > Along with \Recent, which is also magic - it's a read only value where
> > EVERYTHING ELSE in the same space is writable (ACLs permitting), it's
> > defined in a way that makes multi-session a pain.  It's a wart.
> 
> On the other hand, some people (you know, I keep telling myself that I
> cannot possibly be alone in this) use the lack of \Seen to mean "I
> should do something for this mail", which means that we have a lot of
> "unread" messages in our mailboxes. Having a way of automatically
> marking those messages which are *really* fresh arrivals is therefore
> very convenient, and \Recent does exactly that.

Kinda, unless you lose your IMAP connection, or...

In other words, it's a lossy, partial solution - which comes at a
server end penalty of having to WRITE something when a client reads.

> Sure, we can use another flag for this, but having these flags managed
> by the server means that no client has to ever touch it, and therefore
> it will "just work" all the time. That's a pretty big benefit, IMHO.

Except then someone says "but I want to see messages which have not
yet been seen BY THIS DEVICE", so my phone sees things as new even
when I left my desktop logged in and checking for new messages.

Special cases are great when they match your use case, for sure.

> (Yes, I typically run just a single instance of a MUA these days and lug
> my laptop around -- that might explain why this works for me.)

Whereas I work in a couple, and have my phone... so it's pointless
to me.

Bron ( and the management types with their crackb^WiPhones... they want
       something different again )