Java.sun.com, Sun's main Java site, has a fascinating interview,
(http://java.sun.com/developer/technicalArticles/Interviews/bloch2006_qa.html)
with Google Java architect, Joshua Bloch, in which he explores his
latest book, "Java Puzzlers," written with Neal Gafter. He models
puzzlers on optical illusions; you look at a piece of Java code and see
one thing, but then on closer inspection you realize your mind has
deceived you. The idea is to trick the mind into making a mistaken
assumption about code; you see your mistake and will hopefully never
make it again. With optical illusions, Bloch claims, understanding how
the illusion works does not change the illusion. But if you understand
how a "code illusion" works, you can stop making that mistake, and spot
it in future code that you review.

Bloch claims that most of the puzzlers were derived from real
experiences of developers and that the biggest problem for developers
is unwarranted optimism.

Speaking of unwarranted optimism, it's fascinating that they fell into
their own trap in one of the solutions to a puzzler that turned out to
be broken; they had to include an errata in their book correcting it.
Bloch argues that no one is immune. There are also interesting comments
about language design decisions, about.

Finally, I'm curious: Has anyone read the book and found that it
actually protects or inoculates them against making programming errors,
as Bloch claims?