> Just as an aside, the preference is for Apache projects to discuss their way 
> into a decision, rather than appeal to the process.
> 
> Ideally, people should chat about things on the list, until someone does 
> something about it. If no one comments on the commit, then the decision is 
> made. We call this "lazy consensus".
> 
> The key to this, of course, is that all the commits flow through the DEV 
> list. The commits are part of the development discussion too. The only part 
> that matters, really :)
> 
> The Jakarta project tends to be a bit vote-happy. Many Apaches consider all 
> this voting slightly scandalous. Outside of a release, voting is seen as a 
> sign that discussions have broken down. Following the formal process is 
> considered a last resort -- what you do when discussions don't seem to be 
> working, and the team is not of one mind.
> 
> A release is a little different. Since there are different grades of release, 
> a release will always require a vote. But these are usually majority votes, 
> and a -1 is not a showstopper. (Though, the goal is always for the members to 
> be unanimous, or to find a way for them to become unanimous.)
> 
> The bottom line is that a project is not a government, it's a team, and a 
> team should be of one mind.

Well put.  Voting on releases makes sense and the lazy consensus seems
to work well on projects I've been involved with.

Right now I notice that the JIRA issues are not sent to the dev list. 
Nor are the checkins.  Is that going to change?

> -Ted.

sean

Reply via email to