Hi Richard, On Wed, Mar 31, 2021 at 03:21:01PM +0200, Richard Biener via Gcc wrote: > That's all true. It's still true that since GCC is a GNU project, formally > its maintainers are appointed by RMS (I've just read the official governance > structure document!).
I think this is unfair to the steering committee and misrepresents what it means to be a GNU project. That "gnu-stucture" document was written by RMS a couple of months ago and doesn't represent how the GNU project and its maintainers have worked for years. It seems to have been a reaction to various GNU maintainers getting together and discussing how the GNU project should actually be governed and how it should interact with the FSF as summarized here: https://gnu.wildebeest.org/blog/mjw/2019/12/27/proposals-for-the-new-gnu-fsf-relationship/ RMS indeed claims to be the "Chief GNUisance" of the GNU project and that that title somehow makes him the leader of the project and that he appoints GNU maintainers. But that isn't how things work in practice. GNU projects (packages) have almost total autonomy and in general decide themselves who the maintainers are. RMS simply records their decissions in the maintainer file so the FSF knows who the volunteers responsible are. In the past when RMS was the FSF president it was sometimes an advantage to discuss some (legal) issues with him, but in practice the FSF staff often had more time to actually help. I am pretty sure that is how the steering committee has worked too. Yes they involved RMS from time to time to update the FSF records for the current steering committee members and to have a more direct line to the FSF when it needed to involve the FSF for legal guidance, but that was more to make sure the FSF was up to date than to give RMS any leadership role. I do think Nathan is right that the Steering Committee should have been more clear about this up front and especially two years ago when RMS stepped down as president of the FSF. But that is water under the bridge now and the steering committee did clarify the relationship. > It's also true that the SC is only indirectly reachable, > that we didn't vote on our representatives, or that there's no traces of > its work (assuming it does any). Just to point to the pieces that > make it "not open". > > > The reality is that the governance of GCC is extremely open because > > it's performed by the developers in the community, not the GCC SC. > > And GCC is much less bureaucratic than other, large Open Source > > projects. It doesn't have multiple committees and SIGs. Everything > > is worked out among the developers. Projects are started by > > developers who take the initiative to start a project. > > > > Be careful what you wish for because it may be much worse than the > > freedom that you currently enjoy. > > I'm actually enjoying not needing to interact with RMS or the FSF > and indeed the SC appears to handle things well. But since people > are throwing in ideas to disassociate GCC from GNU I wanted to > point out that GCC needs to think of its own governance structure. I do think you have a point here. If GCC is going to disassociate itself from the FSF it needs to find a different fiscal sponsor and legal guardian for the project and that would be a good time to re-formalize the GCC Steering Committee setup. But I also think David is right. Be careful what you wish for :) Cheers, Mark
