Anablog

About a year ago Michael Pollan wrote a great article about food and nutrition. I agree with his conclusions but I also admire his writing style. His entire theses is stated in the first seven words of the article:

Eat food. Not too much. Mostly plants.

 That is what EDN web editor-in-chief Matt Miller has been trying to teach me— the pyramid writing style where the main point is up front and the support comes later. Too many engineers use the drama style, like an Agatha Christie mystery, where you don’t know what they believe or are trying to communicate until the last paragraphs. Maybe that is good in the movies, but it sucks in technical writing.

The other thing I wanted to bring up was that this fellow Pollan is an analog guy. He is brave enough to deal with the subtleness of a world of tradeoffs and compromises and complexity, instead of the purely digital dogma of the one-zero crowd. What a delight. He does not say that you should eat no meat like a vegetarian. He does not say that you should only eat meat like an Aitkin’s diet. The answer is in a blend of solutions. Just like the analog problems you folk deal with every day. The real joy is seeing this guy get his due— I saw him on PBS last night touting some book or something, but it seems that people are finally ready to admit that things are pretty complex and you need to do some thinking and study to understand them. But just like our analog jobs, the satisfaction when you figure things out in this complex world is much more satisfying.

One really remarkable observation he makes regards the limitations of reductionist analysis. Reductionism is the basic tenet in the theory of experiments where you try to isolate variables and look at them just one at a time. Pollan rightly points out that this is a very shallow tool and that most nutritionists realize that it is the complex interactions of the variables that matter. This brings to mind when Bob Pease debunked the Tuguchi method of improvement. Taguchi would take each component in a power supply and loosen its specs to make it cheaper. But he neglected to notice that when you did that to all the components acting together, the design no longer worked.

Another remarkable thing about that article is that he admits that journalists themselves have been a part of the problem. Rather that verifying this by overstaying my welcome, I will leave you to enjoy the article.