3 ms·
> As I've said before - IMAP has a spec for "keywords". That's just because IMAP has a spec for everything. That's the huge problem with IMAP: it can do too m
by pilif 13y ago
> As I've said before - IMAP has a spec for "keywords".
That's just because IMAP has a spec for everything.
That's the huge problem with IMAP: it can do too much and it's one of the more complicated "classic" protocols (from the area where RFCs were much more concise and bare-bones).
There are so many implementation bugs in various clients and servers that IMAP feels like being kept together by duct tape.
As such, people are very reluctant to add more protocol features to their implementations and even if they did, with some likelihood, the feature would only work in very specific client/server combinations.
In case of labels, the incentive for Google was to produce the implementation that works best with the existing clients (use pretend-folders for labels), and the incentive for client is to implement labels the google way (the only widely-deployed server that currently supports them)
Email is such a mess.
- stephenr 13y ago> the incentive for Google was to produce the implementation that works best with the existing clients I doubt that very much. If Google could turn off IMAP tomorrow and not lose a huge chunk of their clients, they would.
- coldtea 13y ago>That's the huge problem with IMAP: it can do too much and it's one of the more complicated "classic" protocols (from the area where RFCs were much more concise and bare-bones). There are so many implementation bugs in various clients and servers that IMAP feels like being kept together by duct tape. IMAP does much because we have many needs. And HTML does even more, but we still have several top notch browsers competing. In any case, if IMAP development is so difficult, have people join ONE IMAP library project that tries to be as complete as it can, instead of each mailer creating its own IMAP client libs. Like we have Webkit/Blink which is used across multiple browsers.