On 03/25/2012 01:14 PM, Peter Stuge wrote: > Jean-Michel Pouré - GOOZE wrote: >> iterative modifications and evolutions. This only happens if the >> first version of a patch is committed fast and spreads using the >> Internet. > WTF? > > This goes diametrically against the goal of software quality.
Please calm down. Software quality is a complex issue, and sometimes does not work the way one would intuitively think. A good example is this false assumption you were building on. History has shown that the bazaar model is capable to provide quality software, and in a quite wide range of settings it is even more capable than cathedral. We also seen examples of very capable programmers unable to keep up with the bugs in their code, because lack of resources. Think about Hurd for a while. Bazaar shown to be more effective in code quality for a number of reasons: - it is more capable to bring in resources. programmers, reviewers, testers - its turn-around time is much faster; the same patch might get written in a faulty way, tested, fixed and enhanced in the same time while in cathedraal it gots written right at the first shot (but still not enhanced) http://www.catb.org/~esr/writings/homesteading/cathedral-bazaar/ with which I guess you are already familiar with. Moreover: code quality is traditionally defined in a non-functional way. However (at least to me) code quality also have a functional ingredient: how the code functionality fits the users' needs and expectations. We have plenty of proofs that bazaar just cannot be beaten on this one. For an example compare the Linux and FreeBSD user base (though I guess even the FreeBSD project does not count as a cathedral anymore). _______________________________________________ opensc-devel mailing list opensc-devel@lists.opensc-project.org http://www.opensc-project.org/mailman/listinfo/opensc-devel
