Re: Desktop Summit Planning

Hi Brian,

On 12/14/2011 04:42 AM, Brian Cameron wrote:
At the October 26th IRC meeting, Jon McCann spoke about his concerns
about the Desktop Summit.[2] At the November 23rd meeting, Dave Neary
(who was also very involved in making the past Desktop Summit happen),
highlighted that there are real benefits to sharing facilities and that
concerns can be fixed.[3] Dave pointed out that it would be best if a
decision could be made sooner than later, since it is hard to start
putting together bids for 2014 if the scope of the event is not clear.

It's useful to go back to what GUADEC is for, I think - and work from there to see whether a Desktop Summit supports or works against that. If we don't agree on why we have a GUADEC, we're not going to agree on the rest.

Back in 2005, the last time we had this discussion (at the time, it was over maintaining User Day, or rather integrating end-user related content and inviting people from outside the GNOME community to participate in GUADEC) the goals we came up with, which are still documented in were:

Primary Goals:

1. To have fun meeting friends
2. To allow developers and contributors to have high-bandwidth discussions.
3. To highlight new ideas and cutting edge developments.
4. To get new contributors and involve current contributors in a higher level.
5. To set the direction of the project for the coming year

Secondary Goals:

1. To create media awareness out of the usual circles
2. To involve corporate partners and facilitate an approach to the community
3. To spread free software to the surrounding region


I don't see how a Desktop Summit affects goal 1.

In my mind, GNOME's gotten to the point where those high-bandwidth, planning the future discussions (points 2 and 5) have gone beyond what we traditionally thought of as the GNOME project.

Just look at where key GNOME contributors are working now: Lennart Poettering is working on audio and system start-up with a clear focus on improving the desktop, the browser and web platform have become key components of the free desktop, Richard Hughes is working on making system-wide colour management a reality, there are kernel hackers and Xorg developers working at that level of the stack to improve the desktop end-user's experience. GNOME has long been good at fixing problems at the right point in the stack, rather than patching around infrastructure issues.

In addition, applications like LibreOffice, Eclipse, Mozilla, and people buiilding on the GNOME platform (Unity and XFCE come to mind) all build on and use our platform, and I think it would be very beneficial to get people from these projects together to see what we can do to make that platform better for them.

So it makes sense for us to have some kerrnel, Xorg, web, and application developers there.

KDE is also looking down the stack at things like metadata, audio, voip, file sharing... - and so I definitely think it makes sense to have some relevant KDE people working with the relevant GNOME people on avoiding duplication of effort where it's possible.

The other goals could potentially be compromised by a desktop summit.

The bigger the conference gets, the more people will tend to stay in smaller groups of people they know - Dunbar's research on communities in action - resulting in it being a harder conference for project newcomers or peripheral contributors to attend and figure out what's going on and how to get involved.

To address this, it would be possible to organise, in a way similar to (say) OSCON, dedicated tracks for smaller subsets of the desktop, which would still allow some cross-pollination, while providing a small enough surface for newer community members to get some traction.

Also, the more stuff is going on in the conference, the harder it is for any one topic or theme to get attention. I tend to think that the cream will rise to the top, and that people will notice exciting work.

Definitely, the main benefit of the conference has been the economies of scale for sponsors - both in organising attendance and in sponsorship. I think that potentially broadening the conference further will bring greater benefits - making the Desktop Summit *the* place to be to talk about the free software desktop would be a success in my mind.

2. It is hard to measure what specific collaborative benefits are being
made possible by the Desktop Summit. It is hard to point to specific
advances that have been accomplished. Some have concerns that not a
lot of collaboration is actually being done.

Specifically to address this criticism: let me remind people that a lot of this has come down to the mandate which the first desktop summit was given ("GUADEC and Akademy co-hosted") which made the lives of the organising team harder, and also made the conference less successful as a whole. This is something which both the GNOME and KDE eV boards agreed on.

In this last conference, there was a decision early on by the program committee, supported by the boards, to limit the scope of the conference to KDE & GNOME. Again, even though the schedule didn't have KDE & GNOME labels on talks, this created a clear "us & them" - it was obvious which talks were KDE ones, and which were GNOME.

The venue, without a nice central hang-out space where people could mingle, didn't lend itself to the type of mingling and collaboration people hoped for either. And the format (which is the classic Akademy format, and not quite what GUADEC had as a format, although the GUADEC format's been in flux for a few years) meant that a lot of the hackfests/BOFs didn't get the awareness or visibility they could have - people found out about them through word of mouth, and the word of mouth, naturally enough, had trouble crossing project boundaries.

It seems like we can do a lot better in both opening the conference beyond GNOME and KDE, and in improving the format to encourage and facilitate collaboration between key community members working on the same thing.

1. To not have a large combined GNOME+KDE event, and to instead have
a smaller Desktop Summit or focused hackfest(s) with a more clear
agenda to work on specific and measurable collaborative tasks.
GUADEC and Akademy would continue as separate events.

Going back to what I said earlier about the need for a place to gather: I wonder if Plumber's hasn't taken on that role from a "kernel-up" position, and whether the Desktop Summit should aim to do so from a "Desktop-down" perspective? I have nothing against smaller events, but the big advantage of a single annual conference is that there are a *lot* of conferences, meetings, hackfests and whatnot happening these days - do you really want to add one more to people's agenda (which potentially could lead them to avoid GUADEC altogether, and kill the conference in 1 or 2 years)?

2. To arrange the Desktop Summit so that it is more of a co-located
event. The GNOME and KDE events are separate but overlap on
certain days. For example, GUADEC could happen first and continue
for several days, then a few combined days of Desktop Summit
followed by several days of Akademy. This setup would likely be
more complicated for bidding, since it would likely require a
more dynamic space to accommodate the shifting needs.

There was a very clear expression in Gran Canaria that this didn't work - and I think it would be a mistake to try it again, for that reason. And for the reason you mention.

3. The GNOME community has been having trouble finding volunteers to
help make events successful lately. Some people like Dave Neary,
Lennart Poettering, and Ekaterina Gerasimova did a great job
volunteering to make the last Desktop Summit a success. However,
the fact that there were too few volunteers engaged caused some real
issues. Many of the things GNOME folks have complained about the
last Desktop Summit were caused more by a lack of GNOME volunteers
helping than anything else. For a Desktop Summit to be successful,
we need to more clearly see that the GNOME community is more
interested to engage and wanting to get involved.

I don't see the dearth of volunteers you're suggesting. I do see a number of things where volunteer labour has proven insufficient, and where I think it would be worthwhile paying money to make things better: network, graphic design, website maintenance, and conference videos. These are all things that we can improve, but I don't think a lack of motivated volunteers is the key to making that happen.

I'd like to add Chris Kühl (who organised the social occasions) Jon Nordby (BOFs and more), Stormy Peters (sponsorship - after leaving as ED), Patricia Cruz (volunteer co-ordination), Andreas Nilsson (some graphic design) and William from Texas (marketing) - and the 5 other GNOME members of the program committee - who also gave a lot of time, often in the shadows, before and during the conference.

Gathering volunteers is usually a local issue - organisational continuity from one year to another is much more important - how do we ensure that GUADEC organisers know what has gone before and don't remake old mistakes?

That was a challenge with Stormy leaving last year and me having period where I had little time, but we've done a pretty good job of documenting the important stuff in the wiki - even if we don't have a process like LCA's Ghosts of Conferences Past. But that's the key issue we need to ensure works well, whether we're having a DS again or not.

To sum up: I don't think having a DS harms what we hope to get out of a GUADEC, I think we can make it vastly better with small changes in format, and the benefits to sponsors and organisers of the economies of scale are real, and proven. A smaller summit is not a bad idea, but I fear it would hurt GUADEC in the long term. And there's a potential to make a truly Free Desktop Conference which would include not only KDE and GNOME, but also platform and application projects, and at that point we could see the collaboration dialled up to 11.

With all that said, DS is a bigger conference, and expectations and the organisational load change with scale. We need to start spending money to ensure that things which are slipping through the cracks with volunteers get done properly.


Dave Neary
GNOME Foundation member
dneary gnome org
Jabber: nearyd gmail com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]