Earlier this week the Open Source Applications Foundation — the organization developing Chandler, whose work I followed for three years and whose story I tell in Dreaming in Code — announced what it called a “restructuring,” which meant laying off roughly 2/3 of its employees. (Infoworld’s story has the basics.)
So does this mean that Mitch Kapor, OSAF’s founder and chief funder, is “pulling the plug” on OSAF, as Techdirt has it, or “bailing on” Chandler, as a CNET blogger put it?
Kapor is indeed leaving the OSAF board and handing the rest of the reins to Katie Capps Parlante, who’s been running the project day-to-day for some time. That’s a big change. From what I can tell, though, this isn’t so much a “plug-pulling” as an amicable parting of the ways.
Kapor’s financial withdrawal is not a big surprise. He had told the Chandler developers several months ago that he intended to fund the project through the end of 2008 but no further; that decision had been discussed on the project’s public mailing list, so the looming changes couldn’t have come as a total shock to anyone who’d been following the story. (I write about these changes in a new epilogue to DREAMING IN CODE that will be included in the forthcoming paperback edition, due out in late February.)
What actually happened at the end of 2007, according to Parlante, who I interviewed today, is that OSAF and Kapor agreed that it would be healthy for the project to move out from under Kapor’s wing faster than they originally envisioned. Over the past year, Kapor has been less involved with Chandler and more focused on new projects (such as Foxmarks), and Parlante says that the project leaders felt that if they were headed towards independence anyway, it made sense to move faster.
“We’ve been joking that we laid off Mitch,” she says.
At that point, Kapor agreed to provide transitional funding for the group — less than he’d originally planned when envisioning paying the entire project’s bills for another year. That triggered the staff cuts.
How much exactly did Kapor commit? Referring to the deal that launched the Mozilla Foundation in 2003, which Kapor brokered when AOL was shutting Netscape down, Parlante says, “He’d told AOL to give Mozilla enough support for it to become viable on its own, and he decided he should follow his own advice with us.” (In 2003, AOL put up $2 million, and Kapor himself put up $300,000, to fund the Mozilla Foundation, which is now a hugely successful enterprise, based on Firefox’s revenue from Google.)
Chandler started in 2002 as a high-profile project that aimed to produce a novel personal-information manager and demonstrate that the open source development methodology could produce innovatively designed desktop software. But the work proceeded at an agonizingly slow pace, and it took about five years for OSAF to ship a usable “Preview” edition last fall.
So there’s no question that the cutbacks represent a come-down for a project that started out with such grand ambitions and golden prospects. But surely there are better metaphors for what’s happening than “pulling the plug”: “leaving the nest,” perhaps. In fact, as of the end of the month, OSAF will move out of its longtime home at Kapor’s offices on Howard Street in San Francisco’s SOMA district and become a virtual team.
From the start of OSAF, Kapor had made clear that he did not envision the project as an open-ended philanthropic obligation, but rather as a test-bed for new ideas in software design and project organization. During the time I spent at OSAF, he would regularly repeat his belief that OSAF ultimately needed to become financially self-sustaining.
Parlante says OSAF will use its transitional funding period to explore lots of different business models — everything from selling advertising or charging for service to community support (donations and fundraising) to business partnerships or deals to bundle the Chandler Hub server with other products. “Open source business models are always a little up in the air, they’re always changing — that’s true even of Mozilla,” Parlante says.
Can Chandler survive on its own? Right now I’d give it good odds for continuing in some form: as long as there are developers interested in continuing work on it, there’s no reason for it not to. It will be harder, but by no means impossible, for the organization to find enough money to support a small full-time staff beyond the transition.
The obituary writers are already chomping at the bit. And of course many of the criticisms of OSAF’s mistakes are accurate. Still, those mistakes are now history. Since shipping the Preview version of Chandler its team has sped up the flow of new releases, new features and bug fixes, suggesting that the pressure of looming independence has already made a difference. The question is, does Chandler today offer enough innovative value to build a thriving community and win support — both volunteer development effort and cash?
Given the history, it would be foolhardy to say “yes” for sure; but I think it’s also a mistake to say “no way.”
I’m a veteran of a company that spent roughly three years disproving premature reports of its demise. Salon is still around and doing good work. It’s certainly going to be challenging for OSAF to continue on its own, and it’s entirely possible that the organization will be gone within a couple of years. But writing it off today seems wrongheaded to me.
Of course, in spending so much time and thought on telling Chandler’s story, I invested something of myself in it, too, so maybe I’m just unwilling to let go. But there are other indicators that the project is unlikely to simply evaporate.
At the very moment that OSAF announced its cutbacks and the obituary writers jumped the gun, geek-hero blogger Cory Doctorow posted an enthusiastic endorsement of Chandler on the extremely popular BoingBoing. Doctorow’s post might be part of an effort to rally support for Chandler among its users now that its future isn’t guaranteed. That’s what happens when an enterprise that people care about is threatened. Or maybe it was a spontaneous coincidence — that happens too!
The big question is whether enough people care enough about Chandler to keep it afloat. In a sense, that is the final experiment in the OSAF lab, and as long as Kapor continued to fund the project out of his own deep pockets, it would never be able to get results.
“I would say I had a lot of ambitions that we wound up, for very good and practical reasons, scaling back on,” Kapor said in an interview Thursday. He described the outcome as “a working subset of a grand vision.”
Kapor said his interest in continuing waned. “We found ourselves in the situation that the team wanted to continue on very much,” he added. “I found myself in a different place. I did not have that same level of commitment and desire, because I had the original dream in mind.”
UPDATE: I did catch up with Mitch Kapor late today. He confirmed the accuracy of this account. “It’s time for me to move on,” he said. See also Kapor’s own blog post on this topic.
I’ve also received a couple of anonymous emails from people involved in the project, suggesting what I’m sure is just the tip of the iceberg of disagreements over Chandler’s direction, its leadership, and the personnel choices involved in the layoffs. Having lived through the difficulty of such events more than once myself, I know that they’re deeply emotional — and that there is often more than one “right” view of the situation. I haven’t been following Chandler closely enough for the last couple of years, since I wrapped up the book, to gauge the accuracy of these perspectives.
I also don’t yet know exactly who’s leaving. I’ll try to link to the participants’ own reports. Here’s Ted Leung’s announcement of his departure. Here’s Matt Eernisse’s. And Mike “Code Bear” Taylor’s. Taylor and Eernisse have both joined Seesmic, Loic LeMeur’s new startup. Brian Moseley, who for a long time was the central developer for the Cosmo Server (later Chandler Server), is also leaving.
Funny footnote: Moseley and I crossed paths a decade before we met at OSAF; he was part of a small development team of Cornell students who wrote the first, somewhat disastrous version of Salon’s Table Talk in 1995. He’s plainly become a very different developer in the interim. His colleagues went on to fame and fortune during the dotcom boom before going bust.
Post Revisions:
There are no revisions for this post.