It would be faster to implement however, I believe then the TB plugin because the only missing piece in this case is adding the IMAP client code to pull flagged messages into the hub.
There's additional work, such as finding a place to store IMAP auth information.
I'm strongly against incorporating IMAP pulling directly into Cosmo code.
If needed, it's a good candidate for an external daemon and service that proxies between an IMAP server and a Cosmo server. Use any implementation you'd like, including reusing python code from the desktop.
But in general, I don't think any web-based service that asks users to enter in their IMAP username/password can not be successful. People are often willing to share information, but the vast majority will not go so far as to put their email username/password out onto the net. People are only comfortable doing this with a desktop client.
I'm in favor of pretty much leaving things the way they are. The level of work being discussed here seems way out of proportion with what's effectively a renaming of "task" to "starred" and feels like a poor allocation of available resources.
-- Jared _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Open Source Applications Foundation "chandler-dev" mailing list http://lists.osafoundation.org/mailman/listinfo/chandler-dev
