At 12:50 09.11.2002 +0000, Stefano Mazzocchi wrote:

But I have been thinking about this for a long time! Let me add that I very much admire your attitude. One can only emphatically support your exhortations for modesty, respect for others and level-headed behavior.

Let us just not over do it. Yes, a leader will be wise to leave when he or she becomes a drag. However, there is no glory in leaving a project for the sake of leaving the project. As for your the statement that a community is healthy only when it can survive the departure of its leader, I think it is somewhat misleading. A community is healthy when its participants are having fun, not necessarily when it can provably survive the demise of its leader. Sometimes a community can survive its leader, only to fall apart a while later. The statement "community is healthy only when it can stand its leader leaving" somehow suggests that a community should force the leader to leave in order to check whether the project is healthy or not. That is probably not what you are saying or what you mean, but suggesting that leaders are somehow dispensable strikes me as over doing it. Its hard to replace people.

It's like preaching abstinence to your children only to discover that they are still virgin at 60. Maybe not what you intended as a parent. :-)

This pun introduces the next topic: software darwinism. The principle of the survival of the fittest is a cruel one. Of course its cruelty does not mean that it is wrong. It just does not apply to Jakarta. We must make every effort that it does *not* apply to Jakarta. Let me explain.

Darwin's thesis is that offspring invariably resemble their ancestors although variations can occur between each generation. Some variations may advantage the survival of the individuals showing the variation. Thus those individuals may have more offspring and may eventually form a new species competing with the original species. I am sure every one already knew this. In his book the "Origin of the Species" Darwin insists that the clarification of species is highly correlated with ancestry, not necessarily with an act of God, which was quite a shocking thought at the time. That is perhaps why he waited over 20 years to publish his ideas.

My point is that natural selection applies to a birth death process. We don't have such a process in software. What we have is a series of evolutions. Natural beings do not change their genes during the course of their life. A lion does not say, hey it would be cool to have a long neck like a giraffe, let me grow a neck. We do that in software all the time. We look at a feature or idea in someone else's code and import it into our own.

Software development is not a life and death process. We do not want that anyway. We do not want software to have 5 forks and die two weeks later. We want out software to resist time but also allow it to evolve and to improve. We want sane evolution not a jungle.

To me, Darwinism suggests either painfully slow evolution or just a pointless life and death cycle.

-- Ceki

TCP implementations will follow a general principle of robustness: be conservative in what you do, be liberal in what you accept from others. -- Jon Postel, RFC 793