How many developers think different?

Posted on March 17, 2018 by Rick Jelliffe

The StackOverflow yearly survey came out today. This week I have been writing some posts thinking about what modes of thinking, jobs, technologies, debugging strategies might be suitable for developers with smaller working memories or below average short-term-memory-consolidation.  And I have suggested that the emphasis in the hiring exams of many high tech companies on missionaries-and-cannibals or Hackerrank-style coding problems favours analytical thinkers over synthetic thinkers (and therefore hackers over software engineers, not to mention impulse over experience, and even therefore instutionalized discrimination of people with a disability.)

So the survey lets us give an estimate to this: 5.9% of developers (over 100,000 respondents) reported they had a memory or concentration disorder (that they would class as bad enough to be a disability). 2.1% identified themselves as autistic: I am not sure that self-diagnosis is very meaningful.

So developers with some problems with analytical tasks (and presumably relative strengths in synthetic tasks) are at least 6% of the developer population. It is not a competition, but that is more than the number of professional developers who reported themselves as gay or black combined: there were almost as many people with memory problems as there were women.  Of course, it is not a race. And does a high number indicate that people with this disability are actually being hired and not discriminated against?

One way to tell is of course to do a sanity check against numbers in the population. So the numbers identifying as gay and lesbian seems to line up with the general population. About 12% of the US population is black, so the low number of black developers indicates some disconnect somewhere.

The numbers of autistic people capable of work is now calculated at about 1% of the population (1 in 45 autistc, with the estimate rising each year as the standard broadens, and with 70% of autistic people not employed), mostly men. Programming seems a good fit for autistic people, so that 2% figure seems small: is there an issue that some non-Western countries may not have decent anti-discrimination ethos, and that therefore offshoring in fact can be seen as an anti-discrimination avoidance mechanism for Western firms? I hope not.

The population estimates for people with ADHD/ADD/dyslexia kind of memory problems is similar to the figure of 6% of developers reporting themselves to have a memory disability is entirely credible.  What it suggusts, I think, is that people with these memory defects can still do useful jobs as programmers, and that hiring test practices that exclude them may cheat companies out of what they have to offer. To successfully hold down their jobs, these people must have figured out strategies and technologies that let them cope and perhaps thrive.

Is this where pair programming fits in?  Sharing the cognative load?