Are old engineers obsolete dinosaurs?
Plenty of anecdotal evidence suggests that employers prefer younger engineers over one who is 50 or older. But a new study suggests the peril of that position.
In “Is Programming Knowledge Related to Age?”, a recent paper by Patrick Morrison and Emerson Murphy-Hill, the authors ran a big-data experiment to see if aging developers have trouble with the latest technology.
The experiment is somewhat crudely-crafted (perhaps the study's authors are graybeards). By tracking responses to questions on Stack Overflow they correlate the site's “reputation” statistics against age. Interestingly the vast majority of participants on that site are youngsters, clustered around 29.
Turns out, old folks rock.
Reputation on Stack Overflow peaks around age 50 and does show a sharp decline by 70. Even at that not-so-advanced point in life, the average is about that of a thirty-year old and is much higher than someone five years younger. While young folks show little standard deviation, oldsters reputation varies wildly, with plenty of data points well above the average (and some well below).
By noting the kinds of questions Stack Overflow participants respond to, the researchers determined that older developers have a significantly wider range of skills than young people. That levels out around age 50 and enters only a modest decline later in life. Again, the standard deviation is huge.
How much of that knowledge is about new technologies? Here the results are less clear, though the authors believe their results show age does not confine one to the tech of yesteryear.
There are some real problems with the study. No raw data is presented; it's all expressed in graphs and reduced summaries. But it sure appears that there are only a handful of people studied over age 45. And the experiment took place within the narrow confines of Stack Overflow, and using reputation as a proxy for knowledge, both of which are somewhat suspect as determining anything about the developer population as a whole.
But as one who will achieve the ripe age of 0x3c shortly, the results are encouraging. I intend to forget any data to the contrary, ignore the shorts caused by my shaking hands soldering SMT components, and continue to design MCUs with vacuum tubes.
What is your experience?
Jack G. Ganssle is a lecturer and consultant on embedded developmentissues. He conducts seminars on embedded systems and helps companieswith their embedded challenges, and works as an expert witness onembedded issues. Contact him at . His website is.