■"Second Thoughts for a Designer of Software That Aids Conservation: Simpler rules may be more effective than computer models," by Eric Schmitt, New York Times, 21 September 2004, p. D2.

Great article on environmental planning that has proceeded with the aid of pioneering computer models and simulations. The program in question, Marxan, was designed by an Aussie grad student with help with a math professor. The program is considered quite good, and as such is having a revolutionary effect on environmental planning around the world.

Now that math professor is raising some caveats about relying too much on the program, saying thatóin effectóno program should override a consistent effort to revise environmental plans along the way on the basis of observed changes stemming from the original remediation steps undertaken. Sounds reasonable, yes?

It's one of the oldest debates and I run into it all the time in government planning: the "perfect plan up front" versus "ad hocism along the way." Guess what? Neither side makes any sense. When you wait on the perfect plan, you often never even get started. Or, if you actually write it up and start implementing it, the sticking-to-the-plan mentality typically becomes the biggest threat to long-term success.

I first bumped into this sort of thinking when I did reengineering work with Africa Bureau in the U.S. Agency for International Development in the mid-1990s (part of Al Gore's reinventing government program, which was really quite visionary in many ways). We're seeing the same dynamics at work now in the debate over the 9/11 Commission's laundry list of changes for the Intelligence Community ("Do it all or be doomed!" we are warned).

The analogy I ginned up for USAID used the model of planning originally developed by Bill Walsh in his legendary coaching stint at San Francisco, otherwise known as the West Coast Offense. The basic description is this: you plot out the first twenty plays of the offense, and you stick to that plan no matter how much success or failure you experience. Twenty plays are typically no more than one-third to one-fourth of the total plays you'll run, so it's like have a firm plan for the first year of a four-year plan, knowing you'll recalibrate at the start of year 2 based on the intell you've gathered in year 1. That's basically what the Walsh method was all about: testing the environment while doing your best to achieve initial success. But once those first 20 scripted plays were done, then it's all about adaptive planning. That's not learning on the fly so much as learning as you go, with appropriate breaks for rethinking and recalibration. That's why West Coast teams tend to score a lot in the second and third quarters, even as they often fall behind early in the game. Because it's all about where you finish, not where you start.

Makes you kind of wish Central Command had Bill Walsh helping them out, doesn't it? Don't worry, they're getting there.