I have snipped most of a very interesting post to add one small tangent:
John Lee wrote:
In short, policy is a substitute for common sense.
I think I'd say this a little differently. If common sense is a road, I think of policy as the rumble strips and crash barriers. They're not a substitute, exactly, but they should follow similar contours and try to limit the scope of problems.
In my field, Alistair Cockburn borrowed the Shu-Ha-Ri model of learning martial arts and applied it to software teams learning new methods. I've found it very useful. The actual model has a rich history, but my cartoon of it goes like this. In going from beginner to master, there are three stages:
1. Follow the rules. 2. Understand the rules. 3. Transcend the rules.
In the first stage, novices have a real hunger for rules. They can't and usually don't want to understand what's going on. They just want to follow some simple steps and get the desired result. In the second stage, they still follow the rules, but not mechanically; here they gain an understanding of the purposes behind the rules. In the third stage, one understands the principles so deeply that one has no need of the rules.
So I generally think of our various policies as a harm limitation device for newbies, an educational tool for those settling in, and a convenient focus of discussion for masters.
William
P.S. Sorry for the duplicate paper link earlier; I should have read all the new posts before replying, not just the ones my mailer thought were part of the thread.