Many people take it as an article of faith that there is a 10x difference in productivity between the best programmers and (take your pick) either average programmers or the worst programmer in their organization. Almost no programmers, however, are compensated at 10x the compensation of the average or lowest performing member of their organization.
That's not an article of faith. It is a repeatedly measured fact. Pick up a copy of Peopleware for an explanation of how that was measured.
That said, in comparisons of programmers from different organizations the single biggest factor in determining performance was the productivity of other programmers at your organization. Part of that is that productivity is highly environment driven. Part is that good programmers like to congregate together. Either way you are very unlikely to see 10 to 1 ratios within an organization. A ratio of 3 or 4 to 1 is much more realistic. (See Software Estimation: Demystifying the Black Art by Steve McConnell for a cite.)
Going back to the theme of the article, there is another reason for high salaries, which is to provide a disincentive for bribery. You see this in an obvious way with traders. Imagine that you're bargaining over the exact price of a $10 million bond. Wouldn't you be tempted to offer $10,000 under the table to the other guy if he'll agree that the bond is worth $100,000 more? If the other guy was receiving a normal salary, wouldn't this offer be hard to resist?
However are you going to risk the outrageous salaries that traders get offered for a small bribe like that? And if you began taking bribes, you can be sure that news would get around. (If for no other reason than that traders trade that kind of valuable information around.) Which would make it hard to ever get another job at that outrageous salary.
That's not an article of faith. It is a repeatedly measured fact. Pick up a copy of Peopleware for an explanation of how that was measured.
That said, in comparisons of programmers from different organizations the single biggest factor in determining performance was the productivity of other programmers at your organization. Part of that is that productivity is highly environment driven. Part is that good programmers like to congregate together. Either way you are very unlikely to see 10 to 1 ratios within an organization. A ratio of 3 or 4 to 1 is much more realistic. (See Software Estimation: Demystifying the Black Art by Steve McConnell for a cite.)
Going back to the theme of the article, there is another reason for high salaries, which is to provide a disincentive for bribery. You see this in an obvious way with traders. Imagine that you're bargaining over the exact price of a $10 million bond. Wouldn't you be tempted to offer $10,000 under the table to the other guy if he'll agree that the bond is worth $100,000 more? If the other guy was receiving a normal salary, wouldn't this offer be hard to resist?
However are you going to risk the outrageous salaries that traders get offered for a small bribe like that? And if you began taking bribes, you can be sure that news would get around. (If for no other reason than that traders trade that kind of valuable information around.) Which would make it hard to ever get another job at that outrageous salary.