Discovering the New Gnome
The last few years have been troubled for the Gnome Project. Once a premier desktop environment for Linux, it has seen its market share diminish amid user dissatisfaction over Gnome 3 and accusations that the project was ignoring users. Yet, over the last six months, something important has been happening: Slowly and quietly, the members of Gnome have started trying to turn the situation around.
The process began at GUADEC, Gnome’s annual conference, in July 2012, with a presentation titled “A Bright Future for Gnome,” which presented an unflinching yet optimistic view of Gnome’s prospects and was followed by a number of discussions about the possibilities. It continued several months later at the annual Boston Summit.
“Since then,” says Karen Sandler, Executive Director of the Gnome Foundation, “these conversations have been continuing in a number of fora, including on foundation-list, our email list for members of the Gnome Foundation, and on desktop-devel-list, which is open to all.”
Perhaps the most important immediate result is a change of leadership. Ever since development on Gnome 3 began several years ago, Gnome’s priorities appear to have been shaped – at least unofficially – by the interface designers and their ideas.
Given the innovations in Gnome 3, that was probably inevitable. However, according to release team public representative Matthias Clasen, “One clear take-away from GUADEC this year was that the release team is expected to provide more direction. The community expects the release team to provide leadership and goals that go beyond just doing half-yearly releases of a desktop with a few cool features.”
One of the first public indications of this change was the announcement by Clasen that Gnome would drop its fallback mode for those who lacked the 3D drivers necessary for Gnome Shell. Instead, Gnome would emphasize a set of core extensions to recreate a Gnome 2-like interface for those who needed or wanted it. The announcement was a reversal of the project’s reaction to both extensions and the continuing interest among users in Gnome 2, both of which had been largely ignored until Clasen’s announcement.
Meanwhile, where just six months ago, many were debating the truth of Benjamin Otte’s depiction of Gnome as demoralized and understaffed, according to Sandler, the last few months have seen the return of “people who used to be much more involved in the Gnome community who have largely moved on as well as newcomers who have recently joined.”
In particular, she notes the success of the Gnome Outreach Program for Women as an important source of new project members. The project has been so successful that in its next round, it will expand to 11 other projects.
All these seem small, but definite signs that Gnome is attempting to reinvent itself at last.
Weaknesses and Strengths
Gnome enters this seemingly new stage with a mixed legacy. Although representatives of the project naturally prefer not to dwell on the problems of the immediate past, Sandler does acknowledge that Gnome 3’s “early releases had a lot of bugs” and suggests that “the negative commentary surrounding that has caused some people not to give Gnome 3 another try.”
On the technical side, Clasen suggests that Gnome suffers from the general weakness in free software of how software is deployed. The developers of an application, he notes, either have to get their work included in a distribution or else produce packages for multiple distributions themselves. “As a consequence, there are far too few interesting applications out there,” he says, and those that projects like Gnome produce are apt to be overlooked.
Both Sandler and Clasen prefer to emphasize Gnome’s strengths. Sandler mentions the flexibility of Gnome’s extensions, as well as its accessibility applications like Orca, which remain the most advanced free software accessibility tools available.
Similarly, Clasen emphasizes the improvements inspired by user feedback in Gnome 3’s overview and message tray. Additionally, as Gnome 3 matures, apps designed to take advantage of its features are starting to be released. “These applications install search providers to make the shell search work as intended, they populate the application menu consistently, and they use notifications in a way that works well with the message tray.”
Even more importantly, he emphasizes, is the smooth-working relation between Gnome’s designers and developers. Clasen describes the current situation as:
a night-and-day difference from how things used to work in earlier days. There is a lot of information that flows in both directions: we partially implement early mockups, take screenshots, or the designers try to build the branches; the developers explain what can’t be implemented with available technology, the developers clarify subtle points of the design, and so on. When you see a new feature appear for the first time in a development release, it has usually gone through several rounds of this process, nowadays.
Another strength Gnome has as it reinvents itself is the increasing separation between the interface – Gnome Shell – and the rest of the desktop environment. Although Gnome itself is primarily concerned with building what Clasen calls a “cohesive product” that combines both, it also accommodates other projects, like Ubuntu’s Unity or Linux Mint’s Mate and Cinnamon, that add their own interfaces while using Gnome 3’s technology. In fact, while Gnome itself has declined in user popularity, when all these usages are combined, Gnome could still claim to be the dominant alternative on the desktop.
The Immediate Road Map
The next year or so promises to bring the first signs of changes to Gnome. As a member of the release team, Clasen himself is dedicated to “continu[ing] to listen to user feedback and make classic mode useful” – that is, identifying the core extensions to provide a Gnome 2-like interface for those who want one. Similarly, Clasen hopes to see more applications redesigned for Gnome 3, with a resulting update to the Human Interface Guidelines (HIG) that describe how a Gnome-based app should look and behave.
More ambitiously, Clasen would like to see support for touch screens in upcoming releases:
It is pretty clear to me that Gnome will stay mainly focused on the laptop form factor. That is what most of us are using Gnome on every day, and that is where most of our users are. Traditional workstations are not going away, either – we have in fact improved our support in this area, with color management and support for graphics tablets. But touch is important, regardless. We need to work well with modern hardware, and touch screens are a part of that. Now that Windows 8 is out, we will increasingly see touch support on laptops and hybrid devices.
Among the project’s contributors, another upcoming change is a switch to “continuous integration and testing,” which is also being called “Testable.” Clasen explains that Gnome has been using JHBuild to build the desktop layer from Git repositories or tarballs, but the tool has serious limitations. “It is particularly problematic for testing any operating system integration features that require newer system services. Developing something like a new login screen is not supported at all by JHBuild.
Instead, Gnome is switching to OSTree, which is capable of building complete operating systems. Other useful features of OSTree include “check[ing] out entire system images with history while being space efficient, and easy bisection and rollback. In a nutshell, OSTree is aiming to be a ‘git for entire filesystems and binaries,’ with up-to-the-minute access to builds of the latest commits.” With the use of OSTree, Gnome should be able to offer a standalone Gnome installation for testing and development. This change has been loosely described as “Gnome OS.” Use of the term has led to the assumption that Gnome is planning to produce its own distribution, which would be a controversial switch of its role as an upstream supplier to distributions.
However, that is not the intention at all. “This term has scared many people,” Clasen notes, “so let me be very clear: This is not an attempt to replace distributions. All the improvements that we hope to make through this initiative will directly improve what the Gnome project is able to offer distributions.” Clasen goes on to observe that “the day-to-day buildability of the Gnome stack has markedly improved since we started this effort.” He also hopes that OSTree will finally eliminate an ongoing problem with interns: “Next year we don’t want to have any Summer of Code student give up or fail because they can’t get through a JHBuild of the entire stack” – a time-consuming task that requires expert knowledge of what can be safely skipped if problems arise.
To Gnome 4 and Beyond
The conversations about Gnome’s future has included some far-ranging speculation, including having laptops or tablets shipped with Gnome preinstalled. However, although the project’s mailing lists have seen some talk about what a version of Gnome for new form factors might look like (e.g., should a Gnome tablet have a terminal?) neither Clasen nor Sandler would commit to making any suggestions about what might happen in the long term.
“We don’t have very detailed roadmaps,” Clasen said frankly.
The most that Sandler would say is that, while the Gnome Foundation is open to the idea of commercial ventures or corporate partnership, “anything undertaken will be done with our values and community at the forefront.”
On the technical side, Clasen accepts the idea that Gnome 4.0 might be released in two or three years but is currently envisioning its main features as increased integration of cloud services. “There will be support for more online accounts, including more specialized ones. All the core applications will provide search results to Gnome Shell’s overview, which is the central entry point to accessing your stuff.”
“All of this,” he adds, “is just the gradual completion of the Gnome 3 design. Beyond that, we will have an SDK that supports independent application development. Important milestones along the way are figuring out if and how we can sandbox applications, and a better way to distribute and deploy applications.”
The skepticism that greeted Clasen’s announcement about replacing fallback mode suggests that Gnome’s problems are not over yet. If nothing else, many users may take some convincing that Gnome is really transforming itself.
However, if the optimism sometimes sounds a little forced or as much hopeful as actual, it does seem that Gnome is taking the first tentative steps to renewal. The process may take months – or more likely years – but the first steps are being taken with some success.