Community
December 4th, 2002
As for the second part of the article, I'm not sure if I agree with Andrew. I think territory is a good thing. Not "this is my code, and you can't touch it", but "this is my code, and if you want to touch it, please pay respect by trying to understand it before you change it". Why one thing was done in one way, and not another one, is often caused by zillions of reasons, some of which are not obvious to one who steps in head over heels. Let's take the oh so popular car analogy: if I look at a car, I see some thing on four wheels. Why four? A plane is formed by 3 points, so I can easily remove one wheel. I might have to move the other 3's position a bit, but I will end up with a car that can drive straight forward. Vwala, I've fixed a bug in the car design! And then the next person steps in, and tries to make a turn... He'll experience why there were 4 wheels in the first place.
I guess what I'm trying to say is this: not all code breaks are maliciously done. In fact, most are not. But many code breaks are caused by somebody "fixing" something without a perfect understanding of why it was that way in the first place. You have to consider that before saying "lets open all code to everybody". Somebody who has proven to have a clear understanding of Log4J, doesn't necessarily have a clear understanding of how a servlet container works.
What does need to be considered, however, is the openness of the "get commit access" procedure: if a Log4J committer comes knocking at Tomcat's door "hi guys, I want to fix code I broke in Tomcat", the treshold should be very low to let him in.
Disclaimer: I'm unknowing of the actual situation in Apacheland. Consider this as "an outsider's theory". I don't want to pretend that I can come in head over heels and fix everything.
Leave a Reply