5 ms·

I've been inspired by Gladwell's more general answer from the book Outliers: "Gladwell repeatedly mentions the "10,000-Hour Rule", claiming that the key to su

• kragen 15y ago
  Like most of Gladwell's ideas, it falls apart under the least examination. Consider a neurosurgeon. Perhaps after ten thousand hours of surgery she was an expert in surgery. But perhaps only a quarter of that time was spent performing neurosurgery. Does that mean she's not an expert neurosurgeon? Perhaps, then, after ten thousand hours of neurosurgery she has become an expert neurosurgeon. But maybe only 5% of that time consisted of excising glioblastomas from the anterior cingulate cortex. Does she need to spend ten thousand hours doing ACC glioblastomectomies in order to be an expert at that task? In short, because what we normally mean by "tasks" or "fields" can be broken down into smaller tasks or fields, the statement is self-contradictory if they mean what they normally means. But if we search for a definition of "task" that makes the statement true, we can trivially find a circular one: a "field" is something that a talented person can achieve mastery of in about ten thousand hours of practice. Once reformulated that way, it becomes clear that this is not a statement about human expertise or expert performance, but about the economic structure of current society, and in particular our degree of specialization: most people (within the professional class Gladwell finds interesting, anyway) work in jobs that require about ten thousand hours of focused practice by a talented person to reach mastery. (I should point out that this occurred to me when I saw Jeff Fox (RIP: http://groups.google.com/group/comp.lang.forth/browse_thread/thread/622263caccba35b5 http://groups.google.com/group/comp.lang.forth/browse_thread...) claiming that the research Gladwell summarized meant that you needed about ten thousand hours of programming in Forth to be good at programming in Forth, so you shouldn't put too much stock in the opinions of experienced programmers who didn't have much Forth experience.) Once stated this way, we can formulate interesting hypotheses about the cause: perhaps when a field starts to require substantially more than ten thousand hours of practice to master it, its practitioners start to specialize in a part of it, so that they can reach mastery sooner.
  • imjk 15y ago
    Is it unreasonable to think that if someone has put in ten thousand hours into a general field such as programming (across various languages, as opposed to just focus on one language) that there are enough general principles to extrapolate for them to be considered an expert programmer? Is your argument against Gladwell's idea (not really even his idea) that the professions he site has too many sub-components that under such a definition no one could achieve ten thousand hours of mastery? Perhaps I'm not understanding your argument. Regardless, it's interesting that you bring up the example of the neurosurgeon because Gladwell's specifically written about the mastery of expert surgeons including the example of Charlie Wilson, a highly regarded neurosurgeon. Included in his telling of the 2987 transsphenoidal resections of pituitary tumors that Wilson performed, Galdwell highlights the seemingly thousands of hours Wilson practiced surgical procedures on rats early in his career. You can read the article here: http://gladwell.com/1999/1999_08_02_a_genius.htm http://gladwell.com/1999/1999_08_02_a_genius.htm
    • kragen 15y ago
      > Is it unreasonable to think that if someone has put in ten thousand hours into a general field such as programming (across various languages, as opposed to just focus on one language) that there are enough general principles to extrapolate for them to be considered an expert programmer? Well, maybe. But maybe Jeff Fox is right, and you really need ten thousand hours of programming in Forth to perform at an expert level programming in Forth. And, given IBM's historical turnover, it wouldn't be surprising if performing at an expert level maintaining DB2 required not just ten thousand hours of programming, but ten thousand hours of programming in PL/M on DB2, in order to know your way around that particular codebase well enough to perform well. Certainly the ramp-up time is likely to be longer on that codebase than on something more modern. But maybe we should go in the other direction — maybe Bill Gates is right, and it only takes a year or two of programming to perform within an epsilon of your best possible performance, and it's really something broader, like "engineering design of artifacts with human users", that needs a whole ten thousand hours of practice to master — not merely programming as such. My point is that these can't all be correct. Being an expert at engineering design of artifacts with human users implies that you're an expert at programming, which implies that you're an expert at programming in any context: in Forth, in PL/M in the core of DB2, on the space shuttle. Being an expert surgeon entails being expert at tying sutures, but it surely doesn't entail spending ten thousand hours practicing the tying of sutures.