With only two out of three days of presentations and missing out on the BOF-sessions altogether I am still happy to have made it to the second Desktop Summit in Berlin.
The event operates on a scale that filters out the kind of experimental work you might come across at LGM or even at FOSDEM; with both KDE and GNOME targeting ‘normal users’, this event cares more about high-level work on community developed software and less about specific tools. As an anti-dote it would have been interesting to find projects outside the KDE/GNOME stratosphere such as the OpenBox window-manager, the Fedora design team or even to hear from Canonical’s Unity. That said, the overall atmosphere at the Desktop Summit was vigorous, frank and energetic. These are clearly interesting times, judging from the more than 700 people actively trying to get their heads around developments that continue to radically change the landscape of free software. It is great to see it so much alive and kicking.
The Desktop Summit brings two largish communities together: KDE and GNOME. Both are in essence window-managing projects for the free desktop, and ‘fight for the same mindshare’ as someone expressed it. But their interest in comparing notes seemed genuinely more important than the need to organise two meetings in parallel. The diplomatic agenda of the summit was apparent from the choice of keynotes that carefully bridged interests (hence stayed rather superficial?) and the decision to schedule talks like Stuart Javis’ Why Are We Here, about the future of KDE or GNOME Shell: Iteration’s what you need as single-track. In this way, the only option was to hear each other out.
Many talks were related to toolkits: QT for KDE and Clutter and GTK for GNOME. For me it was interesting to find out more about these collections of widgets, mixing properties of applications, frameworks and standards.
The Summit clearly functioned as an occasion to scope out overlaps, divide territories and articulate differences. It created an atmosphere of reflection and discussion rather than the usual tribal energy that is all about community-affirmation.
The Summit organisation omitted (on purpose or not) to put projects on the conference-badges, and this must have boosted the sales of T-shirts significantly. But if you were like me, not wearing anything that clarified affiliation, people simply came up asking: So, what do you think about GNOME3?
‘Design’ was present in many of the presentations. Essential, lacking, desirable, impossible, hard to manage, needed, hard to scale, wanted, hated and loved: the success of the free desktop is measured by design. Applications for interface-prototyping were discussed at some point, and Boudewijn Rempt showed off Krita but mainly talks tried to make sense of how design-practice could interface with software development; how to design functionalities that are ‘discoverable’ by users.
The much-debated development of GNOME3 (Gnome Shell) seems to finally have broken the Apple-spell and I enjoyed many interesting and self-confident talks about how to integrate User-Interface-design into development cycles, possible relations between software design patterns and … design patterns, the problems and affordances of doing designing in a community. The discussion went far beyond the fear of ‘design by commission’ and in that respect it was important to hear some serious thought being put into the idea that developers might need to learn how to design. “Some people decided to stop writing software so they could develop ‘a pure design mind’. I think that’s not a good thing. We need people that can wear different hats”1
As a way to start such a learning, more and less digested flavors of ‘design thinking’2 were tested out. This method is probably a sensible choice since it seems to have (superficial?) overlaps with software development: iteration, testing etc. It is also rather post-it note-dependent and I wonder how it helps teams to deal with the hard question: what would a free desktop look like?
Donald Norman suggests that users need a ‘coherent mental model’3 to comfortably navigate digital services spread over multiple devices, and participants at the Desktop Summit seemed to be in agreement about the need for such a thing. How might they differ from a proprietary model? What ideas of inter-usability (Claire Rowland) could you imagine to come out of the principles of free software? The GNOME Shell Design Principles offer a modest start, but there were other interesting ideas around as well. Federico Mena Quintero proposed to think of well-designed free software as having the same ‘qualities without name’ that make organically grown cities comfortable for example. His anti-urbanist approach could do with some nuancing, especially if connected to design-learning as an “informal apprenticeships, like craftsmen teaching skills to each other” but I guess we’ll hear more from him about this one day.
A less positivist explanation of their ‘readiness to behave’ is maybe the apparent awareness of why and how business interests cross paths with free desktop development.
Besides the glib contributions of Canonical’s Mark Shuttleworth in the panel on copyright assignment, there was very little outright commercial speak, and no hysterical trying-to-please-business whatsoever. Many presenters spoke from a clear understanding of the dependency on and deep relations with business interests and the free desktop. As Dirk Hohndel said in his keynote: About 90% of the people actively developing free software are also employed by software companies.
The perversity of the roller-coaster relation with companies such as Google (Android) and Nokia (Symbion, Troltech) was clearly brought out during the Panel on Copyright Assignment. The pressure on free software developers to double-license their code is a result of businesses interest in free software fitting their system. The question is of course, if that is in the interest of free software itself.
The excitement felt at the Desktop Summit seems largely fueled by the explosion of smart devices and I am not sure what role the free desktop is actually plotting out for itself. To me it seems back to a depressing lack of standards, working with very expensive devices who’s architecture is kept secret, consuming tons of developers energy before being replaced by yet other devices. This all at a pace that not many people can afford to code for without remuneration.
In a nice session about connecting the desktop to the web, Xan Lopez quoted Alan Kay saying that “People who are really serious about software should make their own hardware”4. Unfortunately, besides a rather gratuitous lightning talk by Michael Meeks about the RepRap project, hardware wasn’t really on the table. Judging from the quality of the discussions, no doubt it will be there at a next iteration of the summit.
- Comment by Pippin in the session Features follow function – design-driven development of GNOME Shell
- “A mental model is an explanation of someone’s thought process about how something works in the real world. It is a representation of the surrounding world, the relationships between its various parts and a person’s intuitive perception about their own acts and their consequences. Our mental models help shape our behaviour and define our approach to solving problems (akin to a personal algorithm) and carrying out tasks.” http://en.wikipedia.org/wiki/Mental_models In his book The Design of Every Day Things, Donald Norman argues that in a well designed object, the mental model of the design should match the mental model of the user.