Dave Newton wrote:
Jonathan Revusky wrote:
<snip>
I have no publicly-accessible open-source projects. If I did, I would
not give commit access to anybody that asked for it, because I do not
have the time to review the contributions of others and do not trust J.
Random Coder enough to assume that they'll do the Right Thing, because
in general, most people aren't very good programmers.
The whole idea that, when you give somebody commit privileges, that they
just go beserk committing all kinds of code of questionable quality --
this is just not something that really happens. I recognize that it
could happen. Also it could happen that you give commit privileges to
someone who is outright malicious. However, the latter would be so
infrequent really that, IMO, it's not an issue. If a wandering serial
saboteur -- the Ted Bundy of open source coding, if you will -- happens
to get involved in your project, well, I would attribute that to
inordinate bad luck, maybe like walking down the street and getting
struck by lightning. Possible, but so unlikely that it does not
condition your decision making.
What usually happens is that people sound all enthusiastic about doing
stuff and then, when they have the commit access, they simply do
nothing. That is what happens easily the vast majority of times. People
overestimate the time they can devote to something. They underestimate
the investment that it is to really get their heads around the code.
When people do start using their commit privileges they are usually
quite timid about it initially and initiate discussion on your list
prior to doing anything remotely controversial. People typically start
off doing very small localized things. And these things are not very
time consuming for the more established people on the team to review.
One thing that would be possible is to encourage people to get their
legs by doing things like working on unit tests and javadoc comments and
so on. Most projects, unfortunately, have too little of both of those
things and letting people in to initially work on that is quite low risk.
That would provide a way for poeople to gradually get into the swing of
things. I think that any people managing an open source project have to
be thinking about how to get new blood into the project.
Again, YMMV, and hopefully has!
If you have, that's great, and I'm glad it's working for you, and I
hope it continues to.
It's not just working for me. It's working for a lot of people. A lot
of people use FreeMarker, you know.
That's a pretty small sample size, but good :)
Be that as it may, apparently it's infinitely greater than your
experience running open source projects.
Anyway, this is getting sterile. I've made my point. It is my considered
view that this idea that the ability to commit code is something that
needs to be this zealously guarded is not well founded.
Probably a project like Struts would benefit from drastically lowering
the bar to becoming a committer.
The problem is that they've created this political structure where
they've defined committers as people with political power and
non-committers as people with no political power and so it has to do
with a certain clique retaining their power. It has basically nothing to
do with guarding the quality of the code.
Actually, it is probable that being politically correct (less likely to
disagree with the current clique) is a greater factor in becoming a
committer than coding prowess is.
Regards,
Jonathan Revusky
--
lead developer, FreeMarker project, http://freemarker.org/
FreeMarker group blog, http://freemarker.blogspot.com/
Dave
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
