made with ChartBoot
The chart above shows the Bureau of Labor statistics on programmer employment. I am not convinced that these numbers are at all accurate. Getting reliable data on programmer employment is surprisingly difficult.
To begin with, programmer is a vague category, and it is by no means clear that everyone who worked on “programming” defined themselves primarily as a “programmer.” Secondly, the Bureau of Labor Statistics did not beginning tracking programmers until 1972, and in 1983 and again in 2000 they adjusted their categories and methodologies. For the first ten years, three broad categories (“computer specialists”, “computer programmer”, and “computer analysts”) encompassed everyone working in computing.
By 2000, these categories had expanded to include Computer and information research scientists, Computer systems analysts, Information security analysts, Computer programmers, Software developers, applications and systems software, Web developers, Computer support specialists, Database administrators, Network and computer systems administrators, Computer network architects, Computer occupations, and “all other” computer occupations. This seems to explain the decline in the number of programmers post-2000. Some of them simply got recategorized.
Follow
Thank you for the blog and book, I plan on downloading the Kindle edition after this post.
I suspect the programmer employment numbers are low, based on a little bit of digging.
I ran a query against the Stack Overflow (SO) Users table (query URL is in the website field), and it reports that 1,170,181 unique users used the site in 2012. Now, while it’s true that not everyone who uses SO is a programmer, it’s also likely true that not every programmer uses SO (In particular, SO use is worldwide, while the graph is for the US). I further guess that it’s a reasonable proxy for programmer employment, though I wouldn’t yet venture a guess as to how much this is the case. That there are roughly twice as many unique SO users as there are reported programmers does suggest that the reporting may not have caught up with the situation ‘on the ground’.
Patrick — thanks for the data on Stack Overflow. I agree that the Bureau of Labor Statistics numbers skew low, particularly in the data from the past two decades, after they began adding additional categories that cover some of the same territory as “programmer” (what is the difference, for example, between a computer programmer and a software developer, according to the BLS?).
In part the ambiguity around what makes a programmer is what makes this history so interesting. These are debated categories, and different actors and observers use the term in different ways in different contexts. On the other hand, it is frustrating not to have reliable data…