After Oneiric’s release last week I had the chance to sit down with Neil Patel, Unity’s Technical Lead, to talk about how he felt the Oneiric cycle went and to get his thoughts on Precise Pangolin 12.04.
Neil has been working for Canonical for a couple of years, and was previously heavily involved in the original Netbook Remix Launcher which was superseded by Unity originally in Ubuntu Netbook Edition 10.10.
As the Technical Lead for Unity, he oversees a lot of the development on the new shell for Ubuntu.
Neil is the guy you go to with your questions about the development of Unity, and that’s exactly what we did when I sat down for an interview with him last week…
Oneiric introduced a lot of changes, most notably the leap to GNOME 3. How do you feel this cycle has gone? Did you achieve everything you wanted to?
Overall, I’m happy with the cycle.
I think we just missed out on landing some cool new features, but we spent a lot of time going over core parts of Unity and making them much more robust, which was needed. I think people will only see the true fruit of this work in the coming week or so when we get our last SRU out of testing and into the archive. We’ve been working until release day and for the couple of weeks since. It’s always a gruelling task but I feel that, in the end, we’ve done well.
I think the GNOME 3 transition went as well as could be imagined. Let’s say I’m happy we have another six months to fix things before LTS ;)
If you look back on this cycle, the main things we had on our plate for Unity were Compiz stacking issues, memory usage, the new Dash/Lenses ideas and Alt+Tab. I’m happy with our status on the stacking issues and, testing what is in Ubuntu proposed, we’ve been unable to reproduce the issue even when using some of the more nefarious applications!
I think the new Dash and Lenses have been received well by users and developers. In addition to re-thinking the what a Lens is, introducing the ideas of a container (the Lens) and its data sources (Scopes), we also heavily tweaked the Dash to make it nicer to use and easier to find. Having a chance to go over the code, the team did heavy performance tuning and memory-leak plugging at the same time, giving us a lean, mean, but featureful Dash for 11.10.
One of the larger changes was replacing the default Alt+tab experience for 11.10. I’m a huge fan of the new Alt+tab, and I think it’s a big improvement over the previous one. Between Alt+tab (switch between windows + applications) and Alt+` (switch between windows of the current application), I’m set for window-switching! I should say that we’ve got great feedback about it and through that have identified some interesting edge-cases which we hope to solve for 12.04.
Finally, we also spent a lot of time on general memory usage and performance, allowing us a to comb through and find any leaks/bugs in our code. At the same time, we were able to find and fix leaks across the stack, including in parts of graphics drivers which had never been exercised before (a general problem we face with how much Unity pushes desktop graphics). The outcome of that is Unity (and its component daemons/libraries) using far less memory, being more robust and performing better. That said, there are still memory leaks to chase down, crashes to fix and increased performance to be found.
I mean, if we fixed everything, we’d have nothing to do for Pangolin, right? ;)
What will the Unity team be focused on for Pangolin?
The course of Unity was planned as Natty=features, Oneiric=features/bug-fixing, Pangolin=bug-fixing. That’s not to say Unity won’t evolve for Pangolin, but what it really comes down to is that we know and respect that this is an LTS release so, before anything, we must deliver a completely solid desktop experience. A large part of this will equate to reducing our bug-count, which will not only make Unity more stable, but also apply polish across the board, giving users a delightful experience when they install Pangolin.
“…what it really comes down to is that we know and respect that this is an LTS release so, before anything, we must deliver a completely solid desktop experience.”
In addition to that, there is also what we’ve termed a “design debt” that Unity has built up over the past two releases. This is the difference between the specifications that the Design team had produced and what we were able to implement during the two release cycles. This debt is sometimes as small as the Dash not looking quite correct, to as large as there is a significant feature missing. The idea is to reduce this debt as much as possible for Pangolin, allowing users to experience the full Unity user experience as designed, which we feel will answer a lot of questions and plug a lot of holes in day-to-day usage.
In addition to stability and addressing the design debt, we will be heavily focused on power usage, memory usage and performance for Pangolin for Unity, and as a general goal for the release as a whole.
Work in all these areas should give you the most polished desktop experience available for 12.04.
Are there going to be many design changes, or is the basic Unity concept now finalized?
I’d say that the general concept is finalised, but people still have a lot to look forward to as we plug the design debt and as Unity starts becoming more and more stable and polished.
Which areas of Ubuntu niggle you the most?
In the distribution, I’d say our core applications aren’t comparable to the competition, though they might be the best available to us. I wish we would spend more time on working on applications as I think that matters as much as the shell, if not more.
In the project, I’d say our application development story still needs a lot of work. It still doesn’t seem to me that we are making people excited to write applications for Ubuntu. I think that ties into what I mentioned above: if we can make the experience of the core applications absolutely rock, making them delightful to both look at and use, then It think we could get people excited about making similarly beautiful applications.
“I wish we would spend more time on working on applications as I think that matters as much as the shell, if not more.”
Will there be any Canonical made patches to rectify some of the feature regressions that GNOME 3 has introduced?
Oh yes, or at least, I hope so :)
I think there is a bunch of work that needs to be done to get it back into line with what we had before the transition, though it might be an excellent time to also apply some much-needed polish in those areas too.
Although graphics isn’t your focus, do you have any idea whether 12.04 will attempt to fix some of the multiple monitor bugs that Oneiric introduced?
Yes! It’s on everyone’s hit-list as something we (the entire distro) have to nail for 12.04. I know design have been working hard on making sure that Unity’s user experience can correctly carry over onto multiple monitors (outside of just adding another panel like we do now), but also I believe that we’ll be trying to pull that scenario into our distribution testing too, making sure that we try and avoid regressions there.
That said, yes, we are at the mercy of graphics drivers and their state in the next six months (e.g. the issues with radeon this cycle), but we’ll try and do the best we can, and something that would be interesting is to start holding back Xorg/kernel uploads if they are breaking certain chipsets etc, but that’s just me thinking out aloud :)
Can we expect Wayland (the in-development display server) any time soon?
Honestly, I’m really not sure.
All the effort from the DX team so far has been to make sure we try and make the experience on the standard Xorg setup as beautiful, featureful, and stable as possible. I haven’t had time this cycle to really look into Wayland’s status, though I think that it has a session at UDS discussing it in relation to 12.04. It might just be that we need to get Pangolin out the door, with all the goals met, before we have a serious look at Wayland.
And, truthfully, Wayland might need that extra time too, before the Ubuntu masses start reporting bugs against it ;)
Thanks Neil for taking the time for this interview. It sure is sounding like Precise will be a great release, with some much needed polish. You can follow Neil on Twitter @ http://twitter.com/njpatel
Lastly, if you haven’t seen our Ubuntu 11.10 feature overview video, check it out: