3 ms·
The problem is that IMAP is a crappy protocol. How do you layer tags onto a protocol where that was never envisioned? Where a basic function "move between folde
by kalleboo 13y ago
The problem is that IMAP is a crappy protocol. How do you layer tags onto a protocol where that was never envisioned? Where a basic function "move between folders" is an extension and not in the original protocol? IMAP doesn't even guarantee that message IDs are consistent between mailbox listings! (there's a great rant about how terrible both IMAP and all the IMAP servers out there are, can't remember by whom)
I'm just glad Google still support IMAP at all instead of going fully proprietary protocol like they have with calendars, where they gave up on iCal.
- reeses 13y agoI was trying to remember the rant you refer to as well. I vaguely recall the core being that Mark Crispin's[1] RFCs were internally inconsistent, such that it was impossible to implement a standards-compliant IMAP server. Of course, UW-IMAP preceded the RFCs, so there was already a "reference implementation" against which other implementations could be compared and found wanting. Mark was especially obnoxious about protecting his view of IMAP. This was a weird time at the UofW. We had a hypertext system (UWIN) that was a mess and HTTP was banned as a server technology. Run a webserver for any reason, and disusering was a very real threat. They may have allowed gopher, but it's been a while. PINE, of course, came out of this, which wasn't too bad, as it inspired a number of decent successors. Best of all, Ada was the pedagogical programming language for the computer science program. Fortunately, you never used it past the second-semester CS course. [1] I just found out that he passed away late last year.