Glyn Normington wrote: > > People giving up *really* gets to me, especially as I must take partial > responsibility. Not only is Axis the worse, but the people involved > presumably go away feeling their contributions were unwanted, which is a > great shame. > > I'd like to play my part in fixing this, but since Axis is the first open > source project I've worked on, I'm still learning the ropes. What should I > do when someone sends in a plausible patch in an area that's unfamiliar to > me but the other commiters appear to ignore it? Should I vote +1 and if > there are another two positive votes, take that as a mandate to commit the > change? I guess the problem is I'm hesitant to change code that I don't > personally understand well, so I tend to leave this to the experts in a > particular area, even though they may be too busy. Do I need to have > thoroughly understood any change I commit or is it reasonable to put some > trust in the submitter of the change?
Motto: it is easier to get forgiveness than permission. Corollary: it generally is easier to back out a change than it is to get three +1's. I assert that it is reasonable to put some trust in the submitter of the change. Background: the first open source project I was involved with that had this problem was Jakarta Tomcat. That may surprise people who have seen how Jakarta in general and Tomcat in specific operates today. What made me different is that I did not give up. By sheer persistence, I wore down the people who were committers, and they ultimately gave in and made me a committer. Once that was accomplished, I committed anything that anybody contributed that didn't seem obviously broken and didn't cause any tests I ran to fail. After I got more than a couple of patches from any one individual, I nominated them for committer status. Now Tomcat has an abundance of largely inactive committers (including myself). This is not a problem. What would be a problem is a lack of active committers. > On a slightly different tack, I was approached on two separate occasions > with offers of help to implement the JAXM SOAP interfaces, but when I > explained how to make a start, I heard nothing more. Perhaps newcomers > worry about communicating openly on axis-dev? How can we cut through the > mystique and let newcomers know that the rest of us are just ordinary > programmers doing our best, understanding some things but not everything, > and occasionally messing up and helping each other fix things up? For some > reason, just saying this doesn't seem enough. One of the great mysteries of life. - Sam Ruby
