There’s a great article on ReadWriteWeb that illustrates a couple of my recurring themes: how darn difficult it is to measure things, and also the fallacy of the idea that developers merely “go where the money is”.
The piece tries to identify what are, currently, the most popular programming languages. It presents five sets of data: analysis of the languages that are used in the code that is stored on the GitHub code management service correlated with the numbers of questions about particular languages on the dev guru site StackOverflow; the results of a simple survey by Hacker News which asked the question “What is your favourite programming language?”; analysis of job adverts on the Dice.com recruitment service; sales of programming books published by O’Reilly; and analysis of Google search terms.
There is a reasonable analysis of the pros and cons of each of these methods, and it concludes that:
“Java is still top dog, and developers don’t seem to completely hate it, though it’s sure not their favorite. C and C++ remain popular, both with developers and employers. Python and C# also provide a good trade-off in terms of popularity with developers and employer demand.”
It’s that last sentence that I find particularly interesting – the idea of a trade-off between popularity with developers and employer demand. If it were true that developers merely followed the money, then the “favourite” languages in terms of the question posted by Hacker News should just equate straight to the skills most in demand (where the money it). They don’t. By quite a margin (numbers one and two on the Hacker News survey – Python and Ruby – only appear at numbers 10 and 11 on the Dice.com stats).
Anyway, some interesting food for thought, even if one does oneself quite quickly using statistics in the way in which a drunk uses a lamppost – more for support than illumination (a lovely quote from Vin Scully, American sports broadcaster, that I recently heard at an event).
mmitII 