Wednesday, January 9, 2008

RIP Mitch Kapor's Chandler

Update: I have posted some of my additional thoughts and critiques ;-) in my most recent post: "Mitch Kapor's Weekend at Bernie's."

Yesterday I received an email that I have been anticipating would arrive at any time. The Open Source Application Foundation, creators of the collaborative PIM Chandler, announced that Mitch Kapor, its primary financial benefactor, is resigning and will be withdrawing funding at the end of 2008. The OSAF staff will immediately be reduced from 27 people to 10, and Katie Capps Parlante, currently the head of engineering, will be leading the project, with the hope of raising money to continue after Kapor's funding ends.

Mitch Kapor, founder of Lotus and creator of Lotus 1-2-3, founder of the Electronic Frontier Foundation, and former chairman of the Mozilla Foundation, started the Chandler project in 2001 with the hope of building an open source alternative to Outlook and Exchange. However six years later, Chandler has still has not reached 1.0 status. It does appear though that Kapor has been inactive, at least regarding product design and features for some time. In September 2007 they did release a 0.7 "preview" version of the product, though in my view, it is still not yet comfortably usable.

As someone who has spent many years developing PIM software, I have a passion for the category and have followed the Chandler project closely. I have been a member of the Chandler design mailing list for several years and have been rooting for them to succeed. But I recently came to the opinion that failure was inevitable.

From my perspective, Chandler was a rudderless ship. I tried to make suggestions which, though small, I felt could greatly reduce the complexity of the product. But their design process seemed to be insular and, honestly, broken. Even when I suggested they do some user testing, I was told by one engineer that he did not believe in the value of such testing.

The user interface design lead seemed intent on defending design decisions that it appeared few inside or outside the company agreed with, and that were in conflict with basic well understood user experience design theory. Worse yet, and quite amazingly, even as of yesterday on the mailing list, they are still debating who the target market for Chandler is.

Beyond poor design decisions and a seven year old and still incomplete development process, Chandler was hurt by a shift from desktop applications to web applications. Competitively, a slew of applications such as iCal, Google Calendar, and dozens of others also beat Chandler to the punch. And while Chandler does now have a web application companion called Cosmo, it is an entirely different code base from the desktop app, with a similar but different user interface all of which obviously ads substantially to the cost of maintenance and the complexity of the system.

The failure of Chandler is sad. But indeed after six years with no viable product or even strategy, it is finally time to die.

14 comments:

  1. Yes, it is sad that effectively Chandler is doomed, but that probably was the case anyway. Those who want to read some of the story can read the book "Dreaming in code". I tried it, but gave up about a year ago when I realised that I couldn't save data to a single file for backing up.

    ReplyDelete
  2. mitch, mitch, mitch. Mitch Kapor. Give this guy up. 6-7 years to develop an app.

    Come on, the guy is done, had some ok apps a couple of decades ago (which I was too young to use 1979) and hasn't got it in him again. This is what happens when you get old.

    ReplyDelete
  3. @first anonymous: You can back up your data to a single file. File > Export Collections and Settings...

    @second anonymous: One man (or woman) does not an application make in this day and age. Also, would you mind counting the years it took to get the Firefox 1.0 browser developed, just as an example? Let me help: Mosaic 1996, Netscape 4.5 1998, Netscape 6 2000, Firefox 1.0 2004... Depending on how you count, 6-10 years, in a well-understood application space (at least since 1998) that wasn't trying to do major innovations (except slim down) and was meant to do just one thing (as opposed to email, tasks, calendar, notes, ...) with hundreds, sorry, thousands of people working on it over the years!

    ReplyDelete
  4. @third anonymous: where did you come up with that crap?

    Mosaic was a completely different codebase, and a completely different product. Netscape 4.5 and 6.0? Again, completely different codebases. If you want to talk about products from roughly the same codebase, then maybe you can talk Netscape 6 and Firefox 1.0.

    Also, all those products you mentioned (and many other versions of Netscape) were *released*.

    And last I checked, Netscape 4.5 (and a few other revisions in there) did more than just web browsing. It did email, newsgroups, calendaring, contacts.

    Next time try a little research before spouting such stupidity.

    ReplyDelete
  5. For those of you interested in a more detailed analysis of the failure:
    http://gamearchitect.net/Articles/SoftwareIsHard.html

    ReplyDelete
  6. @ third anonymous, your Netscape argument holds little weight. All of those products were shipped, and were evolutionary. The time frame you need to compare is from Netscape's inception 'till their 1.0 release; everything after is irrelevant. With that said, Mosaic was started in December of 1992, and Netscape 1.0 was released in November of 1994. Two years, that's not bad. Now ~six years for a .7, that's not so great. And as the fourth anonymous said, Netscape was more than just a browser.

    ReplyDelete
  7. one meeleeonth anonymousJanuary 12, 2008 at 2:58 PM

    Correct me if I'm wrong, but IIRC Firefox *does* share some direct code ancestry from Netscape. Gecko was new, yes, but there's more to a browser than just the layout engine. I'm pretty sure our modern Spidermonkey is directly descent from the original JS engine in Netscape.

    ReplyDelete
  8. Mosaic was revolutionary, no other browser can claim the same.

    There have been many Chandler releases, 0.1 through 0.7.3. Remember that it is open source. If Mosaic/Netscape was started as an open source effort, there would have been plenty of 0.x releases. Instead, we see 1.0 and so on releases which are dictated by marketing, not readiness of the product (witness Netscape 6 also).

    And yes, quite a bit of code did carry over from Netscape to Mozilla and eventually Firefox. Gecko is new, and most of the other stuff has been at least partially rewritten since then. But Netscape/Mozilla/Firefox did not start completely from scratch.

    ReplyDelete
  9. Sorry I am late to this but this is standard Kapor post -123 which was his masterpiece.

    1. Designed Symphony for Lotus which did ok but a disaster compared to 123 which I was once told Mitch wanted to take off the market. (Lotus sold about $2 bb more of it).
    2. Jazz was all Mitch's, a disaster for the MAC.
    3. Agenda was cool to the three dozen people who could understand it. Talk about no UI.
    4. Mitch starts On Technology as an ambitious effort, wastes a couple of years and quits and disappears. On does a scaled down something or other.
    5. Mitch does Electronic Frontier Found. Good effort, good start, Quits when it gets hard. Usually after the PR buzz is gone.
    6. Does some investing on his own, presumably successfully. Then joins Accel. Quits a year or two later.
    7. Spreads some money at Mozilla, becomes Chairman. Starts Chandler seven years ago. Scales the ambition down and down and down until there is nothing left. And no code that matters. A freakin train-wreck. Good PR buzz for a while but nada.
    8. Now a new startup...Fox something that some guy from his past brought to him. A place to hang his hat and pontificate about usability and design some more.

    GIVE ME A BREAK. NOTHING HE DID ON HIS OWN SINCE 1983. CAN WE HAVE SOME QUIET NOW.

    ReplyDelete
  10. Hank, I was in the same boat as you. I was a very early participant, I had friends working there, and even met with Mitch at their weekly lunches. I poured way too much time into playing with Chandler and making suggestions (mainly because I was so excited by an Open Source alternative to Exchange).

    But alas, when I saw Google Calendar launch, combined with the sluggishness of Chandler's progress* I knew Chandler's future was dicey. *It reminds me of the computer language Ada which tried to be all things to all people and collapsed under its own weight.

    My best to the Chandler team though. And I imagine Mitch will focus more on Linden Labs, home of Second Life.

    ReplyDelete
  11. I lost a job because I thought Chandler would be available about 4 years ago. The world does need an alternative to Exchange.

    ReplyDelete
  12. Hank,

    Thanks for your eye-opening views on Chandler. I recently started using the Chandler Hub. I have tried Palm's software and a few other PIM programs, but have yet to strike gold.

    I'm looking for the best free PIM software; I need it to be web-based or to run on Linux. As an expert in PIM software, what do you recommend for me?

    Thanks!
    Looking for Gold
    jcarroll [at] byu (dot) net

    ReplyDelete
  13. Well, Chandler is released now v1.0.2.

    Time to revisit?

    tOM

    ReplyDelete
  14. calendar; contacts; notes; events; etc all in one place www.callij.com , replacement for chandler users.

    ReplyDelete