Here are my relatively quick thoughts on this. Approval speed: This is a balance between doing unnecessary or unused work vs speed of getting this done. People could generate code and post patches without any prior discussion at all. Doing that has a much higher risk of the community at a whole saying that patch is not interesting, and should not go in CVS. The opposite extreme is that there is a long/thought out approval process, but when a yes decision is reached, its pretty much 100% sure that the code will go in CVS. What's the correct approach? Depends on numerous factors. Taking in unsolicited code certainly works, and has happened in the past. But that situation certainly works much better when you have lots of coders with free time on their hands and who won't complain too much if the patch is not accepted. Opposite case is better when you have fewer coders, so that the limited resources efforts are not wasted. Unfortunately, I think we are more in this case than the former. Given that we are in a limited resource situation (IMO), I tend to like to discuss things more. I've probably seen enough ideas recently to keep several people busy a good many hours a week. My last comment on titles was to address that. I don't see any problem with someone doing the code, but the question is then if it is worth it if no one will use it. If you decide that this is still something to do, and what to do it yourself, I have no problem with that code going in. Unfortunately, it is tends to be very hard to know when someone is proposing idea because they will write the code for it, and when someone is proposing an idea that they think should get done, but are hoping someone else will code it. I think approval and discussions will get cut much shorter if you say 'here is an idea I want to code - what do others think?'. This will at least help me out a lot - some of the lengthy discussions I partake is because that is not clear, so without anyone clearly willing to code it, this means it goes in the main queue of things to do which for the most part leaves it to me to do. And given that, I then need to try and prioritize that queue, so I try and see how important the job is. Related, if you are a coder and see someone else post an idea that you are willing to code, please say so. That will once again help things out. And also remember that I may forget the original message saying that person X was going to code it if the conversation goes on too long (sometimes a conversation/idea is idle for a week, and someone comes back from vacation, brings up some new interesting points, which starts a conversation again.