Mark Galassi <mark(a)galassi.org> skribis:
The entirety of what the GNU project (with the closely associated
is big. It involves at least the following:
1. branding as a GNU project
2. sometimes fiscal sponsorship
3. sometimes actual funding
4. a technical infrastructure with savannah, ftpd, httpd, mailman, ...
5. maintaining the GNU coding standards (there is some good in a
long-lived consistent set of standards, although some evolution might
I would add the Free System Distribution Guidelines (FSDG).
6. stewardship of the GNU General Public License and the other FSF
7. I'm sure there's more
While that was all available and working it was fine for GNU projects
to use the services as needed.
I think we can build things incrementally as we see fit. All these
things take time and energy to build, let’s just see how far we can get.
1. branding as a GNU project: I think that GNU is not trademarked in
the world of software, so we could establish certain criteria
(commitment to s/w freedom; a tech goal of playing well with the rest
of the free s/w world) where we list projects as GNU. Without any
need for a public split, we would just maintain a page of GNU projects
recognized by the GNU assembly, pointing out that these are projects
deeply rooted in software freedom, and with non-toxic governance.
To me, that’s the most urgent thing to work on as it will allow us to
recreate ties with other free software projects and organizations.
(Note that GNU actually is trademarked by the FSF (serial number:
85380218; filing date: July 25, 2011; registration date: April 10,
2012). You can find it by searching for “GNU” among live trademarks at
;. In practice, I think it does not carry
much weight because they haven’t actually defended it, AFAIK.)
2. sometimes fiscal sponsorship: this is a tougher one. (a) not all
projects need it, which solves the problem. (b) fiscal sponsorship
loses money, since the 10% or so that the sponsor takes from your
donations is not really enough to offer the full services of
accounting and legal matters. (c) the organizations that offer it for
free s/w projects (debian, gnome foundation, software freedom
conservancy, ...) have a particular angle and I don't think they take
just anyone. Asking the Conservancy leadership for advice might be a
good idea. (Note that although I'm on the Conservancy board, I am
writing this purely as a GNU contributor, which I have been since
3. sometimes actual funding: I think people might just have to strike
out on their own to get funding. I don't know if there is another
central source of GNU funding.
These are the two most difficult items in my view. I would definitely
delegate fiscal sponsorship to an established entity.
4. a technical infrastructure with...: this one is easy, and we will
come out better for it: not only can you host source code, but, for
example, Fosshost will let you host a jitsi (and maybe big blue
button) instance, as might others. In today's tech world there is no
need for us to maintain that infrastructure, and I recommend that we
not try to. But we could put a bit of continuing work into a good
awareness of the well-behaved free s/w mercurial and git services, and
making sure that one of these options keeps a proper archive of all
GNU project history. mercurial and git make this a straightforward
As far as archiving is concerned, Software Heritage (SWH) is doing the
5. maintaining the GNU coding standards: this could become an actual
GNU project (unlike now where it's a one-person monopoly) with a tech
review process based on the good tech review approaches used in
industry (like language standards committees, ...) with something
maybe akin to the C++ 3-year cycle, or whatever they come up with. We
should have at least 4 people on a standards "board".
That would be useful.
6. stewardship of the GNU General Public License...: this one seems
impossible, or close. The best we can do is get good advice on how to
best handle the "GPL version X or later" clause, and pass that good
advice to projects so they can decide on their licensing stance.
That one is hard because the GPL is no longer “a GNU thing”. This
responsibility has to be shared with the broader free software
OK, enough now, but I wanted to get people thinking about an
almost-neutral" approach to achieving a GNU structure as the FSF+GNU
system might become un-viable.
The only urgent one is making sure someone develop a mirror of all
that's on savannah/ftpd/httpd/mailmain/...
SWH is archiving ftp.gnu.org
already (though the tarball contents rather
than the actual tarballs and signatures) and there are other mirrors
too; Git repos on Savannah are partly archived by SWH, and one can opt
in via the “Save Code Now” API or web interface:
Mailing lists have mirrors, but we could very much set up a cron job at
gnu.tools that downloads mbox files periodically from lists.gnu.org