Let's also imagine many Alans (corporate engineers) and Charles (hackers) and from this bigger sample, measure how consistently each approach produces an acceptable outcome.
It may be that the Charles, on average, produce better solutions with less resources than the Alans, but that individually, more of the Charles projects fail catastrophically.
Now imagine how good Charles would be if he could learn to cooperate and participate in a group like Alan's.
Let's also imagine how good Alan could be if he could learn to see past the inflated corporate policies, procedures, and checklists and focus on the crux of the problem.
Both excellent, I think. But further, let's imagine an organization that arbitrates the risk of each Charles by simultaniously working many cheap, risky Charles on the same problem and managing the results with an Alan-like process.
If the Charles are software startups and the Alan-like process is "the Internet," this could work.
It may be that the Charles, on average, produce better solutions with less resources than the Alans, but that individually, more of the Charles projects fail catastrophically.
Now imagine how good Charles would be if he could learn to cooperate and participate in a group like Alan's.
Let's also imagine how good Alan could be if he could learn to see past the inflated corporate policies, procedures, and checklists and focus on the crux of the problem.
Both excellent, I think. But further, let's imagine an organization that arbitrates the risk of each Charles by simultaniously working many cheap, risky Charles on the same problem and managing the results with an Alan-like process.
If the Charles are software startups and the Alan-like process is "the Internet," this could work.