11.2 Giant Arrow

Free For All. Go to the Table of Contents. Vist the Gifcom.

This bazaar can be a powerful influence on solving problems. Sure, it isn't guided by a talented architect and teams of priests, but it is a great free-for-all. It is quite unlikely, for instance, that the guy with the overloaded printer and modem line will also be a talented programmer with a grand vision to solve the problem. Someone named Arthur only stumbles across the right stone with the right sword every once and a bit. But if the frustrated user can do a good job characterizing it and reporting it, then someone else can solve it.

Dave Hitz was one of the programmers who helped Keith Bostic rewrite UNIX so it could be free of AT&T's copyright. Today, he runs Network Appliance, a company that builds stripped-down file servers that run BSD at their core. He's been writing file systems ever since college, and the free software came in quite handy when he was starting his company. When they started building the big machines, the engineers just reached into the pool of free source code for operating systems and pulled out much of the code that would power his servers. They modified the code heavily, but the body of free software that he helped create was a great starting point.

In his experience, many people would find a bug and patch it with a solution that was good enough for them. Some were just kids in college. Others were programmers who didn't have the time or the energy to read the Source and understand the best way to fix the problem. Some fixed the problem for themselves, but inadvertently created another problem elsewhere. Sorting through all of these problems was hard to do.

But Hitz says, "Even if they fixed it entirely the wrong way, if they found the place where the problem went away, then they put a giant arrow on the problem." Eventually, enough arrows would provide someone with enough information to solve the problem correctly. Many of the new versions written by people may be lost to time, but that doesn't mean that they didn't have an important effect on the evolution of the Source.
"I think it's rarely the case that you get people who make a broad base of source code their life," he said. "There are just a whole bunch of people who are dilettantes. The message is, 'Don't underestimate the dilettantes.'"