Peter Tribble wrote: > On Jan 27, 2008 6:18 AM, Ben Rockwood <benr at cuddletech.com> wrote: > >> Currently Projects aren't explicitly owned by anyone, they can be >> endorsed by any group that wishes to do so, which is as meaningless as >> "Affiliating" with a CG. I think its of interest to know what projects >> a CG is interested in, but it does not denote ownership. >> >> Projects should, and must, be explicitly owned by a singular CG. >> Currently that ownership is simply, at best, implied. >> > > I'm not convinced. In some ways, the pre-constitution mechanism > (float an idea, get some support) was much simpler and more open > to projects. (It also didn't imply that projects were layered under CGs.) > > I'm becoming less and less convinced that the CG structure > is working, and I'm not sure that it can be fixed. So trying to > force everything into that structure no longer seems wise to me. > > Ownership of a project implies responsibility; to take on > responsibility implies having some power. The real power > for many projects will lie in the hands of the manager who > controls the budget, and are they going to be integrated? > > That said, projects should have a single point of contact. > Following the CG model, each project should have a named > facilitator. That could, I guess, be a CG. > >
Excellent feedback Peter, thank you! I definitely can see your point about power lying in the hands of managers... but that is I think too narrow a view because it assumes that all development must inevitably go up-stream. If you take that view (most of us do to some extent) its really counter to the ideals of a diverse development community. Our focus should be on encouraging development and collaboration, not in integrating changes back. If someone wants to extent SMF it's our duty to support their effort, support them, supply resources where we can, and make that work available to the broadest possible audience. To some extent the most interesting work will likely be opposed by the managers! Lets take an example. I would very much like to explore and begin working on tight integration between SMA and SMF. If I make an assumption that I need to work with the SMF team to accomplish that work, to get it approved and such, I'm just going to make life difficult for everyone. A much better method is to share my desire to work on the problem, request a Project from the SMF CG, get a repository and start working. If that work ever gets integrated back into SMF really isn't my problem... whats important is that the work occur and that people be able to use it. Think about many of the great things that have been contributed to the Linux Kernel. For a long time if you wanted a pre-empt kernel you needed to use a branch... over time it got more testing, more support, and more interest and then made its way into the main kernel. In an open community budgets are really not an issue. Having the ability and support to work is all thats important... otherwise we're not open source developers, we're consultants working for free. benr.
