Complexity

I’ve noticed that I have a bias toward complexity. I tend to try to look at problems from all the angles I can think of. Most problems are complex from at least some points of view; otherwise, they would not be problems at all. Looking at problems from all angles encourages a focus on their complex aspects rather than their simple ones.

I do profoundly believe H.L. Mencken’s classic quote: “For every complex problem, there is a solution that is simple, neat, and wrong.” For any problem, it’s very tempting to shear away the complexity and push forward with a solution that solves the simpler part. Unfortunately solving the simpler part of the problem tends to make the complex part both worse and harder to solve later.

This is a common problem in programming. I think it is the primary reason that programs become harder to maintain over time. It is very hard to keep a focus on solving complex problems rather than ignoring the complexity in order to solve a simpler problem. This is also a common problem in politics, and I’m sure we can all find our favorite examples there.

At the same time, I’ve come to realize that focusing too much on complexity is a different and more subtle trap. It encourages you to focus on the specific issues rather than the general ones. Sometimes a complex problem can be solved by solving a different, larger, problem. Sometimes you need to take a step back. The trick is to not ignore the complexity, but to be open to the possibility that it simple doesn’t matter.

When I was younger, I always felt that Alexander cheated when he simply cut the Gordian Knot. He didn’t actually solve the problem, and indeed I think the analogy is often used poorly, to suggest that it’s somehow smarter to not solve the problem. However, I’ve slowly come to realize that there is a good use of the analogy. I don’t know whether Alexander’s solution was a good one; he didn’t live long enough to prove it one way or another. But there are times when that is the right approach.

I apologize for the abstract nature of this post. I’ve been thinking of this in the context of working on Go. I think that working on the project has helped me a lot in terms of looking for the larger simplicity without ignoring the real complexity.