Lately I’ve been reading Just one thing. This is not only an excellent and enjoyable book about investing money, but there are some essays which are beautiful written and interesting for software developers. There is an essay about biases of investors. All those biases apply to software developers when they talk about their tools and about which I’ve written before.
The most interesting essay is by James Montier and is called “Psychology matters”. He quotes the main areas of biases an investor shows as Self-deception, Heuristic simplification, Emotion/affect and Social. The emotion bias certainly applies to software developers and the heuristic simplification is apparent. There are no heuristics, facts or numbers in tools, processes, language and framework comparisons, just guts. And it doesn’t matter if you take the RUP/XP discussion, the writing/generating (MDA) discussion, the Java/Rails discussion, the Groovy/JRuby discussion, the text/graphics (UML) discussion, the Spring/Java EE, the action (WebWork)/component (Tapestry, Wicket)/continuations (RIFE, Seaside) discussion, the static (Java)/dynamic (Script) discussion, the open source and license discussion or one of the countless others going on right now about what is the next big thing.
Though the most important bias seems to be self-deception. Under self-deception he lists Overoptimism, Illusion of control, Illusion of knowledge, Overconfidence, Self-attribution bias, Confirmation bias, Hindsight bias, Cognitive dissonance and Conservatism bias.
And while he’s talking about investors, all those can be applied to language evangelists and their high priests. Overconfidence comes from people with few projects. Self-attribution is their way of attributing success to their skills and tools while it was plain luck. The best quote about tool evangelism in the essay is “Heads was skill, tails was bad luck.” And overoptimism is the one thing that describes software developers best while illusion of control plagues most managers of IT projects.
So next time you read about how successfull a tool or framework is, think about the biases the author has fallen for. Good luck.