MASON MARK (.COM)

the best IMAP client in human history sure sucks

2010-01-17

Brent Simmons posted something interesting today: "Email init". The gist of the idea is that we[1] need a really good IMAP client; no such thing exists; so fuck it, let's make one.

The bummer is that The Community has to build it, which is always a slightly dubious prospect. But once in a while, that works out.

There's just no economic justification for the investment (of money/effort) required to re-engineer a complex application that is, after all, a core part of every modern OS. That's why we haven't seen, and almost certainly won't see, a commercial solution. (I personally would buy Brent's mythical $500 IMAP client, in a heartbeat. But that only makes about three of us.)

I've been on a technology pilgrimage to IMAP mecca since around 1997. (Actually, a bit before, but memories get hazy -- the best carbon-dating I can do is remember how excited I was about Ethan's pitch for a Newton OS IMAP client, and the Newton got Steved in '97.)

Nowadays (and for the past several years), if you aren't using IMAP for your email, then you're doing it wrong. But the pilgrimage continues, because while things are much better than before IMAP was widely implemented, none of the email clients have really nailed it yet.

Over the years, I think I've at least launched every single IMAP client developed for Mac, Windows, Linux, and Palm OS (lol I'm serious), and fiddled around a bit further with most of them. Off the top of my head: Mulberry, mutt, Eudhorfa, Outlook Express, Entourage, Outlook, pine, Outlook full turd, Opera, Mail.app, Thunderbird and friends, Windows Live or Whatever The Fuck It Is Called, Chatter, PowerMail, Outspring Mail, ... well, the list is too long to even fit on the top of my head.

But ugh, they all suck. In different ways of course. And everybody has slightly different needs. For me, absolute minimum requirements include:

  • reliable offline mode
  • Japanese compatibility and UTF-8 support
  • indexed search
  • rule-based triggering of custom processes
  • multiple identities (addresses/servers)
  • SSL/TLS support

Even this short-ass list, sadly, isn't satisfied by most existing clients.

Throw in some other requirements, like:

  • good UI
  • tree-display threaded view
  • high performance
  • custom IMAP flag support
  • so-called "smart mailboxes" (i.e., canned searches)
  • decent scriptability/automation support
  • attachment removal (leaving the message on the server)

...and you are left with ZERO EMAIL CLIENTS. Not one client in existence satisfies all of those. For me, the one which comes closest is (sigh) Mail.app. It does most of those things, but is has some very major problems. Serious bugs. A crappy threaded view. It can handle my 60,000-message personal/business archive, but not that plus the 500,000 or so messages of archived mailing list mail (had to move those to gmail, which also sucks). With a bunch of accounts, it sometimes takes over ten minutes to quit cleanly (on an 8-core '09 Mac Pro), but reacts badly to being force-quit. And on and on, blah blah blah.

And this application is, on balance and IMNSHO, the best IMAP client in human history. But I yearn for something better. As far as software goes, this is probably my deepest and most long-standing wish.

So, do I want Brent's Magical Pony IMAP.app? Fuck yeah! Fuck yeah.

But is this the rare project that The Community can actually pull off? Well... I mean, frankly I doubt it, but there's at least some hope. Brent is a well-known and well-respected dude, so out of the gate the project has more momentum behind it than normal. As I write this, the mailing list has 171 messages posted to it, after existing for only 12 hours.

I would love to work on this, too. I mean LOVE it. Writing code, I mean. But I recently did this scary analysis where the cumulative behind-schedule-ness of all my active projects is something like four years[2]. So will I really be able to? Uhm... I uh,... I dunno. I assume most of the folks getting all hyped up about this idea are in a similar position. But I hope so.

And there it is: the reason I have a bit of hope for this project is that most of the people who actually really need a superpower IMAP client are people who can contribute something to making it real.

So will we? I guess we'll probably know in about a year.

UPDATE FROM THE FUTURE:
The answer was, predictably, "No."

[1]: we the email power-user subset of Mac users, that is; mainly businesspeople who rely on email

[2]: Yeah, I have thirteen actives that are between two and fourteen months behind initial schedule. And that's only work projects--I am not including personal projects like my Hasbro Baby's First Open Source Release that I hoped to ship in 2004, or cancelled projects like the distributed native-Cocoa issue tracking app some of you wasted an hour listening to me rave about in 2008. (And, ahem, that's not to say I am a total slacker, either; I did finish a few projects in 2009, a couple of which may have even been on time.)

PREVIOUSLY:

2010-01-13
2010-01-09
2009-03-25
2008-11-23
2008-11-05
2008-10-25
2008-08-20
2008-05-12
2008-04-08
2008-03-21
2008-01-29
1979-11-20

SUBSEQUENTLY:

2010-04-18
2010-04-18
2010-05-15
2010-08-25
2010-09-15
2010-11-03
2010-12-25
2011-03-09
2011-06-14
2011-08-10
2011-10-12
2011-10-12
2011-11-11
2011-12-05
2011-12-31
2012-03-16
2012-11-14