Staring Into The Abyss: Some Thoughts

Last week Benjamin Otte shared some thoughts about GNOME that were pretty stark. It gathered some steam and hit Slashdot and this all happened the week GUADEC was taking place in A Coruña. I wasn’t at GUADEC :-( but I can imagine there was some fervent discussion about the blog entry.

The gist of Benjamin’s blog was that people are leaving GNOME, that the project is understaffed, and arguably the reason for this is that GNOME has lost its direction and Red Hat have overtaken the project as the primary contributor-base. Of course I am summarizing, but check out the original post if you feel I am not representing Benjamin’s views fairly.

I wanted to share a few thoughts. To be clear: these thoughts are my own, and I am not speaking on behalf of Canonical, but I am speaking from my experiences as someone who has primarily been affiliated with Ubuntu and as a Canonical employee. My feedback is going to be frank but I really do care about GNOME as a project, and this feedback is intended from a position of love for the project and to be open and transparent about my own experiences as just one set of eyeballs in this story.

Actual eyeballs.

Fortunately, I think all of these problems are solvable, but for them to be solved GNOME is going to need to do a little soul searching to discover and focus on the right problems and explore and deliver the right solutions.

A Little History

To provide a little context, my interest in GNOME pre-dates my involvement in Ubuntu. I have worked on a few applications that use the GNOME platform (Jokosher, Acire, Lernid, and most recently Ubuntu Accomplishments) and I have had a long interest in where the project is moving forward and as a core part of Ubuntu. I used to go to GUADEC every year, and I consider many folks in the GNOME project to be good friends.

While I care about where the project moves forward I too have also become concerned about the direction it is going in, not in terms of the design and user experience of GNOME (there are other, better versed people to assess this work), but instead in terms of how the project works with others such as companies, developers, and other partners.

In my mind GNOME has become bittersweet. I remember back at GUADEC in Stuttgart in 2005, discussions started happening about what form GNOME 3 would be in. As the years progressed the project struggled to decide on a final vision for what GNOME 3 would look like. This is not surprising: GNOME 2 was such a smashing success that GNOME 3 was going to be difficult second album time. Ideas were shared and bike-shedding occurred, but ultimately it seemed that the project was lacking leadership to take take all of these ideas and flip the switch to a vision and design and move it forward.

Around this time Ubuntu had become arguably the most popular way in which people were consuming GNOME and we (Canonical) were hiring more and more people to perform this integration work (which is no light task, as any distro developer will tell you).

If all else fails, bribing people with bubble-wrap grows popularity.

Back then Canonical was taking quite a bit of heat for “never writing code and just shipping other people’s work” (which I always found a misguided viewpoint as integrating and delivering a solid Free Software Operating System is significant work and a great contribution to the wider Free Software commons).

We were starting to find though that there were areas of GNOME 2 that we felt could be improved and expanded (largely based on feedback from our users). We started growing a design competence and hiring developers to build new code to add improvements to the experience. Many technologies were created such as the messaging menu, notify-osd, dbusmenu and the global menu, control center improvements, and ultimately Unity as an additional shell for GNOME.

I remember this time vividly. I was in weekly discussions with Mark Shuttleworth, Rick Spencer (Ubuntu desktop team leader), Ivanka Majic (head of design), and David Barth (head of engineering these components). Our goal was simple: be able to showcase these technologies in Ubuntu and bring value to Ubuntu users, but to also ensure they were contributed to the wider GNOME project as technology that could help the general project in moving forward.

I personally saw this all boil down into pretty simple parts: Canonical and GNOME were partners and it was a mutually beneficial relationship – the GNOME desktop with barely any users defeats it’s purpose and Canonical was helping to deliver it to millions of users in Ubuntu, but Canonical could not build an awesome Ubuntu without the wonderful components in the GNOME desktop to fill in the many different pieces in an OS.

My simple philosophy was also marinaded in the gift culture of Open Source and Free Software: Canonical was paying designers and developers to produce new code that could be of value (and thus offered as a gift to the GNOME commons) and as with all gifts, while it may not be exactly what you want (and may need some adjustments and improvements), I presumed there would be a polite, respectful, and open discourse to take these contributions and bring them into the shared commons that was GNOME, particularly as they were created with GNOME in mind.

This was not my experience of what happened.

Partners

I was really disappointed with what resulted. After years of Canonical and Ubuntu being criticized for not contributing code, when we then engaged in writing code we were met with a frosty, suspicious, and at times, frankly entitled attitude from some parts of the GNOME camp.

Now, don’t get me wrong, Canonical was not perfect here either. I fully admit that some of this relationship could have been handled better (and I am partially to blame here). We made some mistakes early on in which code was released too late and there was sometimes not enough open discussion. Retrospectively, we could definitely have done better in being more pro-active in some parts of the relationship too. At the time we were still learning how to do this, and as such we made some mistakes too.

Canonical wanted to strike the right balance of bringing innovation to Ubuntu releases with new features, but to also openly engage and contribute that innovation to upstreams such as GNOME. My goal here is not to open up a blame game of who did what and when (I will leave that to the commentators ;-) ), but what disappointed me most about the whole situation was that from my personal perspective it seemed that some influential members of the GNOME project were treating Canonical’s contributions more critically and suspiciously than others.

Now I haver never subscribed to conspiracy theories, and I don’t believe that there was a shadowy GNOME Illuminati that was meeting together in a hollowed out volcano to plan how to keep Canonical and their contributions out of GNOME, but I was surprised and disappointed at the attitude that came out of parts of the GNOME project to us, when we were ultimately delivering GNOME to millions of users as well as writing new code that could enhance GNOME. It just seemed incredibly entitled.

The shadowy GNOME Illuminati

There were three things that really blew my mind about all of this:

  1. From my experience of working on volunteer Open Source projects, new volunteers and their code contributions are tremendously valuable. As an example, if someone comes to my current project (Ubuntu Accomplishments) and is willing to propose new, disruptive ideas, and willing to contribute chunks of code, I will treat those people with open arms. Being challenged is a good thing: it keeps us fresh, and a challenging, innovative idea followed up with running code is awesome. Now, of course, this is not to say that writing code automatically gets the contribution into the core project, but I would treat the entire social engagement with someone offering such a gift with positive open discussion to see how we could find a great solution that makes everyone happy. This seems an area where things could be improved with GNOME.
  2. If I was also running a project that was understaffed and struggling to define its direction (which I would argue was the case with GNOME at the time) I would treat such new contributions as wonderful ways of solving problems and building a new direction for the project, particularly if our major distributor was going to be delivering that technology anyway. Code is the currency of Open Source, and rejecting chunks of this currency because they don’t fit an as-yet incomplete jigsaw puzzle of a vision just doesn’t make sense.
  3. Without sounding egotistical from the perspective of an Ubuntu guy, I would argue that the vast majority of GNOME consumers were getting GNOME in Ubuntu. Of course, there was and continues to be the wonderful work going into Debian, Fedora, OpenSuSE and others, but it seemed that Ubuntu was the most commonly-used GNOME distribution (I suspect it still is). Again, I saw this as a partnership but from my perspective it seemed like parts of the GNOME project saw Ubuntu as fundamentally subservient to GNOME; as if we had an obligation to deliver whatever the GNOME project saw fit, irrespective of our own ideas and feedback from our users. In my position as an Ubuntu guy, I have always tried to treat our upstreams with maximum respect as they are a big part of who we are; Ubuntu is nothing without awesome apps, and a wonderful integrated experience. I guess I just expected a more positive and collaborative experience with GNOME than I experienced…the kind of collaborative experience that I had known and loved in the earlier days of GNOME.

Of course, it takes two to tango and we at Canonical could have no doubt done better to improve our relationship with GNOME, but I remember back then feeling like no matter what we tried to do, we came up against resistance from the GNOME project, and this was de-motivating and no-doubt added stress to our relationship.

GNOME 3

To shift gears a little, one of the points in Benjamin’s post was that GNOME 3 is a Red Hat project. To me this is a bit of a double-edged sword.

On one hand, the crux of his point is entirely valid: most people contributing to GNOME seem to be a clique of Red Hat folks. What concerns me a little are the concerns in parts of the community that Red Hat is “running the show” and that much of the decision-making has been private to Red Hat staff.

Here’s the thing: I don’t doubt that this is probably happening, but this is not necessarily a bad thing. These concerns again highlight what I think continues to be an unrealistic expectation in parts of the GNOME project with those who are willing to invest in the platform (in this case, Red Hat). If Red Hat have decided to invest in a team of developers to work on and bring value to GNOME, building Free Software that can be shared with everyone, these contributions should be received with open arms. Leadership is leadership, irrespective of the employer.

Of course, there needs to be a culture of openness and transparency, and I suspect a certain amount of internal water-cooler chat is happening in Red Hat, but you will find that with any commercial team that is actively engaged in a Free Software project; we just need to always try to keep things as open as possible. GNOME is definitely going to need to ensure that the openness and values of open collaboration are not compromised, and an open and frank discussion with the Red Hat team about resolving these concerns is no doubt the best step forward.

Pictured: a proven conflict resolution technique.

I personally think it is wonderful that Red Hat are investing so much in GNOME and they have arguably led in much of the direction and leadership in delivering GNOME Shell and the various other parts of the platform. What seems ironic to me is that the same criticisms that were thrown at Canonical with Unity (as a perceived competitor to GNOME Shell, which it was never intended to be) are now being leveled at GNOME Shell (“you don’t care about our needs”, “you are pushing your own agenda” etc).

Maybe a solution to this problem is to be open and frank about the relationship with Red Hat. As an example, we always try to be open about our relationship between Ubuntu and Canonical; there is no doubt that Canonical drives a lot of the development and innovation in Ubuntu, although this leadership and innovation is firmly rooted in expectations around openness and collaboration. We don’t try to hide the influence Canonical has on Ubuntu, and I wonder whether the wider GNOME community feels comfortable in accepting the influence Red Hat has on the project. This is always a delicate balance.

I would agree with Benjamin that GNOME is essentially a Red Hat project these days, but as I say this is double-edged: the wonderful benefits of the investment from Red Hat will be tinged with the challenges of how vendor-neutral the project wants to remain.

The Future

So what is the future of GNOME and how can these problems be solved? Can they even be solved in the first place?

I think so.

I love GNOME as a project, and I love the folks involved in it. While we don’t always agree, the core ethos and goal of GNOME is admirable: to bring an awesome Free Software desktop to everyone. While I personally prefer Unity as a shell, I think the work that has gone into GNOME Shell has been a wonderful rebirth of the motivation and focus of GNOME. The architects of this vision should be credited in getting GNOME out of the slump I mentioned earlier that seemed to stem from 2005. Of course, I will always be disappointed that GNOME seemed quite so resistant to much of the contributions we wished to make, and I think we could have helped to have moved things along a little faster, but I am delighted that GNOME 3 has got to the point it has got to.

As I mentioned earlier, my feedback here really has nothing to do with the design and technical direction of GNOME, and others can provide more insightful commentary than me. I do though think this people-problem issue of GNOME being a rather difficult project to work and interface with at times is a problem that has not yet been confronted and resolved. While this problem continues to exist, I worry that it will eat away at GNOME more and more.

GNOME is blessed with some wonderful leaders, and I hope that the content in this post can act as some food for thought: I am not expecting everyone to agree with me, but if this opens up a discussion about these topics I will be happy.

What is not a solution is for us to give up on GNOME. I know some folks are moving on from the project and moving onto other things, and we have more competition than ever for desktops, but I still see GNOME as an important foundational component of the Free Software and Open Source desktop today.

Now, I am sure this blog entry is going to result in some folks screaming from the rafters that I am misrepresenting GNOME and it is all Canonical’s fault, and you are entitled to your view. Traditionally I have not wanted to raise these concerns publicly as I didn’t want to cause any further harm in the relationship between Ubuntu and GNOME, but Benjamin’s blog post seemed to offer a good opportunity to throw out some feedback that might be helpful in constructing a solution.

While I don’t have much time to contribute to GNOME formally these days, I am more than happy to talk more, provide any further feedback, and help where else I can. I would love to see the GNOME project that we know and love be back in a healthier state. Thanks for reading.

  • 92ertele

    Please, do not start something like Apple did 20 years or so ago. Now they have lot of bucks but also lot of bugs – nobody cares but it even more cares that it`s nothing more than a buck-shitting-machine

  • Olav Vitters

    Speaking as a release team dude, a lot of the improvements from Canonical come out as total surprises. The same thing you say about Red Hat I thought applied to Canonical. Meaning: no clue what is going on, then a surprise just before a 2.x.0 release.

    And some things just can be done way quicker as a hack than as something proper.

  • Anonymous

    Thanks for the thoughts, Olav. I definitely agree that we could have timed some of these contributions better. I would like to think that as time went on this got better though.

  • Jo-Erlend Schinstad

    What we have here is failure to communicate.

    I think we need to start designing the language we use. For instance, when people claim to “hate Gnome 3″, are they then talking about the shell or the platform? I’ve participated in so many of those discussions and it’s often very difficult to even understand what’s being discussed. We know that a lot of people were confused about the changes in Gnome Panel 3, that you need to press and hold the Alt-key, etc. Those are very small changes, so that cannot really be what “Gnome 3 haters” are talking about, except most people weren’t told that. The panels suddenly looked quite different and there’s no obvious way to customize them anymore (Alt+Right-click instead of right-click. Super+Alt+Right-click if you’re using Compiz). But then you see Gnome.org uses Gnome 3 as a synonym for Gnome Shell.

    When did we start to use the term “shell” for these panels anyway? Do they do that in KDE and other communities as well, or is it a pure Gnome thing? Because if so, then that’s destructive too. I don’t mind the terminology, of course. Quite the contrary. But I think we need to start communicating communication itself. We need a community of communicators who are able to explain these things and bridge the gaps.

    I feel certain that if the changes in Gnome Panel had been explained well in advance, much of the controversy would never have occurred in the first place. Many of the most angry people I’ve argued with this past year, really didn’t know that they were quite literally a single key-press away from accomplishing what they wanted. A lot of people actually thought the ability to customize had been taken away from them for no good reason at all. Linus Torvalds recommended that people who wanted classic panels should switch to Xfce. He should’ve told people to press and hold the Alt-key!

    That’s one misconception that caused anger. To make things worse, these people were actively being confused about “Gnome 3″ as a name. They went on to attack Gnome 3, when all they wanted was really to move the clock back to the corner and add some panel applets. I think a lot of people got the impression that Gnome Panel 3 was just Gnome Shell customized to look similar to the classic panels, when in fact, it’s the opposite that is true.

    I still find it extremely difficult to communicate anything related to Gnome 3, and I try to avoid that term as far as possible. Because I know what it means to me, but I have no idea what it means to you. I’ve used Gnome since 1999/2000 or something, and I can’t remember ever having experienced so much confusion. I hope I never have to again. It has been very expensive.

  • Efrain Valles

    Jono, no matter how neutral one tries to be in these kind of posts, biases will always be in the readers mind. Great post. I would add that GNOME is going through a state of change and the people that do leave will leave and the new people along with the ones that really love the proyect will stay to see it shine as one of the diamond pillars of the free software ecosystem. The only ones needed are the new contributors and the ones that really love the project. This also happened in Ubuntu during the Unity switch

  • http://www.facebook.com/profile.php?id=100002487259611 Ethan Anderson

    With the Ubuntu Software Center and STEAM, I think we’ll see a lot more linux users become used to regularly paying for software, and users from other platforms bringing their old usage patterns with them– as the flow of money through Canonical ramps up, Canonical will have greater financial control over the entire open source software ecosystem. Are there any general principles or plans in place to handle this well or use it to drive Free Software forward?

  • Anonymous

     Our goal has always been to deliver a strong Free Software platform and to harness that platform to encourage wider Free Software development and adoption, but to not restrict proprietary software running on the platform too. So in essence, we will never ship proprietary games on the CD, but you can get them in the Ubuntu Software Center with an equal level of visibility as Free Software games.

  • Cody Russell

    Hey Jono, nice post!

    I have a little different perspective on some of this.  I’m a former GNOME contributor, a former Ubuntu contributor, and a former Canonical employee.  Hopefully not all of those will remain “former”, I’d like to get back to contributing to GNOME and Ubuntu someday. :) Anyway, I think it’s a little unfair to say “we offered cool shit to GNOME and they turned us down”.  We (Canonical, and I’ll include myself in that “we” since I was working there when a lot of this started) usually developed these features mostly in secrecy.  We didn’t allow for any design-time input from the GNOME community.  We didn’t allow for any technical coordination with them, even when we were fully aware that we were developing something that was related to something that GNOME developers were working on.  Then we would suddenly announce, “hey look at what we’ve been working on!” and the general reaction from the GNOME community seemed to be one of annoyance because they were working on things publicly, and we knew what they were working on, but we were working on things behind a wall.  We were aware that there was a duplication of effort going on, but they weren’t aware. / Cody

  • Anonymous

     Thanks for the comment, Efrain.

    I agree, and as I mentioned, my aim here is not to throw blame out, but just to share one perspective. I agree that the change in GNOME is very similar if not the same as the one that has happened with Unity: change is emotional, but if the right steps are taken it can invigorate and enhance a community and project.

  • Anonymous

    I think you raise some awesome points, Jo-Erland. As I mentioned in my post, I don’t really have any strong views about the design and technical change in GNOME from 2 to 3. My take it that I prefer Unity, but I totally accept that some folks will prefer GNOME Shell or GNOME 2. Taste and preferences is individual.

    I think you make a valid point about the shell vs. platform. Many folks still ask me why we don’t ship GNOME in Ubuntu and I have to explain that we do shop GNOME, and it is fantastic, we chose to use a different shell (Unity).

  • Anonymous

    Hey Cody!

    I agree and slightly disagree depending on the project. As an example, for notify-osd I think you are entirely correct. Later projects such as indicators were released much earlier in the development cycle. Where I do I agree with you is that the level of design integration with upstream for many of these projects was minimal.

  • Stephan Adig

    Cody,

    isn’t that the challenge for companies like Canonical and Red Hat? Working on things to make revenue later, and being so much transparent, that nobody can take away the idea?

    On the other side, it’s good to have different choices. 

    Anyhow, I think the whole issue comes down to: Why is Gnome pushing back, and projects like KDE very welcoming to new contributors?

  • https://launchpad.net/~j-johan-edwards J Johan Edwards

    This post strikes me as a little dishonest. It’s true that Jon McCann and crew designed Gnome Shell in an ivory tower and refused to let Canonical in: browse the 2008-9 archives of gnome-shell-list, and you’re find that nearly every design question is answered with “the team discussed decided this on IRC”. No logs linked.

    However, this is also true of Unity. Mark Shuttleworth unveiled “Ubuntu Light” at UDS-L, already fully designed. The same is true of Notify-OSD, Indicators, HUD, the spread, etc. Meanwhile, ayatana-design — the supposed open forum — goes almost wholly ignored in the design process.

    That isn’t to say ivory towers are bad. Design by committee, cooperation, and compromise has a long and storied history of failure. Both Canonical and Gnome acted understandably, and arguably wisely. However, I think it’s unfair to criticise Gnome for refusing foreign design contributions when Canonical does the same.

    Unity is succeeding, in my mind, primarily because Canonical can afford to out-engineer Gnome. Red Hat seems to view their contributions to Gnome as volunteering weekends at the soup kitchen: good to do, profitless, and peripheral to the rest of their livelihood. In other words, Canonical survived the 11.04 transition because they could pay their way through the perilous time when the community recoiled. Gnome could not.

    So Gnome doesn’t have the developer, testing, or design muscle to carry out their vision. Arguably. That doesn’t mean their decision to do design in-house was a bad one.

  • Cody Russell

    I’m actually not trying to be accusatory in what I posted. It’s not an attempt to blame anyone. Just a comment directed specifically at the notion that Canonical wrote some code and then GNOME didn’t seem very enthusiastic about it later.

    A lot of people at Canonical seemed kind of annoyed that GNOME wasn’t excited to accept our projects, but it seemed kind of obvious to me. I think there’s more to working in the community than simply writing code and trying to dump it on people. There’s an actual social, community aspect of it that Canonical was trying to avoid participating in at that time. In a way, it seems almost as though we were trying to become the center of the desktop community rather than just participating in what was already there.

    This seems to be the case in almost every open source community. I remember working on some code for enabling RGBA colormaps by default in GTK+ when I was working on client-side window decorations, and I wanted to create a new window manager hint to optimize rendering by marking which regions of a window were known to be opaque (even though the colormap was RGBA) so that we could avoid compositing those regions when their alpha byte was opaque. So I implement the code and post on the mailing list with this idea, but I’ve never worked on any window managers and nobody had any idea who I am. I was just some random dude with some idea to create a new WM spec that all these other window manager developers would potentially need to implement, and I was not greeted very warmly. :)

  • https://launchpad.net/~nathwill Nathan W

    great post, Jono. My heart also goes out to the GNOME team, and their work is very much appreciated. My big worry is about what seems to be a fair amount of divergence between GNOME as delivered by Ubuntu, and GNOME as envisioned by the GNOME project.

    What I would like to see is that GNOME and Ubuntu start working together more collaboratively in the design process, so that the duplication of effort can stop, and we can celebrate our shared vision (Free Software) instead of arguing over petty differences of aesthetic (who gives a damn if the clock is in the middle or on the right, anyways?).

  • Anonymous

    Agreed 100%.

    I don’t criticize a company/community for assigning designers to design something so crucial. If everyone had a say in the direction of Shell or Unity, none would be as good as they are.

    You can’t have too many cooks, otherwise dinner will be ruined.

  • Anonymous

     Great feedback, J Johan, and I largely agree.

    My goal of this post was not to throw blame at GNOME for acting in a manner that isn’t too different from how Canonical approached Unity, it was more that the way in which it was handled seemed unnecessarily hostile to me.

    I am certainly not denying that much of the Unity design and development was done in-house, and this is an area we need to continue to work on improving.

  • Guest

    problem with that is that Canonical clearly states that Unity is under its control (which we all know is not a bad thing) while GNOME in general (probably we should write g-s here, but things are spreading to other apps now too) proclaims itself “open to everyone” and “accepting other people’s feedback” while it’s not

  • Guest

    a quick note about the clock: ever tried an i-device? :)

  • Anonymous

    Gnome’s circumstance – especially when contrasted with Canonical/Ubuntu can only make sense.

    Even if Canonical ruffled some feathers over at Gnome, the product Canonical is working on is destined to be better.  This is not subjective, but if you insist that it is, then even by that standard.  Canonical is more grounded in feedback, testing, use cases and the community.If you’re going to ram a big change (Unity, Gnome Shell) through, you still need to find out what’s wrong within it by engaging your users.

    This fundamentally healthy valuation of the very people using what you make is not a mystery.  I’ve never once felt engaged by the gnome community over my years of using it – it’s no surprise then that it was only after Canonical that I could easily submit bug reports.

    Assuming my duty as the “commentator” mentioned in the post, My suspicion is that the Gnome community owns the lion’s share of the blame.  Not because decision makers were from the same project (Red Hat).  But because it became insular.  They enforced stigma against Canonical and maintained barriers between the users of Ubuntu and the Gnome desktop.  So, it’s no surprise Canonical picked up their toys and left. I could talk about how not a single person I know likes the latest Gnome desktop – by whatever name you want to call it.  But this and other similar issues are just symptoms of a bigger problem. We are at a painful but also monumental stage of the evolution of these particular FOSS projects and Canonical’s approach just happens to work better.

  • Stef Walter

    Before joining Red Hat I sort of imagined there was a lot of “internal water-cooler chat happening at Red Hat”. I was surprised by how little I’ve seen among the desktop guys. Sure it’s natural to talk with those sitting nearby in the same office, and some of the gnome-shell guys work together like that, but far far less than I imagined.

    IRC really is the ‘water cooler’ in this case. Anyway, would have been cool to see you here at GUADEC, especially in the BOFs.

  • Olav Vitters

     It just took way too long from both sides to get comfortable and understand eachother. In my view, Ubuntu is about trying stuff and see if things work. Within GNOME, it is more about always pushing from GNOME to distributions and doing things the ‘right’ way first time (which sometimes can take 2 years instead of 6 months). Such different views and assumptions created surprises and misunderstandings. It is really difficult to work together if you don’t understand what the other person is about.

    Anyway, have an idea on how to solve this. But unfortunately, it’ll take a while…

  • http://twitter.com/marcoshamas Marco Shamas

    I read this sentence a couple of days ago: “Nothing is more destructive to good design than group thinking and collective decision making. Why? As I said, to most people good design is invisible. Group decisions focus on the visible, bad aspects of design.” (Oliver Reichenstein)

     and I started to think if the open/group approach is one of the causes of bad Design on the Linux Desktop. 

  • Olav Vitters

    Some people do not see or know the difference between Unity and GNOME 3. They call Unity GNOME 3 basically. It usually takes another comment from them to notice that.

  • http://duffy.id.fedoraproject.org/ Máirín Duffy

    “As an example, if someone comes to my current project (Ubuntu Accomplishments) and is willing to propose new, disruptive ideas, and willing to contribute chunks of code, I will treat those people with open arms. “

    ‘Open arms’ is not how I would classify my discussion with you about potentially collaborating on a federated cross-FLOSS project accomplishments system, Jono. You’ve very carefully qualified your statement there in a manner that’s not really all that open or welcoming.

  • Jo-Erlend Schinstad

    You know, I’m not really talking about Gnome Shell or Unity here, but rather about the official Gnome 3 desktops; Gnome Shell and Gnome Panel. Classic and Modern. Both are still available, but people doesn’t know this, precisely because people refer to Gnome Panel as Gnome 2, even if it’s also part of Gnome 1 and Gnome 3. So that’s extremely confusing.

    The image below depicts the classic Gnome 3 desktop. You know, the one people are now convinced doesn’t exist anymore. When I show this to people, they tend to think it’s A) a heavily customized Gnome Shell, B) Not true (I’m faking it for some reason) or C) I’m confused about what Gnome 3 is, since that’s obviously not Gnome Shell.. The latter is exactly my point; it’s not Gnome Shell and it’s not Unity, but it’s still Gnome 3.

    When even screenshots isn’t enough, how can we hope to understand what we’re talking about? We can’t allow communication to just happen all by itself. We need to actively design it. We need to define what things actually do mean. It just can’t be a matter of opinion.

  • http://twitter.com/jspaleta Jef Spaleta

    Máirín,  Was the discussion in a publicly archived communication medium? I’d like to read back over any discussion for that federated cross project accomplishment system idea.

    -jef

  • http://twitter.com/jspaleta Jef Spaleta

     Archived irc logs that can be reviewed by people outside the design team  to better understand design decision making process might be part of the solution here.  Working out in the open of irc is not a bad thing. But the syncronous nature of unarchived irc  (the you had to be there for the discussion aspect of the discussions)  are still problematic.

    I can’t speak for anyone else, but for me personally, archived discussion among stakeholders are absolutely invaluable for me to understand final technical and design decisions. Looking back at the discussion that drives a final decision adds a huge amount of context that helps digest the decision and more importantly helps me understand what the stakeholders are thinking and where they are coming from. So when I need to interact with a stakeholder to address a concern I can be some what more confident I’m not retreading the tire of a previous discussion.

    It’s much harder to do that when most of the decision making process is unarchived “water cooler.” Doesn’t matter if its in a company hallway, at a conference, or on an unlog irc channel. Its the lack of reviewable archive of the discussion that is problematic for me….not my inability to inject my personal opinion into the discussion.

    -jef

  • Drycrust3

    Like you said, Gnome 2 was a big success. It was one of the reasons I got hooked into Ubuntu. To be fair, my interest is largely as a user, so the programming and inner workings don’t interest me. From that point of view, if Canonical adopted a similar layout and user friendliness to that of the Gnome 2 desktop then I would happily use it.  While Unity looks really attractive, I have found it inferior to Gnome 2, which is why I downloaded the Gnome desktop and use the Gnome-Classic desktop.  If you look at the Distribution Watch “hit list” you will notice that Linux Mint is first and Ubuntu is second, but prior to the release of Ubuntu without Gnome 2 this order was the other way around: Ubuntu was the most popular version of Linux, with Linux Mint second. There could be many reasons for this change, but one possible reason is people have found the Unity desktop isn’t up to their needs, so they look for distributions with better desktops.

  • Anonymous

     I don’t recall our conversation word for word, but I remember you asking.

    From what I remember you asked if I would be interested in making it a federated system and you didn’t have a spec, designs, or code. I interpreted your question as “would you be interested in such a federated system” and I seem to remember telling you that I was not really interested in working on that kind of problem and I wanted to get a first cut of my spec up and running first.

    The point I was making in my post was with regards to gift culture and someone who brings content (e.g. a spec, designs, or code) that can bring a challenging or new perspective. From what I remember of our conversation there was no such content involved.

  • Anonymous

     I think a federated system would be awesome too, and if someone is willing to put the time in to build it, lets have a discussion and see how  it could work. :-)

  • Anonymous

     I think you make a good point, Olav. There has always been a natural tension between distributions and upstreams, but I think one of the challenges both feel is a sense of entitlement of their position in the pie (e.g. an upstream feels they have to define the vision and direction, or a distro feels like they bring all the users).

    I would love to hear your ideas on how this could be solved; I am certainly happy to chime in with feedback and ideas if it helps.

  • Anonymous

    DistroWatch is not really a metric. Firstly, it is based on page views and many existing users of Linux don’t need to go and look at DistroWatch very often, and secondly it targets a very specific demographic — Linux enthusiasts — and Ubuntu is targeting not only Linux enthusiasts but general consumers too.

  • http://twitter.com/jspaleta Jef Spaleta

     I did not state an opinion as to whether it would be awesome or not so I’m not sure why you said “awesome too.” The language of your reply  seems to assume that I am pro-federate system when in fact I have not made a judgment either way. I’d appreciate it if you you’d refrain from reading value judgments into questions I ask. If I think something is a good idea I will relate that.  In this case I’m asking for reference material for existing discussion to read up on ahead of deciding if the idea is a good idea to support or not.

    -jef

  • Anonymous

     OK then. :-)

  • Chell

    As much as I’m interested in seeing how communications on a collaborative cross-FLOSS project went, I’m more interested in seeing constructive efforts and a willingness to move the platform forward. It is a pivotal piece of technology for many projects and companies.

    GTK, Gnome and the gnome-shell are wonderful gifts to the users and developers. Shouldn’t the focus of two of the most notable companies (Redhat and Canonical) be how to make it more sustainable – together? Even if that would mean burying he hatchet?

  • Sesivany

    “So Gnome doesn’t have the developer, testing, or design muscle to carry out their vision.” No open source desktop project has muscles to fully carry out its vision. All projects are more or less understaffed. Even Red Hat, which has multiple times more developers than Canonical, would appreciate more manpower in most of its projects. Imagine webOS had over 500 full-time developers in HP when it was killed and it was still considered a dwarf compared to Android, iOS etc. No truly open source desktop project would never ever have so much manpower. So saying that GNOME is understaffed because it’s made a bad decision is a bit pointless. And I don’t know what “in-house design” you’re talking about. GNOME designers might seem dismissive to complains and suggestions, but the design process is completely open. All mockups are on wiki, the same applies to arguments supporting mockups, critical feedback etc. I wish Unity was open like this. Of course, then someone has to come and say it’s gonna be this way because design is about compromises much which many people don’t understand. I think that what Canonical didn’t understand back then is that upstream (GNOME project in this case) is not a marketplace for solutions where you come and say “hey, I’ve made this solution. Take it or leave it”, it’s a place where solutions are born. I truly believe that everything in FLOSS has to start from every first beginning in upstream. It’s the only way to make sure that everyone is on the same page and balance between all parties is maintained. That’s why I think that approach Canonical took (inventing in downstream and pushing stuff upstream later) is wrong and doesn’t support healthy collaboration in the open source community. And no wonder it caused conflicts between GNOME and Canonical.  

  • http://twitter.com/jspaleta Jef Spaleta

     I’d love to see Canonical publish its methodology for counting usage for review and discussion.  Where should I go for an open discussion on that?

    -jef

  • Stefano Maffulli
  • http://duffy.id.fedoraproject.org/ Máirín Duffy

    Jef, it was a private conversation:

    The idea originally came from the Fedora RPG idea proposed at FUDcon Tempe in 2009: https://fedoraproject.org/wiki/Fedora_RPG_OLD My team quickly realized, while it’s a cool idea, Fedora doesn’t exist in a vacuum and all FLOSS contributions, whether or not they are Fedora-specific, are worth encouraging. We thought a federated system would be a great solution, not the ‘one size fits-all bikeshed’ Jono claimed in this conversation that OpenBadges is. Then some months later I heard about Jono’s ‘Accomplishments’ system. Three projects with extremely similar goals, all around the same timeframe seemed like a great opportunity to avoid NIH, so I thought a federated system built on the Open Badges API & spec Mozilla put together would be an elegant solution and a great way for several prominent FLOSS projects to collaborate together.

    So, yes, I most certainly did have a spec that I referenced: the Mozilla Open Badges one (https://wiki.mozilla.org/Badges/infrastructure-tech-docs)

    Jono, you specifically invited people to participate who can submit patches. Thank you for correcting that qualification with the notion that other forms of contribution are valid. That being said, it doesn’t change the fact I tried to collaborate with you because I was interested in building a badge system for platforms other than only Ubuntu (eg tied specifically to Ubuntu One), and your core objection seemed to be a system that was anything but tied to Ubuntu.

    Oh, well, a lost opportunity, happens every day. No hard feelings, it just seems prudent to be honest about how open and welcoming you are when asking similar things of GNOME.

  • Fe

     you’d love, but who said you can sir?

  • Michael Zucchi

    You cannot compare a set of system libraries to a standalone application: you can’t go accepting patches willy nilly to parts that can have far reaching implications.  Nor can you compare it to mission critical applications that handle people’s precious data.  Hostility is also a product of a reduced developer base …

    I think GNOME was a victim of it’s own success and it involves technical, commercial, and political aspects well beyond changes to the window manager, which is just the most visible recent event.

    Commercial: from very early it was driven by commercial interests and paid developers due to the .com boom.  When that busted, and over time, those paid developers were lost, and now you’re just left with redhat and canonical.

    Political: Bullies came in and forced their apple-mac centric view of the world onto every one else, alienating users and developers alike.  They gained entry because the developer base was already shrunken due to the .com bust and other competing technology.  And Miguel lured another bunch away with this mono stuff (and really, GNOME itself was similarly helped by his ability to get people to do stuff for him, until he got bored with it and left it behind).

    (GNOME itself was started as a copy of microsoft com technology, evolution was a clone of outlook, and then it moved to macos as many of the developers didn’t even use gnu/linux …)

    Technical: C was a good decision back at the start, and would remain so for portability and multi-language support reasons if the developers just focussed on some core capabilities and a limited language set.  Few developers want to write desktop applications in C, and about the same for languages like python (even smaller for the custom gnome languages).   And mono?  Even if it were technically brilliant (which it isn’t), basing a GNU technology on .exe files and MS tech is just political suicide.  I was always a super `hardcore’ C hacker, and I still get a kick out of realising an algorithm in C, but

    I wouldn’t dream of using it for a GUI – I use java now, even on desktops, absolutely loathe c++.

    The early days was filled with simply filling out the basic capabilities of the libraries – you know, like rendering text and graphics.  But these are all ‘solved’ problems now (i don’t think people remember how archaic things were pre 2000), so a large part of the early appeal for those type of developers is gone.

    I don’t really have a recipe for success – to me GNOME 2 was too bloated and I was already on XFCE back then.  I only ever used the gnome desktop (despite developing for it for years) when I was working for Novell.  GNOME was at it’s peak about 2003 and it’s just been a down-hill slide since.

  • Michael Zucchi

     ’group think’ isn’t collective decision making.  It’s group who bully their decisions onto others by silencing dissent.

    There’s bad design eveywhere, not just in gnu/linux desktops.  It’s a hard problem: you need both good technical and good design skills, but most people who put themselves in those positions have neither.

    (most of those with the required skills just aren’t interested in giving their skills away for free or even at the pay rates such a position typically earns).

  • Michael Zucchi

    I was working for the same company – but being 10 hours of timezone away meant I missed out on all of these ‘little, minor conversations’. 

    It shuts one right out of the conversation itself, and hides any details from those not present.  Even if people are irc/email hounds, even without the 10 hours of timezone difference.

    They matter a lot, even though they don’t seem to at the time.

  • Michael Zucchi

    i first heard `shell’ wrt windows 95.  Although ‘technically’ it’s a ‘shell’ in the same sense a ‘bourne shell’ is a shell, just a GUI based rather than CLI one – I find the term strange.  It is not a gnome specific term, although it wasn’t a common gnu/unix one.

    Personally I find calling it a ‘shell’ a turn-off, and prefer ‘desktop (enviromment)’, or even ‘window manager’.

    About the only thing it does above a traditional window manager is that it mounts removable media when you plug it in – and TBH i’d prefer that’s ALL it did.

  • http://benjaminkerensa.com/ Benjamin Kerensa

     Jef has a good point.

  • Jo-Erlend Schinstad

     But a WM is not the same. And a DE is much more. The problem is that Gnome used to have only one shell – the panels. Now it has two; Gnome Panel and Gnome Shell. Both are in Gnome 3. And it has two WMs as well; Mutter and Metacity. They’re also both in Gnome 3. In addition to that, there are several third-party shells, including Unity – which of course is neither a WM or a DE. And Compiz is not part of Gnome.

    The situation is very confused, and that means users will be confused as well. That’s no good.

  • Home user

    This discussion is confusing. Most of the discussion on slashdot was about design choices. People seem to complain about usability and configurability. it looks like users and distros are dropping gnome because of design choices.

    My point is that an understaffed project with closed leadership can make popular design choices whereas a fully staffed project or a project with open leadership can make unpopular design choices. (But I am not a gnome developer, so I may be wrong)

  • http://adamwill.id.fedoraproject.org/ Adam Williamson

    I think Benjamin’s concern about Red Hat isn’t undue influence by RH on GNOME, but rather – as he explicitly states in the close of that paragraph – the ‘bus factor’. No project should rely too much on any one individual or company, because if something bad happens to that individual or company, the project is in trouble: that’s the ‘bus factor’.

    The problem as I see it isn’t that RH makes significant contributions to GNOME – it’s that, increasingly, no-one else does the same. RH has long sponsored significant chunks of GNOME development, but other companies used to do it too. There seems to have been a gradual trend recently for other companies to invest less in sponsoring GNOME development, and RH has been almost by default ‘picking up the slack’ – no-one else seems willing to give GNOME devs paychecks any more, so RH has been doing it. Which is probably better than the alternative of the number of paid GNOME devs dropping even more significantly, but obviously isn’t as good as the situation where there are multiple vendors investing in GNOME.

    Unfortunately, it’s not a problem RH can really solve: all we could do is stop paying all the GNOME devs we currently pay, which doesn’t seem like something anyone would appreciate. (Least of all the devs). It’s a problem other companies have to solve, by stepping up and hiring GNOME developers, quite simply. We need someone to replace the investments that used to come from Novell/SUSE and Sun/Oracle, to boil it down.

  • http://adamwill.id.fedoraproject.org/ Adam Williamson

     ”When did we start to use the term “shell” for these panels anyway? Do they do that in KDE and other communities as well, or is it a pure Gnome thing?”

    It’s standard industry terminology. The ‘shell’ is the basic desktop interface. It’s not GNOME-specific.

    In a sense I see where you’re going, but I’m not sure the GNOME team would agree. AIUI, part of the direction of GNOME 3 is exactly supposed to be towards GNOME being more of a unitary ‘thing’ and less of a platform. The GNOME team, to a great extent, really does see GNOME 3 as being ‘the entire stack’ – if you’re not running Shell you’re not running GNOME 3. The fallback mode is a kludge they want to get rid of, not an alternate shell with full status. GNOME as a project has consistently argued that running Unity does not constitute running GNOME 3, in contrast to how Canonical looks at it.

  • http://adamwill.id.fedoraproject.org/ Adam Williamson

     Y’know, we should provide templates for ‘we’re #1 on Distrowatch! Woo!’ and ‘Well, you know Distrowatch rankings don’t really mean anything, right?’ to emerging Linux distributions, for use at the appropriate points in their life cycles. I used to have the latter on my copy/paste cheat sheet at Mandriva…=)

  • http://www.fewt.com/ Fewt

    One thing I have noticed over the last year or two – and find pretty disturbing – is the amount of change that RedHat seems to be forcing upstream.  The Systemd / UDEV merge, GNOME 3, turning Syslog into a blob, restructuring the filesystem.

    These things may be good for RedHat / Fedora however I’m growing increasingly concerned about the downstream impact.  For example, we shouldn’t be forced to adopt a new init system to continue to use UDEV.  I know that Fedora has claimed we don’t have too, but in practice I don’t see how you can possibly merge the two without the goal to be a complete take over.  If that wasn’t the goal – it wouldn’t have been merged.

    Just some random thoughts on the subject.

  • Guest

    Just fork all of GNOME.

    The GNOME 3 maintainers are fucking idiots and cannot be trusted with the trust of maintaining it.

  • Marcus

    If you really think that innovation has to be done behind closed doors, you have not understood the open source ecosystem @Stepan

    It’s never about taking away ideas, but collaboration.

  • http://twitter.com/marcoshamas Marco Shamas

    I think a good metric is Wikipedia since has a huge number of users that are not related to any particular category: http://stats.wikimedia.org/archive/squid_reports/2012-06/SquidReportOperatingSystems.htm It looks like that in the latest 2 years nothing has changed. Ubuntu is always between 0.6% and 0.7% , while Linux Mint 0.01%

  • http://twitter.com/jspaleta Jef Spaleta

    While the wikimedia stats are probably the best public stats we have to work with, there are still limits.

    By default the browsers in Mint still use a user agent string with Ubuntu in it.  I know, I checked. I booted a Mint 12 installation and checked its useragent string myself. I guess I should check Mint 13 now as well.  Wikimedia can’t tell the difference between a typical Ubuntu and a typical Mint client, making any comparison between the Ubuntu and Mint numbers suspect. Based on what I know of the default useragent strings, and my review of the scripts wikimedia uses to digest their logs.. I will say that noone really has any idea how much of that Ubuntu usage in the stats is really Mint usage.  Noone.

    I would argue that wikimedia stats are only viable to track “linux” usage with all distros combined, which isn’t showing any growth either across the span of time the wikimedia stats cover. At best, what the stats show is that all the distros are doing really is competing for the same stagnant userbase demographic. Our slogan as linux distro users should be: “We are the 2 percent”. 

    The only “linux” that has leaped across into mainstream adoption according to the wikimedia stats is Android.  

    -jef

  • http://www.facebook.com/profile.php?id=100002487259611 Ethan Anderson

    I mean more along the lines of donations to Free Software applications available in the software center– when that’s implemented, visibility in the software center, reviews, and ratings will begin to have an actual direct or indirect financial effect on free software projects, and part of that effect will be the result of how they’re able to present themselves there. That would give Canonical a degree of power over these projects, would it not? For example, excluding improperly packaged applications will deprive their developers of some donation revenue. Do you expect that to improve things at all?

  • http://twitter.com/marcoshamas Marco Shamas

    I think it will remain at 2% for a very long time. Software availability drives OS selection. For example I can’t see a manager pushing for Linux for any reason if the software isn’t there already. It’s more of a “We know AutoCAD, what do we need for that?” kind of deal, not “I like CentOS. What CAD software is available for that?”. Since most of the software for the Linux Desktop is open source, usually the best tools are also ported or already available for Windows. But the best proprietary software for Windows isn’t available for Linux. Even if now Valve will ship Steam for Linux we would have only a little part of what Window’s users and OSX users already have. The only way to attract more users would be to have good programs that can run exclusively on Linux, but I’m sure a lot of people don’t like the idea.