How to read reviews

In my roles as author, columnist, instructor, track chair, I have gotten a lot of practice interpreting reviews. Here's a few rules of thumb:

  • If you are reading numeric scores, pay close attention to extreme deviation. I would rather see a talk that got 20 fives and 10 ones than one that got 30 fours. Ted Neward is a champion in this regard, consistently inspiring reviewers to write in new numbers on both ends of the scale.
  • If you are reading prose reviews, learn how to spot tells. If reviewer X says several illogical or ignorant things, then it is a reasonable optimization to stop crediting them on other subjects. This will not be 100% accurate, and may be unfair to individuals, but remember your time is limited.

I was reminded of all this by Giles's observation that

SICP has 82 5-star reviews on Amazon, 53 1-star reviews, and virtually nothing in between.

Of course I had to go read a few of the one-stars. The very first one begins:

I don't understand what others see in this book. I realize that it is *supposed* to be about foundations of good programming style, and not about any language in particular, but being language specific is exactly what it does. Many things it teaches can only be done in Scheme.

Apply the rules of thumb above...

Get In Touch