Friday, July 15, 2005

Dave on Treehouses

Last summer I built this treehouse with my kids. We based it on a design we found in David Stiles's Treehouses, Huts and Forts.

The house has an 8x5 floor plan with a 3x5 overhanging deck (click here for a better look). It has six big windows to let the fresh air in and screens to keep the bugs out. Security features include a trap door accessible only from the ladder under the house and a dead bolt on the front door. It was a lot of fun to build and I hope my kids and their friends will have lots of fun using it.

How can I say this without sounding immodest? I humbly believe I am now one of the world's foremost authorities on building treehouses. I have big plans. In addition, to the exciting Dave on Treehouses web site (coming soon!), I am working on the following:
  • Speaking engagements at major treehouse building conferences around the country.
  • An anthology of the best writing about building treehouses.
  • Summer of 2006 internships for worthy candidates. Each intern will take part in the building of a real treehouse.
Of course I'm joking. That is, my kids' treehouse is real, but I have no plans to promote myself as an authority on treehouses. My one treehouse building experience might be archetypal, but most likely you have different trees, different plans, and a different budget.

It is the same with software development. Every software project targets a specific platform, has its own mission, and must conform to its own constraints. Yet there are lots of self-appointed authorities on software development. They apparently believe their experiences are archetypal -- that their specific methods apply to software development in general. We follow their guidance at our own risk.

Does this mean we shouldn't be consulting the experts? Far from it. I just think it is up to us to sift out the relevant, essential advice from prejudice and mere opinion. As Mr. Ed says in Thought Leaders and Thought Followers, an appeal to authority is no replacement for a well-reasoned argument.

How do you sift the good advice from the rest?
  • Know the expert's credentials. What has he done in his career? Where is he now? Although I was just poking fun at Joel Spolsky, I think he has good credentials. So do the other experts listed under Software Development at the right.
  • Understand the expert's biases. Is he biased toward .NET or J2EE? Is he for Agile Methods or against? When you understand a person's biases, it helps you sort well-reasoned argument from knee-jerk reaction.
  • Seek out opinions from people with different biases. I think this is the most important point. If you just listen to the J2EE crowd, you run the risk of becoming a J2EE acolyte. It's like listening exclusively to "red state" talk radio. Do that long enough and you'll be able to recite the party line. Your world will look a lot redder. But when you turn on NPR, you'll see the "blue staters" have some good points too. The world will begin to look purple -- which in fact it is.
So there you have it. More talk about software development with a little politics thrown in. Sorry if you thought this was going to be about treehouses, but you have to admit, that is a fine looking treehouse. Isn't it?

No comments: