[ora600]

The gap between exceptional developers and average developers was proven in some planned experiments, as described in Peopleware.

This article, on the other hand, contains lots of philosophy, imagining and wishful thinkings - but not an ounce of proof.

The only fact is about analysts - not developers. The financial field does not make a good comparison - random chance plays a much bigger role in wall-street than in software development.

I recall reading in "Random Walk Down Wall-Street" that it has been shown that even the best managed funds don't do better than well-diversified random stock selection. Taking the high impact of chance in the investment sector into account, if someone was a star analyst at one place, it is likely that he'll perform worse at his next assignment due to regression to mean.

[ap22213]

Personally, I think that the data in Peopleware is somewhat weak, and I'd like to see updated studies. However, surprisingly what Peopleware says (iirc) is that the #1 way to increase productivity of developers is to improve the working conditions.