< Previous page

Unlike Chandler, most commercial software projects are profit-motivated and limited by money or time or both. Furthermore, they usually start not as a software vision but as set of concrete features or specifications that customers want. Fear, greed and pride usually drive programmers in such undertakings, especially one-shot, entrepreneurial projects like Chandler: Fear for your job or your company if the project comes in late or flops completely; greed for a big payoff if the project succeeds; pride in completing a tough job on time in the face of unyielding deadlines. Without those drivers, the Chandler team was free to engage in an outrageous amount of "axe sharpening."

Give a person six hours to cut down a tree, the saying goes, and she will spend the first four sharpening the axe. In other words, most of us would rather spend time improving the tools that make a job easier than getting on with the job itself. One of many such axe sharpenings at Chandler occurred when the team, already a year behind schedule, decided to wrap its code in packages that would make it easy for nonprogrammers to add features in the future. Hard-nosed software managers simply will not put up with this kind of axe sharpening. Let's get something out the door, they intone sternly, before we worry too much about what someone might want in the distant future.

In structuring Chandler as a programmer's paradise without the discipline of do-or-die deadlines, Kapor eliminated the all-nighters, the time's-running-out frenzy that pushes a team to finish a job. And because Chandler is financed by a nonprofit organization, no big payoffs were coming. The consequences of Kapor's good intentions were unfortunate but predictable.

The project began in spring 2001. A year later, little code had been written. "Optimistically," Kapor wrote in his blog, the team could have a fully functional version of Chandler available "by the end of 2003. Pessimistically, it will be 2004."

Kapor's projections proved to be badly off the mark. After almost three years of living with a still-unfinished project that was supposed to have been done in a year, Rosenberg realized that he could not stick around to see the project through. He gave up and went home to write his book.

However, he does provide the address of Chandler's Web site so readers can see how the project has progressed. When I checked the site on Jan. 1, the fully functional 1.0 release of Chandler was still not available.

< Previous page