GUADEC 2012: Discussions!

“If I had asked people what they wanted, they would have said ‘a faster horse’” — Henry Ford

GUADEC 2012 logo

Continuing to write about this year’s GUADEC conference that ended a few days ago. It should be obvious that opinions in this blogpost are my own.

The state of GNOME

Many people discussed the concerns expressed by Benjamin in a recent blog post. I know that Benjamin deeply cares about GNOME but I am not sure with some of the conclusions – some feel unproven. Dropping some thoughts here to comment some of the statements.

  • For the (non-)variety of people working on GTK+ anybody can get his/her own impression by taking a look at the GTK+ commit log (though that is just the master branch). Same for any other project. If you want to get numbers, git clone a module repository and run git log –after=2010-08-01 –author=” –pretty=format:”%ae” | sort -u | wc -l to see how many different authors got code changes committed in the last two years. We can all still differ in interpreting those numbers of course. ;)
  • With regard to the statement that “[m]ost important desktop applications have not made the switch to GNOME 3″, I am not sure which extend “GNOME 3″ is meant to imply.
    Picking one of the core parts by taking a look at GTK+3 acceptance, you can follow the ongoing efforts of Mozilla porting to GTK+3 in this bug report, see on page 19 of Michael’s slides that there is “an improving prototype” of gtk+3 for LibreOffice, enable Inkscape’s experimental GTK+3 build when compiling, or try out GIMP’s branch for the GTK+3 port.
    Work is clearly in progress and I traditionally don’t consider big complex projects as early adapters.
  • Not sure how to interpret “losing mindshare” either, so I can only state that GNOME received about 41000 changes (only counting the master branches in the GNOME Git repository, work can also happen in other branches plus some teams also use external infrastructure) by approximately 1275 people for version 3.4, and about 38500 changes by approximately 1270 people for 3.2. I know this does not tell anything in the long term. Unfortunately I do not have numbers from before 3.0 handy (I gathered the aforementioned data for writing the last release notes), but iterating over all GNOME Git repositories (list available here) and using git log –after=yyyy-mm-dd –before=yyyy-mm-dd –pretty=oneline | wc -l syntax should be doable for anybody interested. Note though that the list linked above does not include modules that were archived in Git (see the Git web interface) so this would require a bit more work to also include them.
    Apart from sheer numbers, GNOME provides and takes part in several outreach programs (Google Summer of Code, GNOME Women Outreach Program, Google Code-In) with a high number of contributors staying in our community after programs have ended (I don’t remember exact numbers but Marina mentioned this in her talk about the Women Outreach Program). You don’t need to wait for a program though to get involved – GnomeLove and its large number of mentors let your start your journey at any time!
  • Dominance of one company: Page 16 of Dave Neary’s GNOME Census from two years ago lists 16.3% commits by Red Hat employees. Judging is up to each individual, but it would require updated data to really judge the situation in 2012.
    Still, likely everybody agrees that GNOME would benefit from more companies involved. Every project out there probably would. We have seen companies cutting involvement in the past (IBM, Sun, Nokia) so that is nothing new. Companies need compelling reasons why to invest in the GNOME platform, the community, and GNOME’s future. If current reasons and future plans are not well-defined and convincing this is something to discuss and improve, on several levels (mostly advisory board, foundation board, release-team, but also all the other hard-working teams that make GNOME the awesome project that it is).
    Picking up two examples: Do we advertise enough our awesome translator community with its high quality translations for dozens of languages? Is our developer story convincing enough? Surely there is always room for improvement: The translation community plans to improve outreach and make it easier to start translating by helping with setting priorities. New tutorials for developers are in the making for 3.6 (and any developer is welcome to help by providing short code snippets). And when I take a look at the new features coming in the next release it clearly feels like the most active development cycle in the GNOME 3 era so far (Allan named some already in his post).
  • In order to respond to expectations expressed by some community members towards the release team (mostly in regard to leadership in case of potential conflicts or confusion about direction), the GNOME release team asked the community: “Which role do you expect the release team to have?” (slides are available). Frédéric’s blogpost covered this topic already.
    As written in the slides, the release team serves our community but it’s up to our community to decide to which extent. Its current self-understanding is to “try to not get into the way“. The release team did not express its position in order to initiate an open discussion.
    If the community thinks that there are ways in which the release team can help the GNOME project to perform even better, then the release team will be happy to do so. The direct feedback at the conference seemed to be very positive and expressed lots of trust in the release team’s work, but more feedback (especially by those not attending GUADEC) is needed before proposing potential changes.
    (Disclaimer: I am a member of the GNOME release team.)
  • With regard to criticism which sometimes comes up on the transparency of decisions: It is a fact that many discussions happen in real time on IRC (or via other channels, like Google hangouts), in the timezone of the developers, and not on mailing lists only.
    IRC makes it harder for interested people to follow those discussions if you live on the other side of the world or are not online all of the time. My very personal opinion is that IRC logging might help to be able to get a better understanding of the reasons why and how some decisions were taken.
    The fast pace of GNOME’s development is impressive, combined with summarizing and communicating plans early through further communication channels (mailing lists, blogs) so people can also provide feedback if they cannot follow GNOME development that closely. In this field, Allan does an awesome job with his regular blogposts on what the design team is up to so people can chime in to contribute and get involved.

Other random short bits

  • Proposed with some friends Brno (CZ) to host GUADEC 2013. Strasbourg (FR) is the other option (and also a great city!). Final decision to be made by the GNOME Foundation Board in September or so.
  • I was part of the papers committee, deciding which talks to have (we did not refuse many as the offer was quite convincing), and trying to schedule them. Though we clearly had absolutely no influence on it, I’ve been told by a few people that this year’s talks were of high quality, so congratulations and thanks to all speakers for interesting topics and good presentations!
  • 17% of GUADEC attendees were female! I hope that number will continue to increase.
  • Lots of very passionate Women Outreach Program (GWOP) and Summer of Code (GSoC) students attended the conference, with an awesome diversity. It seems that our community was perceived as very welcoming. I hope that many people will stay involved and help GNOME to evolve by participating in discussions on GNOME’s future and direction. Folks, you are the future of GNOME!

Thanks to everybody who traveled to A Coruña in order to participate, and to all our sponsors, making this the best GUADEC ever!

Attendees group picture

12 Responses to “GUADEC 2012: Discussions!”

  1. Thanks for the writeup. A few comments:

    I think the term 2 years is a bit long. I like to go release-by-release. So I usually don’t use –after=$DATE but 3.2.0..3.4.0 to restrict the comments. (This of course requires adapting to the tags used by the respective projects, so is a bit more work…)

    The total amount of authors is instructive, but I was worried about core developers. And the authors list includes all translators or submitters of one-off patches. I tend to use
    git log 3.2.0..3.4.0 –pretty=%an | sort | uniq -c | sort -n
    to get an idea of projects.

    Interesting question btw if %ae or %an is better at guesstimating uniqueness of persons.

    Isn’t http://git.gnome.org/repositories.txt the list of all repositories, even the toy ones we hack on? So I’d guess those are a bit overkill?

    Dave’s census was a summary over the lifetime of GNOME, so its numbers are a summary of 10+ years of development, and it ended with 2.30 or so, no GNOME 3 in there. Since then, a lot has changed (shell vs panel for example). But even in that report, there’s a graph about who owns what (page 19). And the core infrastructure is pretty much already Red Hat only in there.

    Last nitpick: All the big projects you list have an ongoing port towards GNOME 3, but none of them had their port near a releasable state last I checked. Also, none of the projects considered the port a priority at that time, and I haven’t heard anything else so far.

  2. Mardy says:

    IMHO even if the consensus on a decision is reached in IRC, it should still be posted and motivated in a mailing list.
    At least this should hold for all UI decisions and those technical decisions which might impact other projects.

  3. not important says:

    Amazing that you can say “our community was perceived as very welcoming”… as a former contributor I’d say GNOME is one of the worst project in terms of friendliness, transparency and overall atmosphere that I have came in contact with. I won’t comment on the other assertions as it seems this post is very biased. Benjamin’s post comments also speak a lot about what users are feeling. But still, let’s continue to say that we are great people making great software.

  4. aklapper says:

    @not important: Sorry to hear that you had bad experiences, but I guess that I can say that it felt welcoming, simply because it is my impression (and as I didn’t run a survey among people who went to GUADEC for the first time). It is unfortunately likely that “feeling welcome” does not apply for 100% of all cases for many reasons, be it timing or different attitudes or expectations. But it’s also amazing that you know “what users are feeling” – I assume that you didn’t want to imply that 100% of the users commented on Benjamin’s post either. ;) I consider it normal that disagreement is expressed more vocal and louder than agreement.

  5. Tobias says:

    The results of the design process should be expounded on the mailing lists, and only there, because the people who blogged in the past days, Allan Day and Jon McCann filter out anything that sounds critical and only let the Jubelpersers comment on their blogs.

    This is what Otte meant by »self-congratulatory echo chamber«.

    On Allan Day’s blog I commented that he sounds like the Baghdad Bob of Gnome. That was censored. On McCann’s blog I commented that his blog post contains contradictions that are frustrating in the light of what is happening really. That was censored as well.

    And on Bugzilla commenting on general phenomena is not allowed. I guess you know that.

    Hence there is no place to contact the responsible parties themselves and /even/ challenge them, as outrageous as that might seem.

    So, how is Gnome welcoming again? How is Gnome, in the words of Jon, hearing the complaints loud and clear on mailing lists? It’s bizarre when you see that in context.

    The thing is, unless you are a part of the inner circle, you are excluded from the Gnome castle and can only shout across the nice little moat.

  6. Tobias says:

    As a follow-up, I just found an online comment that is relevant to you, André:

    http://lwn.net/Articles/509555/

    But I must say, I was genuinely grateful that you let it “slide” then.

  7. Michael Hill says:

    André, thanks for the terrific report. My experiences have been nothing but welcoming. I find all that’s required is to show up and I’m invited out to eat with a bunch of friendly people. It was true in Toronto last year when I met the Doc Team (“let’s go for sushi!”), in Berlin as part of the Desktop Summit volunteer crew, and especially this year in Prague and Brno.

  8. aklapper says:

    @Tobias: With regard to the LWN article: I fully agree with what bkor says. Sure it’s not the best style to write a comment that only refers to the expression instead of the content/criticism itself. However as I’m not a developer but a bugmaster, I won’t and can’t comment on decision making itself as I often don’t know enough about it.

  9. aklapper says:

    @Tobias:
    I agree with you on using mailing lists to publish results. I don’t agree with your impression of filtering everything bad out. Plus would be interesting to know what blocks you from trying to become “part of the inner circle”, as you call it.

    “Censorship”: If I run a website then it’s me who decides what will be published on it. Same for my blog and comments on it. My blog does not say “All comments will always be published”. :)
    I don’t know anything about the specific case that you mention, or language, expression and content involved so I don’t feel like being in a position to comment on it. It’s up to each blog/website admin what s/he does or not on her/his site.
    Still, “censorship” is something totally different. I won’t (and can’t) stop you from posting your opinion anywhere, except for if it was on my site where I’d decide what is published and what isn’t.

    “Bugzilla”: Yes, because Bugzilla is not for general discussions but for specific, well-defined bugs and enhancement requests. Go to mailing lists if you want a general discussions. That is and will be my point of view.

  10. Tobias says:

    The mailing lists should not only serve as a publication outlet, they should also be a place for peer review. All the constructive input coming in in that particular bug amounts to “Yes interesting, but …”. And those are pretty noteworthy “Buts”.

    However, they are at a point of no return, if you take what Cosimo is saying. This is too late. Essentially they are handwaving that it will all be okay (“Mit dem Angriff Steiners wird das alles in Ordnung kommen”, if you know what I mean …).

    And unless they go ahead and finish implementing their changes there’s no telling if it works also for the endangered use case in the end. But if they push through even further there will be no going back, it will be set in stone. So, this is not peer review at all.

    And as for the blog posting. I have a fairly strong sense that my complaining about the particular developer staying in hiding essentially I prodded him into responding at all and even prodded him to lay out this long-overdue explanations in the blog post he wrote. I read it and found it well formulated but with flawed reasoning in some regards; and it was again handwavy. So I commented, but the comment didn’t appear.

    On a mailing list, before the fact, this would not be possible, because it is an open, if lightly moderated, forum. If someone manipulates a public asset, which nautilus is, he should be asked to defend and expound his actions before he performs them.

    But they shy away from that because they sense that what they want to do is not what the people want them to do and there would be significant push-back. And so they labor away in the least visible manner possible. Lichtscheue Gesellen essentially.

  11. aklapper says:

    @Tobias: I don’t think that anybody is hiding from conversations on purpose, however if you want to get work done you don’t spend the entire day discussing with every single person either, as everybody nowadays can easily have an opinion and share it (bikeshedding, anyone?).
    This is especially true if a non-small amount of commenters clearly misses basic manners (I took a quick look at the LWN.net link that you posted here and I was reminded why I normally keep out of public forums: some people on the internet behave in a way they’d likely never ever dare if it was a face to face conversation).
    Or to put it into Winston Churchill’s words: “You will never reach your destination if you stop and throw stones at every dog that barks”. That means that you should neither talk to everybody nor to nobody of course. :)

    PS: Not sure if dropping German language bits helps international readers to get your points.

  12. Tobias says:

    I see the conflict and the difficulty in seperating the wheat from the chaff, or the bikeshedders from the genuine valid concerns. So I think mailing lists would be a good forum because unlike comment threads you actually have to care enough to sign up and go through some hassle which filters out the drive-by commenting to some degree.

    In science you cannot publish before peer review, and in normal software development models there are public reviews as well. On the Linux kernel all patches go through the mailing lists. There’s some abrasive but /necessary/ discussion there, but after hearing the debate the maintainer is free to arbitrate and has the final say.

    This is what I wanted the maintainer in question to do, to hear the other side and decide. But it appeared that in spite of himself being skeptical initially he was somehow convinced (in the background, and he’s not impartial because he works for the same employer, unlike Linus Torvalds who is independent).

    About the movie quote. I find it particularly poignant because Gnomers actually used this film to make fun of outsiders during the previous Guadec, when in actuality it is much more fitting for the attitudes displayed in the “cabal” of the inner circle: “I’m sitting in my bunker and I wave my hands at bad news and opposing views”.

    Anyway, what is I think significant is that the fervent, uncivilised, mudflinging on the third-party forums is to a large degree the consequence and effect of the practices in modern-day Gnome. People reas on some shitty ad-infested Ubuntu blog: “They are killing features again, no discussions, and they won’t listen!!!!!!111″

    If, for instance, that McCann blog had come earlier, that would have ameliorated the shitstorm. Even better to discuss it beforehand, even if there’s opposition. That should be expected. And on the kernel this works out fine.

    Now they have an ivory tower of development and complain about rejection. I cannot feel sorry for them.