“The Dollars and Sense of Continuing Education”, 1966-03 (; backlinks):
[from 2013 discussion] An engineer’s “half-life of knowledge”, an expression coined in 1962 by economist Fritz Machlup to describe the time it takes for half the knowledge in a particular domain to be superseded, everyone seems to agree, has been steadily dropping. For instance, a 1966 story in IEEE Spectrum titled, “Technical Obsolescence”, postulated that the half-life of an engineering degree in the late 1920s was about 35 years; for a degree from 1960, it was thought to be about a decade.
Thomas Jones, then an IEEE Fellow and President of the University of South Carolina wrote a paper in 1966 for the IEEE Transactions on Aerospace and Electronic Systems titled, “The Dollars and Cents of Continuing Education”, in which he agreed with the 10 year half-life estimate. Jones went on to roughly calculate what effort it would take for a working engineer to remain current in his or her field.
Jones postulated that a typical undergraduate engineer invested some 40 hours a week of study over 120 weeks in his or her degree, or about 4,800 hours total. Assuming a half-life of 10 years, Jones said about 2400 hours of undergraduate knowledge has probably been superseded. To replace that obsolete knowledge and assuming there was 48 weeks a year in which to devote on knowledge replacement, Jones reasoned that an engineer would need to spend 5 hours each of those weeks gaining new technology, mathematics and scientific knowledge if he or she wished to remain technically current. That, of course, assumed the engineer didn’t forget any previously learned knowledge that was still relevant.
Jones emphasized in his article that, “Life-long learning of engineering is possible only by disciplined life-long study and thought.” Over a 40 year engineering career, a person would need to spend 9,600 hours in study to remain current, or the time needed to earn two undergraduate degrees.
Jones hinted in his paper about the continuing issue of accelerating “knowledge decay”, which can be seen rising again as an issue in a 1991 New York Times article, “Engineer Supply Affects America”. The Times article cites the IEEE as a source when it reported that the half-life of engineering skills at that time was now estimated to be less than 5 years, and for a software engineer, it was less than 3. A few years later in 1996, Craig Barrett, president and co-founder of Intel, lent credence to that belief when he stated, “The half-life of an engineer, software or hardware, is only a few years.” In 2002, William Wulf, the president of the National Academy of Engineering, was quoted as saying that “The half-life of engineering knowledge… is 7–2½ years.” More recent estimates emphasize the low end of the range, especially for those working in IT.
Philippe Kruchten, a 30-year software engineering practitioner and manager before he became a professor of software engineering at the University of British Columbia in Vancouver, took an informal stab in 2008 at the half-life of software engineering ideas by re-examining 1988 issues of IEEE Software and trying to see which “are still important today or at least recognizable.” Kruchten conjectured in a paper he wrote for IEEE Software that the half-life of software engineering ideas is likely not much more than 5 years.
If we take Krutchen’s half-life of knowledge of 5 years estimate, and apply Jones’s formula, an engineer or IT professional today would have to spend roughly 10 hours a week studying new knowledge to stay current (or upskilling, in the current lingo). One may quibble that your study productivity is much higher than when you were in college or university, but even cutting the time needed by a quarter to 7.5 hours a week of intense study 48 weeks every year that Jones said was needed in 1966 would tax many working engineers and IT professionals today. The workload needed to keep current helps explain why the half-life of an engineer or IT professional’s career is now about 10–12 years or even less.