Interview with Vincent Untz at Gnome Asia Summit 2011

No Comments

Vincent Untz is the uber-geek release manager for Gnome. We catch him at the Gnome Asia Summit in Bangalore for a quick chat on all things Gnome 3 (and some things not).

Mayank Sharma: What's your role in the GNOME project?

Vincent Untz: My current role is Release Manager of GNOME, which more or less means cheering the release team, because in the release team we all work together, and we are all on the same equal footing.

The release team exists to make sure that we release Gnome on time with the quality we all expect, and to make important decisions when decisions need to be taken, like which new modules are to be part of Gnome, to publish the schedule for the release cycle, and everything that is related to release engineering a big project like Gnome. Because Gnome is actually made of more than 100 modules, so it's really a lot of coordination between the modules to make sure everything works.

So that's my main role. I am also maintaining a few modules, and doing many other things.


MS: How has this role evolved from Gnome 2 to Gnome 3?

VU: That's quite interesting because during the whole GNOME 2 history, the role of the release team moved slowly to a role which is um...everybody got used to their roles and we didn't have to do a lot of work because everybody was so used and accepting those roles that they were just self-behaving themselves. So that was really all fine.

The decision to do Gnome 3 was a big decision. We have been talking about that since 2005 I believe, or even earlier than that. And coming with a plan on how to do it, and what was the idea behind Gnome 3 is something that big projects like Gnome can take in a distributed way. Some people have to lead to explain the vision. So that was one of the big changes in the release team of Gnome 3.0.

Another thing that has changed considerably is that since Gnome 3 is going to be a big release we wanted it to be really "perfect". I mean it's not going to be perfect, there are going to be bugs, but we have been pushing to do many many changes, even at the last minute. Like, usually we really freeze most of the code that then doesn't change any more. But for this cycle specifically we've had many more requests to change the code just to make sure that Gnome 3.0 will really look like what we wanted it to look like. So that required a lot more time than usual.

The release team also discussed how to reorganize the module, how to define what is part of Gnome, what is a Gnome module. So we led some discussion in the community about that.

MS: What sort of reorganization?

VU: Usually before in Gnome 2, we had a really strict definition of "What is Gnome". So we had a list of modules, a list of applications, and we were saying "This is Gnome". And you also had projects like GIMP, Pidgin, and any other project that is using Gnome technologies, and we found out that these projects didn't feel like part of Gnome because they weren't part of the official Gnome release.

So we wanted to address that and we proposed a reorganisation where we have the core Gnome, which is what you need to run the desktop, and then you have applications. We want application maintainers to feel part of the Gnome project, of the Gnome community, so we told that all of the applications which are not part of Gnome we welcome them as Gnome applications and we want them to be proud to be Gnome applications.

So we feature some specific applications like Banshee, Rhythmbox, Gimp, Inkscape. This is just a subset of really really good applications that people love that weren't necessarily part of Gnome before but that we feel are really a good part of the Gnome experience for our users. Because the experience is not just the desktop but also the applications that they use.

MS: Did KDE have any influence on this reorganization?

VU: Our model is a bit similar, but the way KDE does it is different because when they release a new version of KDE they release all applications at the same time, it's all synchronized and all that.

We want to encourage that but we don't want to force application developers to feel that they have to do exactly the same thing as Gnome. So we encourage people to follow our development cycle because we think it's a good cycle, distributions are using it also so it works really well for everybody. We even encourage application developers to use our freezes to ensure that the quality of the applications is always really good. But we don't enforce all that on applications.

MS: So it's like a set of guidelines?

VU: Yes.

MS: How do balance your work at Novell on OpenSUSE and your work with Gnome?

VU: So my position at Novell is an OpenSUSE booster. The role of our team is to make sure that we attract a lot of people inside OpenSUSE community; that we encourage the development of OpenSUSE.

So obviously there's no direct connection in my work with Gnome upstream and my work in OpenSUSE. But we consider that the work we do upstream gives some visibility to OpenSUSE. That it helps people see that OpenSUSE is a good distribution for Gnome, for example in my case but also for KDE in the case of my colleagues.

Being actively involved upstream is a good way to know what's going on to stay informed of that and to guarantee the integration of Gnome inside of OpenSUSE is really good. So it's really important to stay involved upstream. So I do consider that as part of my job and I think Novell also considers this as part of my job.

MS: Because everybody thinks of OpenSUSE as a KDE distro.

VU: Yeah but actually we are really strong on both KDE and Gnome.

MS: So part of your job is also to change that perception?

VU: My goal is not to change the perception of Gnome in OpenSUSE. My goal when I work on Gnome is to make sure that Gnome is absolutely fantastic. But to achieve that I also want to make Gnome absolutely fantastic in OpenSUSE, so that's also an activity I am working on. But I don't consider that part of job as something I do upstream. I don't want to actively change the perception of OpenSUSE upstream. That's just a consequence of my activity upstream.


MS: What is the Gnome "philosophy" that you and Allan (Day, from Gnome marketing) keep referring to?

VU: The Gnome philosophy covers a lot of things. It's about our values and we have values that haven't changed with Gnome 3. Like we care deeply about freedom, and we want it to be accessible to as many users as possible, that includes accessibility for handicapped users but also localization, the usability of the software, and more.

Also the release engineering methods we use, like the 6 month cycle, is really something that really good part of the free software ecosystem adopts following what Gnome did. So we also try to keep this development cycle and guarantee that what we have even in Git, not just in tarballs that people use, but also in Git is always usable. So we guarantee a minimal quality.

I think what Allan refered to in his blog post was the philosophy with respect to design. What's interesting is that in Gnome 2.x we were talking a lot about usability. We have the human interface guidelines which say about the spacing between the widgets, and the arrangement of windows, and that kind of stuff. In Gnome 3 what we changed is that we still believe in usability that it is still something that we are working on but if you step back a bit there's the user experience as a whole and you cannot define a user's experience just based on a set of rules.

This is something we didn't really have in Gnome 2, and that we are really changing in Gnome 3. So I would say that's really the big philosophy shift between Gnome 2 and Gnome 3.

MS: Plasma can be used to build interfaces such as the Plasma netbook, and the Plasma mobile. How does Gnome 3 adapt to these environments?

VU: So the design of Gnome 3 is actually produced to work fine on multiple set of devices, multiple size factors. So we have active developers working on the Gnome Shell to make sure it's usable on the netbooks.

MS: So the appearance of the desktop is consistent across these devices?VU: KDE has a different user interface for the netbook and a normal desktop. We don't think it's necessary. But if you go to other sizes such as tablets or mobile phones obviously that's not the current target of Gnome. We are not working actively on porting Gnome to those devices. Some people are really interested in making sure that Gnome 3 is working fine with touch screens, for example, so it's going to happen.

What's also important is that even if you have what Plasma does -- different users interfaces depending on the devices, you still need the applications to be adapted, and we don't have that right now. The applications are really targeted at the desktop.

But that being said the Gnome technologies outside of the desktop, are being used by many many different kind of devices. There are some phones by Nokia, we have OLPC, we have Amazon Kindle, and lots more.

MS: And what about running KDE apps in GNOME 3?

VU: On Gnome 2, KDE apps were working really well as long as you have the right theme that Qt understands. On Gnome 3, that's not the case yet I assume, because we have moved to Gtk 3, so the way the apps are themed is different. So I believe that KDE apps will not be as well integrated right now in Gnome 3 as they were in Gnome 2. It's going to be fixed obviously.

That being said they are working anyway. They might have a different look and feel but you can use everything in the application. So there's no integration problem in terms of functionality, it's just really the look and feel.


MS: What's the latest on the cross distribution application centre?

VU: We started discussing about it last year in October and we did a meeting with people from Debian, Mageia, Fedora, OpenSUSE, and Ubuntu in January in Germany. The way to find and install new applications, is something that all distros wanted to fix.

We have great package managers, but users don't understand what a package is. They don't want to know that you need this library to use this application. So it's really about focussing on applications instead of packages and adding on top of that what everybody expects nowadays like comments, ratings, screenshots, and that kind of metadata about the apps.

We all agreed on the way forward, which is a bit surprising because collaboration efforts between multiple distros is always difficult. So we developed a roadmap of what we want to do and the goal is to have something usable for the distributions that are going to be released at the end of this year.

Right now we are a bit late on the schedule mostly because we have some people working on the release of Gnome 3, and on the release of OpenSUSE 11.4 last month, and on Ubuntu 11.04 this month. But it's going on and the idea is to provide an application centre where people can easily just look for apps, find new apps, find apps that might be related to apps they already use, and really offer those kind of features to find apps to users.

MS: How difficult is it to implement technically?

VU: It's actually not that difficult. For the UI we are going to use the Ubuntu Software Centre which is really a good piece of software. It needs to be adapted to work on different distros obviously. We also need some additional pieces in the architecture like the backend for storing the metadata that the user contributes. So we are using a protocol that was defined before the appstream project which is called the Open Collaboration Specification. So we are actually not inventing new things but adapting what exists, so it's much easier to achieve. But we still need to write some pieces of code, but it's working well.

The goal is to have it at the end of the year. We have explicitly set a goal which is ambitious because we don't want to wait too long.

MS: Given your dislike on how Ubuntu handled the Banshee revenue stream issue, how can you collaborate with Ubuntu on the application centre?

VU: They are really two different things. When we work with people we work with people and not companies. I do have many many good friends at Canonical and I have no issues working with them and I have lots of respect for them. The Banshee decision is a company decision it's really made by Canonical, not by the people inside Canonical.

So that's one of the things in free software -- we are able to work together even if we disagree on some other things. That's because we really work with people instead of companies. We don't work with RedHat, we don't work with Novell, we work with people inside those companies. So there's absolutely no issue for collaboration.

Images: Shashank Sharma

Comments are closed for this post