Wednesday September 05, 2012

✦ Software Engineering, The Telephone Game, and The Whippersnappers

A while back I linked to Graham Lee’s book on Test Driven Design for iOS applications. I think it’s a good book to learn from. He demonstrates insightful ways to think through how to test different pieces of software. There’s lots of examples and a bit of analysis of the cost and savings involved when unit testing the building blocks of your applications. Graham has since found that some of the cited analysis is wrong.

First of all, kudos to him for recognizing his error and going on the public record to make the retraction. Hey, even the experts make mistakes. What sets the good experts apart is this kind of action. Trust is not built on getting every thing right the first time, but quickly clarifying if something is wrong.

But the most interesting thing I got from reading his apology was how complicated it was to dig into scholarly research about the software industry to reconstruct the data he suspected was faulty. He spotted mis-citations and confusion of terms. There were circular references between papers that didn’t help advance their arguments. Some references lead to dead links and cyber squatters.

Each layer he peeled back revealed how distant he was from the actual numbers. Our industry fast became a critical piece of the invisible infrastructure of society. Yet the body of work we stand on is quite young compared to other disciplines. There’s an unfortunate telephone game going on about good software methodologies. We have much to learn in order to grow software in a way that is stable, secure, dependable, and adaptable to the quickly changing environments around it.

Does this mean we need a license to code? Should software only be developed by the wise sages locked away in the towers on the hilltops? I don’t think so, but it’s frightening to see the amount of fluff and hubris fueling large projects that affect our health, economy, and daily lives.

By the way, Graham first suspected his error while reading another book, The Leprechauns of Software Engineering by Laurent Bossavit. It looks like an excellent expose on the telephone game in our industry. I bought myself a copy.