4 ms·

An excellent programmer can routinely do the work of ten mediocre programmers Quit repeating this crap. Quit it. Look at your hands, count the fingers. How

An excellent programmer can routinely do the work of ten mediocre programmers Quit repeating this crap. Quit it. Look at your hands, count the fingers. How many? Ten fingers. Ten fingers is the reason that the number "ten" appears in that sentence and that is it. In fact, an "excellent" programmer can be from anywhere to less valuable to infinitely more valuable than a mediocre programmer. But if the market needs 5 such people and there are 6 available, then the prices will reflect that rather than some cliche that gets endlessly repeated. can go out and pound the pavement or join a start-up Thus resolving all of your and the original author's objections. Yep, that's how the market works, that's how it's always worked and always will. There's simply no use expecting every company to magically know the "right" price. There's no such thing, ever.

lallysingh 16y ago
Actually the number from my old software engineering class is on the order of 36:1 in productivity from the most to least productive members of a team.
- javanix 16y ago
  Because if there is any training that stays relevant forever it is old college software engineering courses ...
  - ColinDabritz 16y ago
    I agree that recalled college classes aren't the sturdiest reference. Here is a better cited source: "Frequently Forgotten Fundamental Facts about Software Engineering" by Robert L. Glass at the IEEE Computer Society website: http://www.computer.org/portal/web/buildyourcareer/fa035 http://www.computer.org/portal/web/buildyourcareer/fa035 "P2. Good programmers are up to 30 times better than mediocre programmers, according to "individual differences" research. Given that their pay is never commensurate, they are the biggest bargains in the software field." He also has a wonderful published book, "Facts and Fallacies of Software Engineering" I believe the number there is "28 times" as mentioned on Jeff Atwoods blog: http://www.codinghorror.com/blog/2008/03/revisiting-the-facts-and-fallacies-of-software-engineering.html http://www.codinghorror.com/blog/2008/03/revisiting-the-fact... "2. The best programmers are up to 28 times better than the worst programmers." I believe the source studies are cited in the book. I would add that the book is a great and thought provoking discussion of exactly this sort of 'old wives tale' about software engineering. Edit: I found the page with citations on google books. http://books.google.com/books?id=3Ntz-UJzZN0C&lpg=PA14&dq=facts%20and%20fallacies%20of%20software%20engineering%20robert%20glass%2028%20times&pg=PA15#v=onepage&q&f=false http://books.google.com/books?id=3Ntz-UJzZN0C&lpg=PA14&#...
    - btilly 16y ago
      There were a number of problems with the quality of the individual differences research. I would not trust that particular number. Another frequently cited source that was methodologically much better is the fact that 10 to 1 productivity differences are common even across small groups of programmers. Tom DeMarco and Timothy Lister performed this research and cited it in Peopleware. The two illuminating facts they also discovered at the same time were that the best predictor of programmer performance was the performance of other programmers at the same company, and that much of the productivity difference could be accounted for from differences in environment. (Sufficient office space, quiet office space, low interruption rates, etc.) I would trust that number a lot more. And I would strongly recommend that anyone who is interested in it read Peopleware to get some ideas about how they can improve productivity.
- dunk010 16y ago
  The source of this, afaik, is The Mythical Man-Month. You should read it if you haven't. It does get quoted and mis-quoted quite a lot.
jasim 16y ago
The productivity measure of 1 to 10 also includes the penalty of the mediocre programmers making enough mistakes that a good programmer won't make. Considering long term productivity improvements a good programmer can make - one good programmer is definitely better than 10 mediocre code monkeys.
motters 16y ago
Whenever I hear someone say "that's how it's always worked and always will", that's a red flag right there.

lallysingh 16y ago

Actually the number from my old software engineering class is on the order of 36:1 in productivity from the most to least productive members of a team.

javanix 16y ago
Because if there is any training that stays relevant forever it is old college software engineering courses ...
- ColinDabritz 16y ago
  I agree that recalled college classes aren't the sturdiest reference. Here is a better cited source: "Frequently Forgotten Fundamental Facts about Software Engineering" by Robert L. Glass at the IEEE Computer Society website: http://www.computer.org/portal/web/buildyourcareer/fa035 http://www.computer.org/portal/web/buildyourcareer/fa035 "P2. Good programmers are up to 30 times better than mediocre programmers, according to "individual differences" research. Given that their pay is never commensurate, they are the biggest bargains in the software field." He also has a wonderful published book, "Facts and Fallacies of Software Engineering" I believe the number there is "28 times" as mentioned on Jeff Atwoods blog: http://www.codinghorror.com/blog/2008/03/revisiting-the-facts-and-fallacies-of-software-engineering.html http://www.codinghorror.com/blog/2008/03/revisiting-the-fact... "2. The best programmers are up to 28 times better than the worst programmers." I believe the source studies are cited in the book. I would add that the book is a great and thought provoking discussion of exactly this sort of 'old wives tale' about software engineering. Edit: I found the page with citations on google books. http://books.google.com/books?id=3Ntz-UJzZN0C&lpg=PA14&dq=facts%20and%20fallacies%20of%20software%20engineering%20robert%20glass%2028%20times&pg=PA15#v=onepage&q&f=false http://books.google.com/books?id=3Ntz-UJzZN0C&lpg=PA14&#...
  - btilly 16y ago
    There were a number of problems with the quality of the individual differences research. I would not trust that particular number. Another frequently cited source that was methodologically much better is the fact that 10 to 1 productivity differences are common even across small groups of programmers. Tom DeMarco and Timothy Lister performed this research and cited it in Peopleware. The two illuminating facts they also discovered at the same time were that the best predictor of programmer performance was the performance of other programmers at the same company, and that much of the productivity difference could be accounted for from differences in environment. (Sufficient office space, quiet office space, low interruption rates, etc.) I would trust that number a lot more. And I would strongly recommend that anyone who is interested in it read Peopleware to get some ideas about how they can improve productivity.
dunk010 16y ago
The source of this, afaik, is The Mythical Man-Month. You should read it if you haven't. It does get quoted and mis-quoted quite a lot.

jasim 16y ago

The productivity measure of 1 to 10 also includes the penalty of the mediocre programmers making enough mistakes that a good programmer won't make. Considering long term productivity improvements a good programmer can make - one good programmer is definitely better than 10 mediocre code monkeys.

motters 16y ago

Whenever I hear someone say "that's how it's always worked and always will", that's a red flag right there.