I have been meaning to write this article for a couple weeks now, ever since the recent GNU Hackers Meeting over in Göteborg. I'd rather be hacking tonight, but some other circumstances make such a report a bit more timely; so, here goes!
I arrived a little late, having missed the first day's talks. I was particularly irked to have missed Bruno Haible's talk on modularity and extensibility, but I understand that there will be some video up soon.
I didn't know enough at that time to miss having seen José Marchesi's talk, but now that I know the fellow, it is a shame indeed. Actually that was one of my biggest take-aways from that event: that I never really knew GNU as a community of maintainers. GNU in 2009 is not the kind of organization that flies people all over to meet each other, and it's to our loss. I hope to see much more of GNU hackers in the future.
So everyone that goes to conferences knows why they are great: the hallway track. Or the bar track, as it might be. So what were the topics?
GNU is not FSF
This point really surprised me: that GNU is not the Free Software Foundation. There a relationship, but they are not the same thing, not by a long shot.
So here's the deal. In the beginning, there was GNU. But Richard realized that in the world of 1985, with proprietary software on all sides, it would be easier to defend the small but growing software commons if hackers collaborated by assigning their copyright to one U.S.-based organization. The FSF was set up as the legal entity associated with GNU, with RMS at its head.
25 years later it's still like that. The copyright assignment paperwork that every GNU contributor signs has some clauses that obligate the FSF to keep their conributions free (in the Free Software sense), but ultimately trust in the FSF is a trust in RMS, and in his principles. It is a testament to RMS's character that there are 35000 lines in fencepost.gnu.org:/gd/gnuorg/copyright.list, representing at least 5000 contributors.
But what's up now? As free software became more and more successful, it became clear that there were other ways to get involved than just writing code. There's all kinds of advocacy work that needs doing, for example. The FSF was a natural concentration point for these efforts.
However, also inevitably, with the influx of people, the composition of the FSF changed. There are very few hackers in the FSF now. There is RMS of course, whose work these days doesn't involve programming, but that's about it. Recently Mako was added, which is an important step to redressing that situation, and more on that later.
So the advocacy and campaigning is with the FSF, and the hackers are with GNU. I think every GNU hacker is really down with the message of freedom; I mean, we are the ones hacking the hack. But, as a majority, the GNU hackers are not down with "Windows 7 Sins" or "Bad Vista" or most of these negative campaigns the FSF has been running recently.
I say this as a GNU maintainer, not as a representative of GNU, but I believe I have the facts right.
GNU and RMS
All GNU hackers respect RMS. We respect his ethical principles, his vision, his tenacity, and his hack. I mean, GNU Emacs: this, for most of us, is the best software in existence. The early days of GCC. Texinfo. Really remarkable contributions, these, without which the world would be a poorer place.
(If you disagree, that's cool; but I'm just trying to explain where we come from. I guess also I should clarify what I mean by GNU here. I mean, "people who have signed paperwork to assign copyright to the FSF". I know that mentioning the FSF there is a bit ironic, but it's a strict definition.)
But nowadays, while RMS's principles remain strong (thankfully), he's a bit absent, technically. That would be OK -- he has certainly put in his hack-time -- but that the GNU project doesn't really have a means to make technical decisions on its own right now. Each maintainer can mostly make decisions about their modules, and we can talk to each other (mostly via gnu-prog-discuss), but there is no governance structure for the GNU project itself.
Worse, sometimes RMS decides things without any input at all, when such input really is needed. The decision to bless Bazaar as the official GNU version control system, for example, really sits poorly with a lot of people. The adoption of SCM and MIT Scheme as two additional Schemes in the GNU project also are real WTF decisions.
These "blessings" don't have much of an effect on people's behavior -- most active GNU modules use git, for example -- but they make people lose faith in GNU's technical coherence.
The issue here is that the GNU project is a community of people working for software freedom, yes, but we have some specific values binding us together. One, the ethical principles of supporting user freedom; more on that later. Secondly, there is a technical vision of a "GNU project" as a cohesive, well-thought-out, integrated technical whole. One can now argue about the extent to which that is currently true, but it is a very important value to GNU hackers.
Things were more or less fine when there was more of UNIX that needed to be reimplemented, and when RMS himself was more on the ball. But now there is a significant measure of dissatisfaction within GNU itself about the way decisions are made. There are some steps being taken to fix this (a recently created GNU advisory board, for example), but it is ironic that some decisions in GNU really do come from one man.
At the same time there is actually a widespread concern within GNU about what would happen if decision making were to be made more open. No one wants outsiders making decisions, it would still have to be maintainers; but there is concern that things might get out of hand, that there might be too many pointless discussions, that bad decisions could be made, that there would be a tyranny of the talkers, etc. This is why things are changing really slowly. Everyone wants more autonomy, but they don't want to lose the solidarity.
GNU and GNOME
So what's the deal with GNU and GNOME? Everything, and nothing. Allow me to explain.
GNU people see GNOME as an "outside" thing at this point. To most GNU hackers, GNOME is not quite GNU. GNOME doesn't follow GNU's coding standards (recall the point about technical integrations), assigns no copyright to the FSF, and seems to prefer LGPLv2+ and not GPLv3+. There is hardly any communication between GNOME people and GNU people. In GNOME, people look at problems their systems are having, and decide to solve them within GNOME -- PolicyKit, for example.
(I don't want to pick on David's excellent software; but I'm equally sure that it never crossed David's mind to suggest PolicyKit for inclusion into the GNU project. There are many similar examples.)
Hell, many GNU people even use things like Sawfish or Xmonad or StumpWM or other such things. Admittedly, GNU folks would be more likely to install GNOME on their cousin's computer; but as for themselves, Emacs is the only program many people need.
GNOME people on the other hand are in two camps, more or less. Broadly speaking these camps correspond to Free Software and to Open Source, respectively. The former feel a strong heart-tie with the GNU project; perhaps they started working on GNOME back when it really was a GNU project, or perhaps they started with it because they believed in GNU's ethical principles and chose that environment because they wanted to spread user freedom to their friends. As I say, most GNOME people are disconnected from GNU per se, so this tie really is more cultural than practical, but it is there, and it is strong in its own way.
The Open Source people tend to value GNOME in particular from the days when Qt was GPL-only, or even proprietary, and GNOME allowed them to develop open source applications as well as proprietary ones. Open Source people appreciate the technical comraderie, as well as the technical excellence of the platform, but often disagree with the GNU project's copyleft principles, instead appealing to individual choice as to whether to use or make proprietary software or not.
I hope I haven't been uncharitable to anyone here. Please correct me in the comments if so.
is GNOME GNU?
So, Andre posts, wondering about the relationship between GNU and GNOME. I hope I have been able to add some broader context to the question.
I think that Andre's characterization of RMS as "fascistic" is totally wrong. There are serious problems of decision-making within the GNU project, yes, but "fascistic" takes it a bit too far, and almost brushes Godwin's Law ;)
The particular context in which the discussion has been brought up, that nasty thread on foundation-list, is unlikely to bring about any mutual understanding. It is ironic that the topic is the code of conduct, which was designed to promote understanding.
Bottom line, GNOME can be GNU if it wants to. But I don't think a decision is necessary right now, and certainly not on foundation-list, not in that thread of the talkers.
I promised to talk more about hallway tracks at the GHM, but I've run out of semicolons. Until next time!