> We have started to recognize people as having release-upload rights, even if > they aren't formal maintainers (this is really fairly recent and so far has been > handled on a case-by-case basis).
Is there a need to distinguish "releasers" and "maintainers"? I would say that once the glibc community trusts someone enough to make their releases, this person should obtain upload rights on the GNU ftp server. And then it would be logical to call this person a "maintainer", since uploading a new release is formally the most powerful act in maintaining a package (while practically, of course inspiring the development, organising the community and so on are also very important). Being a GNU maintainer has a very specific meaning, and a very specific responsibility role. It means that one needs to follow and implement GNU policies. So just because someone can upload a tarball (or push a commit), doesn't necesserily mean that one wants to take on that responsibility -- they might disagree with the GNU project but still be a valuable, appreciated and prolific contributor. We should also think about the number of maintainers in a redefined governance structure. Is there a need to restrict their number, or could each and every developer have the same maintainer responsability? If there is a fixed number, how could it be related to the size of the project? It seems natural to me that the GNU libc or GNU Guix should have more maintainers than GNU MPC, for instance. This is already decided on a case by case basis by RMS/GAC. As you say, some project have different needs than others.