The Full Story Behind the Canonical vs. GNOME Drama

supersloshy writes this followup to our Thursday discussion of friction between Canonical and GNOME: "I've seen a lot of GNOME bashing for various reasons here on Slashdot as well as several other websites. The problem with all of this is that you never hear GNOME's side of the situation, making a lot of disrespectful comments about GNOME (or the others involved) rather baseless and illogical. Dave Neary has an extremely thorough blog post which details problems on all sides that make the issue much more complicated than 'GNOME is being idiotic by not accepting our technology.' The points covered in the blog post include, among others, how Freedesktop.org is broken as a standards body, that Mark Shuttleworth doesn't understand how GNOME works, that GNOME is not easy to understand, and that open discussions from the very beginning are important for specification development and adoption. Another blog post by 'Sankar' also covers similar points while defending GNOME."

Comments

For those without the patience...

[Spyware23]

For those without the patience to read this article (which is much longer than I intended it to be when I started!), here are the headline points:

-FreeDesktop.org is broken as a standards body
-Mark Shuttleworth doesn’t understand how GNOME works
-GNOME is not easy to understand
-Deep mistrust has developed between Canonical, GNOME & KDE
-Difficult people are prominent in each of these projects
-Behind closed doors conversations are poison
-For people to work together, they need to be in the same place

Pulled from http://blogs.gnome.org/bolsh/2011/03/11/lessons-learned/

Re:For those without the patience...

[segedunum]

Which is kind of what you'd expect from Gnome - claim that Freedesktop doesn't work and that no ne understand how Gnome works, it's all just a big misunderstanding and everyone is equally to blame.

Bullshit.

Re:For those without the patience...

[Anonymous Coward]

Freedesktop.org is not a standards body. From their website, "freedesktop.org is not a formal standards organization[...]" Anyone can get subscribe to the mailing list and propose a "standard." Anyone can get a Bugzilla account and a git repository there. They encourage "de facto specifications" - if you're already doing it and think someone else should do it too, write it up and send it out.

Mark Shuttleworth does not understand how GNOME works. He believes there is technical leadership, either somewhere in the GNOME foundation or elsewhere. There isn't. New directions are taken in GNOME by people who are able to motivate. If you can't code and can't convince someone else to code with you, you won't get anything done in GNOME.

GNOME is not easy to understand. This is evidenced by Mark Shuttleworth, a very smart man and a leader of one of the largest GNOME-pushing Linux distributions, not understanding how GNOME works. For that to happen, GNOME must be difficult to grok. It's also evidenced by the chatter on Slashdot and elsewhere that suggests that GNOME's leaders should step up and lead. There are no leaders. No one is in control. Maintainers of individual modules do what they want, guided by consensus, intuition, and experience.

Deep mistrust has developed between Canonical, KDE, and GNOME. The last few weeks' explosion has made that clear. Everyone mistrusts everyone else.

Difficult people are prominent in each of these projects. GNOME has their fair share of jerks. Check the mailing lists around the time of the StatusNotifier spec proposals and you'll see who's who. KDE also has their fair share of jerks. See the same mailing lists at around the same time. Canonical also has jerks, although this may be less clear. Their employees tend to have less of a presence on public mailing lists than GNOME or KDE people. To many in the free software community, not developing upstream or contributing upstream makes you a jerk. That's up to you to decide.

Behind doors conversations are poison, as we can see from Mark Shuttleworth's blog post. He claims that Ted Gould (a Canonical employee) had a conversation with Jon McCann (a GNOME developer) and that in that conversation, Jon said libappindicator sounded great. Mark wasn't present at this conversation. Jon McCann claims the conversation never happened. As far as I know, Ted hasn't weighed in yet (but I doubt he'll contradict Mark). We outsiders can have no idea what really happened. What we do know is that, without a public record, this is a huge clusterfuck of a communication problem. If they had held the conversation on IRC with a logger, or on a mailing list, or documented the conversation on a mailing list or a Wiki afterwards, this whole blow up might not be so painful. in the free software world, conversations must be public.

For people to work together, they must be in the same place. The Ubuntu Developer Summit, GUADEC (GNOME coference), aKademy (KDE conference), and regular hackfests (often sponsored jointly by Canonical and the GNOME foundation) show that being together in the same place helps.

I'm not sure what you're calling bullshit in your post. You didn't really explain. I've tried to show that the claims in the original article seem to be true. Perhaps you think the claims don't go far enough, and the author should have singled out GNOME for punishment or blame?

The problem is that both sides are wrong ...

[tomhudson]

GNOME has a long history of "NIH" (not invented here), and Canonical has developed a reputation for trying to boss developers around (like when they wanted all the major projects to sync their release schedule with Ubuntu).

In the end, they're both going to be irrelevant. GNOME shell is too late, and doing it their own way, going further away from what most people want in a desktop, and Unity is already outdated when you compare it to what's happening in the tablet world.

So a pox on both their houses. They sort of deserve each other.

Re:The problem is that both sides are wrong ...

[GooberToo]

GNOME has a long history of "NIH"

That was all started back when Miguel was creating the project. Don't get me wrong, I prefer Gnome to KDE but Miguel really screwed things up with one exceptionally poor decision after another - almost all of which is firmly rooted in the NIH-mantra. Its a tradition many were seemingly happy to carry on. Sadly, NIH is almost always a sign of significant personality flaws. This is turn easily feeds back into the lack of leadership, lack of formalization, lack of documentation, and lack of communication. All are traits of people who are actively trying to hide their work from peers until, hopefully, it can be forced down their throat within minimal review and criticism. Though admittedly, there are other factors which can justify this type of behavior too; especially if you're trying to jam it down the throats of completely unreasonable NIH personalities. Doesn't make it right, but it doesn't make it understandable.

Again, I use Gnome daily but it doesn't change the fact that Miguel created a really bad architecture, creating poor implementations of various technologies because of NIH, then created a collection of like-minded yes men, who then proceeded to create one kludge after another trying to fix the cluster fuck of bad ideas and implementations originally created by Miguel. Literally, the project has been primed for this cluster fuck since its inception. And you note, both editorials clearly say, "NIH" and that the GNOME project is structurally broken. I've been literally saying this for years now. I've not followed Gnome for several years now and none of this is the least bit surprising to me. Not one bit.

While the telecommute comments are pretty dumb, he is correct in that communication for any large project is required - though physical presence absolutely is not., contrary to his assertions. Regardless, they desperately need to adopt some formalized process if they hope to salvage the project. Looking at projects like Python's PEPs would be an excellent start. But then again, for them to move in that directly likely isn't possible with the current mix of personalities because it means being reviewed and criticisms early and often, which all too often is in stark contrast with the personalities involved.

I seriously hope they get things resolved. But if they don't, KDE is looking pretty strong these days. I just don't want to have to have both sets of widgets and frameworks loaded to maintain my preferred applications.

Fork it, minus all the whining

[ChunderDownunder]

Gnome weren't interested? If it matters to Canonical so much, why not just contribute the necessary support into the core libs? Refactor the gnome library so it supports both the gnome way of doing things and this new-fangled KDE/unity way and can be pluggable. Strict Gnome implementations can do it their way or link your lib.

If Ubuntu Gnome desktop (even running gnome-shell) is nicer that official Gnome, your fork will be adopted by other distros and thus 'win'.

Aaron Seigo (from KDE)'s perspective

[chrisl456]

blog post 1 and blog post 2.

Enjoy.

Re:Aaron Seigo (from KDE)'s perspective

[segedunum]

Well, it is his business considering that he and lots of other people within KDE have contributed greatly to Freedesktop and it has meant that those of us trying to use desktops with applications with differing toolkits that can actually work reasonably together. For some Gnome people to then wander along and say 'Freedesktop is broken and doesn't work' is simply not helpful in the slightest, nor does it cover them in any glory.

Oh, and Aaron has consistently been critical of Canonical over a long period of time over a lot of what they've done. That hasn't changed, although they share a little common ground here.

Poorly if at all

[sjbe]

If people need to be in the same place to work how does telecommuting work?

Poorly in most cases. Telecommuting can work in some cases but only for cases where the need to communicate is either minimal or well defined. I've telecommuted (worked from home) and I'm nowhere near as productive. Most jobs involve a significant amount of communication and it is MUCH easier to communicate in person. Emails and phone calls are great but there is no substitute for face to face communication and close proximity when collaborating on a project. There are exceptions where telecommuting works great but they are and will remain exceptions for most of us.

Difficult people are behind every project it is called pride, get over it

There is pride and there is arrogance and they are not the same thing. Being proud of what you have done doesn't give you or anyone else the right to be a jerk.

Mistrust develops because one side does all the work while the other complains about it.

Mistrust is caused by many things. Every argument has two sides and in almost every case both sides have a least some (though rarely equal) legitimacy to their arguments. People are political animals by nature and if they aren't able to talk about what they are doing AND their motivations for doing so in an efficient manner, mistrust is the inevitable result.

Re:Kubuntu

[clang_jangle]

I don't use ubuntu but I support a bunch of people who do, and usually recommend xubuntu, which has the xfce4 desktop. Ubuntu users might want to get familiar with it now so that when gnome follows kde in snatching defeat from the jaws of victory you'll be undisturbed. sudo apt-get install xubuntu-desktop for ubuntu or kubuntu users.

Re:Gnome does it again.

[Gaygirlie]

KDE is interesting and all but GNOME lets me actually get work done; it wasn't designed to appeal to a 3-year-old.

Indeed. I have myself tried KDE every now and then, all the way from KDE 3.x, and I still don't quite like it. It feels like they are simply trying to include way too much stuff and every imaginable configuration option. GNOME on the other hand feels much more consistent and clean and thus it suits my taste much better. I have only wished they'd switch over to Qt because it's quite a bit more powerful than GTK+, but in such a way as to keep the current feel to it all.

However, it is a moot point now: with GNOME3 removing minimize and maximize buttons and more-or-less forcing people into using workspaces I will have to seek a replacement. I just happen to like minimize and maximize buttons and my current workflow, I don't want to have to learn a new one just for the sake of it being new. It would be a different matter if it was somehow more powerful and efficient than my current one, but it isn't.

Oh well.

GNOME's own alternative?

[ThePhilips]

Dave Neary has an extremely thorough blog post which details problems on all sides that make the issue much more complicated than 'GNOME is being idiotic by not accepting our technology.'

Let's cut the chase: does GNOME provide an alternative notification area spec?

From all written, I can really comment only on the part about "fd.o is broken as a standards body". And all I can say is that pretty much all standard bodies work like that: they rely on cooperation. GNOME didn't take part in talk and later sent list of complaints - instead of drafting new (version of) spec. And GNOME has stopped there, at sending complaints. Standards and specs are not immovable targets, while apparently GNOME childishly refuses to take part in the process by only complaining and calling it "broken." Or I miss something?

Where's the benevolent dictator?

[Troll-Under-D'Bridge]

From TFA:

But then again, over the years I have heard similar feedback from GNOME Mobile participants, and people in Nokia --so it's not all Mark's fault. As Jono says here: GNOME does have a reputation of being hard to work with for companies -- no point in denying it (then again, so does the kernel, and they seem to get along fine).

Leaving aside the question of whether it's good for an open source project to have macho leadership, I think the comparison with Linux (the kernel) isn't valid. Linux, as every slashgeek well knows, is ruled by benevolent dictator. What Linus wants, Linux gets. Or you fork the kernel, which is what most everybody does. I think the last Gnome BD was a guy named Miguel, who has since gone on to other interests.

But perhaps more substantively, Linux differs from Gnome in that Gnome tends to be modular, while Linux is modular only in the sense you can do "modprobe fu ; rmmod bar". So even if Linux didn't have Linus, people are forced by the monolithic nature of the kernel to be more careful with the bits they insert or remove from the kernel. Modifying the kernel is a more surgical operation when compared to the more Lego-like nature of Gnome.

Gnome's modular nature thus makes casual forking (as practiced by Canonical, et al) easier than it is in projects of a more monolithic nature like Linux and, to a lesser extent, KDE.

Re:Gnome does it again.

[bcrowell]

However, it is a moot point now: with GNOME3 removing minimize and maximize buttons and more-or-less forcing people into using workspaces I will have to seek a replacement. I just happen to like minimize and maximize buttons and my current workflow, I don't want to have to learn a new one just for the sake of it being new.

Yeah, it is annoying how they keep taking functionality out when there's no rational justification in terms of usability. My 11-year-old daughter was running Gnome on Ubuntu Lucid, and she had her login screen all customized so it looked cool according to her 11-year-old criteria. Then I upgraded her to Maverick, and her customization went away. I spent some time trying to help her get it working again, and basically learned that it's impossible (or would require more wizardry than I possess). Try explaining to an 11-year-old why an upgrade has removed a feature that she was using and wanted. Is there some usability improvement due to removing this capability for customization? Of course not.

But that's the beauty of open source. We have fluxbox, xfce, KDE, ...

Gnome and the gulag mentality

[Anne Honime]

The biggest trouble I have with gnome is the designers constant push to force me using my computer their way 'for my own good'. No, sorry, I won't, thank you very much. I've used almost every other GUI / desktop manager around, none has tried so constantly to take away my freedom to organize the way I want to work. To add insult to injury, Gnome color schemes and icon design always seem to lag 10 years behind current fashion. Gnome reminds me of my childhood in the cold war era. Although I was born in western europe, it feels like soviets are rolling their tank divisions through my computer. You wait months in a line waiting for next release, just to hear : 'there's no more resize button, get away, and if you're not happy, praise tell me, comrade, why would you need one ? Didn't you know resize buttons are antisocial ?'. And you end up living in a concrete shack, decorated by shades of gray, praising the vision of the komintern. Just say no.

Re:Gnome and the gulag mentality

[Anne Honime]

Case in point, why on earth then, did they dropped the minimize / maximize button ? Oh, yeah... sorry to speak about consistency, everybody should know by now that consistency isn't part of Gnome's roadmap. I'm not arguing that things should be kept around forever by default, neither do I think things must change for the sake of changing. What I want is the freedom to decide what I want to keep, and what i want to drop, and the way it's displayed. Easy enough. KDE, for all its shortcomings, does that very well. You have personalities, pagers, as much buttons as you like in title bars, wherever you like, arranged as you want them to be ordered, and you can suit your keyboard shortcuts to your needs. Stuff that, Gnome.