governance should probably come first - past models and next steps
by Mark Galassi
Today I had jitsi breakfast with a group of hackers in another city -- old free s/w friends I had not seen for years, and since covid we meet frequently.
They were all happy at the creation of the Assembly and virtual backslapping went on.
A very good piece of advice given was:
"Define your governance before you make any significant decisions. Otherwise those decisions will come from the early volunteers instead of from the accepted governing body.
So maybe we do need to put something in place quickly.
It was also proposed that we take Gnome, Debian, and Python as examples and distill the best elements from those. It would be great to have a political science student write an undergraduate thesis on this, but there is probably not enough time.
I went and re-read my email archive on the Gnome foundation deliberations in 2000-07 (my recent move to emacs+notmuch makes this easy) and was able to find web archives from *almost* the start of Gnome. (Sadly the internal email links don't work anymore, but one can still form the picture.)
The whole thing started when a company Andover donated $50k to Gnome in 2000-02. Gnome had no place to put it, and Richard Stallman said he would take it and disburse it if he thought reasonable. Here is a bit of Miguel de Icaza's email on that subject:
Andover was sending the money to the FSF and the issue was being
discussed on the FSF mailing list. Richard pointed out that they
would keep the money and be willing to spend it on things Richard
considered "ok", this the actual text:
> GNOME does not have any other organization, so the funds will be sent
> by Andover.net to the FSF to be held in reserve for GNOME. Miguel
> said gnome-hackers are discussing how they would like to use the funds.
> Ok, that sounds good. Please run the final or semi-final
> recommendations past me; if they are reasonable, I will say yes.
I did not like this, first of all, because I do not think all of
the GNOME hackers might want to get Richard's clearance. I can ask
Andover to change the check and send it to Software in the Public
Interest (SPI, the dudes behind debian) if people want.
I hate more than anyone to be in the middle of this. So what do
you guys think? FSF with RMS as a filter for spending the money, or
Finally we went our way with Gnome and went through the soul-searching process and came up with a charter. I'll point here to some early drafts of that because some of it applies. (I don't think the "advisory board" of companies applies, but who knows? I would think of it as a future direction if ever.)
It's interesting to note that even then there were discussions of keeping the FSF at arm's length. At the time I weighed in with: "don't worry: FSF is there to give us leadership on legal license issues, but they won't interfere with technical direction".
Jim Gettys (one of the X authors and interested in early Gnome stuff) pointed out a different and important angle:
Also understand that some of RMS's actions in the past have alienated
major companies trying to behave in good faith. As a result, many companies
(and people) will not deal with FSF, even those who have made major
contributions to free software in the past. An arms-length approach to
the FSF is most likely to garner general support (while not sacrificing
the goals of a free desktop).
Jim was quite prophetic on this, more than I was!
This email is longish and about the past, so I'll discuss next steps in a follow-up.
Here are the links to Gnome prehistory:
* Bar Decrem posts "Charter 3rd draft":
* Summarized by Havoc Pennington at:
* Jim Getty's comments in response to my post:
* Bart and George on who is a member:
* Nat Friedman - the first practical enumeration of who is a member
* Alan Cox:
Alan makes lots of points, not necessarily that clear or on-point, but it mentions something that made my ears perk up: he discusses the LSB (Linux Standards Base). I do a lot of release engineering, and I have been very annoyed at what seems to be the the disappearance of the LSB. Coherent definition of free s/w interface points is important, and the GNU Assembly could play an interesting role.
1 year, 6 months