Yukon syl's List: opensource

• • Built by teams of workers, craftspeople, volunteers (granted indulgences by the church), supervised by architect/designers - not very isolated
• built like cathedrals, carefully crafted by individual wizards or small bands of mages working in splendid isolation

• • That's what I remember from responses to MS releases.
• Early and frequent releases are a critical part of the Linux development model. Most developers (including me) used to believe this was bad policy for larger than trivial projects, because early versions are almost by definition buggy versions and you don't want to wear out the patience of your users.
• Because he cultivated his base of co-developers and leveraged the Internet for collaboration harder than anyone else, this worked.
• Linus was keeping his hacker/users constantly stimulated and rewarded—stimulated by the prospect of having an ego-satisfying piece of the action, rewarded by the sight of constant (even daily) improvement in their work.
• In the cathedral-builder view of programming, bugs and development problems are tricky, insidious, deep phenomena. It takes months of scrutiny by a dedicated few to develop confidence that you've winkled them all out
• Sociologists years ago discovered that the averaged opinion of a mass of equally expert (or equally ignorant) observers is quite a bit more reliable a predictor than the opinion of a single randomly-chosen one of the observers. They called this the Delphi effect
• contributors for any given project are self-selected. An early respondent pointed out that contributions are received not from a random sample, but from people who are interested enough to use the software, learn about how it works, attempt to find solutions to problems they encounter, and actually produce an apparently reasonable fix. Anyone who passes all these filters is highly likely to have something useful to contribute.
• • I think this i s what makes it more powerful than the Delphi method - self-selection requires interest, knowledge and persistence in this open selection group approach.

6 more annotations...