I haven't been around munch since GNOME 3.0 launched because I've been working really (really!) hard by joining Zave Networks and working to make this happen: Google Acquires Digital Coupons/Incentives Platform Zave Networks To Bulk Up Commerce
I'm joining Google's ranks and am looking forward to many more wonderful things in the years to come!
I'm too exhausted to make a giant blog post about it right now, but I'm proud of what we, the GNOME community, have achieved. It's time to celebrate! Congratulations, everyone, and thank you for the amazing amount of hard work and can-do attitude that kept us on-track these past few weeks!
I have never been so proud to be a contributor to GNOME as I am right now. So many people are working so hard to deliver an awesome 3.0 release it would be impossible to name them all. In many cases people are taking time away from their families or personal lives to give our community the best launch experience—on schedule!—that we can possibly achieve. This is going to be the most user-focused release of GNOME, ever.
The past few weekends I've been skipping time away from friends and family to achieve the most visible contribution that I have made this cycle which has been to spin some marketing videos which we are going to use to promote GNOME 3.0 on launch day. The thinking is that short videos will increase the audience—especially among those who don't like reading release notes.
The first two videos are posted up on our YouTube account. We're using YouTube's HTML5 embedding mode to stream the videos in WebM on gnome3.org and work is underway to use Universal Subtitles for i18n and a11y purposes. That should be ready by launch day.
I'm going to use every last spare moment this weekend to finish as many as I possibly can in advance of our launch next week.
Let's get the word out about GNOME 3's better user experience!
GNOME Shell (both master and the overview-relayout branch) present the list of launchable applications to us as a large, non-hierarchical, immediately searchable list. We can hit the logo key and start typing a few letters and immediately find what we're looking for. KDE has it. Windows Vista and 7 have it. OS X has it (via Spotlight). And now we have it.
On Debian this is a problem because menu and menu-xdg convert an ancient, Debian, unified menu system that predates the existence of the cross-desktop XDG menu specs in to XDG menus and stores them in
/var/lib/menu-xdg duplicating everything that's already in the standard
/usr/share/applications. This wasn't really a problem because all the duplication was crammed in to an easily avoidable "Debian" menu. The idea was that you'd get the same Debian menu in all desktops. Unfortunately, this isn't getting killed off fast enough. However, we can change it for our own systems.
sudo update-menus --remove
sudo apt-get purge menu
The first step is required because
/var/lib/menu-xdg is not considered configuration and therefore the contents of this directory are not removed on package removal. It is necessary to the remove the menu package to prevent package installation from triggering the recreation of these menus via the dpkg triggers mechanism.
I spent most of the day doing content production, so no session blogs from me.
Last night's Beer Summit hosted by Collabora: photos
Prototype (must all be re-filmed in three months) GNOME 3 launch video: WebM OGV (music CC licensed under Sampling Plus 1.0 by Jonathan Yamoty but may ultimately be replaced by a new composition by Joey)
Vincent opened up the discussion by doing a long review of the proposal which was discussed on desktop-devel (http://mail.gnome.org/archives/devel-ann
|From GNOME Summit 2010|
There were two major objections: translators were worried about quality of translations and having to create a new account to patch or contribute to projects that aren't hosted on gnome.org. To address translator objections, r-t is working with translation teams to come up with a more stringent quality guidelines and processes for Featured Applications.
Ted Gould brought up his proposal from the mailing list to define a core GNOME desktop moduleset that doesn't include Shell so that Ubuntu can continue to say that they embrace GNOME. There was some disagreement about this and the conversation was postponed until tomorrow.
What happened next was, in a nutshell, a one and half hour long disagreement about the definition of Featured Applications, whether it makes sense to expend the energy on the delineation at all, and whether the designation has any value for GNOME and for the application receiving it. It would be too difficult to cover all aspects of this discussion among the eight parties whom were principally involved. And, since I was one of those parties, it wouldn't be entirely objective. However, there was some agreement and some things left unresolved.
In the category of agreement the following aspects were clear: the definition of GNOME is the Core and Platform modulesets and the wider GNOME infrastructure hosted and supported modules are--regardless of whatever else they are designated--at the very least GNOME Project. There was also agreement that we, as a project, want to promote good applications in our ecosystem, in the short term, and, in the long term, go the direction of something like an app. store.
With regard to areas of disagreement, the session ended without much conclusions, however, a discussion afterward between a few parties sounded optimistic and so we may see a way forward announced in the coming weeks.
Ted Gould opened the session by introducing the stack and its relationship to the stack which is now based Compiz.
He introduced the Unity notification approach. In the next Unity release, it will allow libnotifier and only traditional status icon messages from Wine and Java (all other status icon clients will be rejected).
Ted gave an overview of some of the use cases in Ubuntu for libindicate: Evolution, Gwibber and Empathy. A critical point is that libindicate is a persistent: if an app crashes, the indicator is taken with it. Indicator also has MPRIS support and Ubuntu's patches GTK+ and Qt have AppMenu support.
To do application tracking, they developed a library called BAMF which talks to X11, DBUS, and desktop files.
Application menus are split in to offline and online division: offline derived from the .desktop file and online derived from the Ubuntu AppIndicator spec.
Ted surveyed Places which is a mechanism for defining search and browseable categories for click-able items: applications, Tracker file results, web history, application history.
There were questions about the decision to move to Compiz from Mutter; nothing new was said that hasn't been written elsewhere and so I will not rehash it here.
I asked about the long-term vision of Unity; Ted replied that they really isn't any huge long-term goals beyond what has already been announced for Natty. I asked if there was an extensibility plan; Ted replied that extensions should be done through AppIndicators.
Federico started a long side conversation about how to get one, unified Activity Journal interface shared between Shell and Unity. There was much discussion about technical feasibility of various approaches and the design philosophy of Shell. Because of the Compiz/Clutter split, the only lowest common denominator would be to XEmbed an alpha transparent window rendered by AJ. Owen said that this approach would have all kinds of problematic side effects like, for example, not working with the search results. There was a lot of desire to come up with some kind of solution so we may see something specifically for AJ.
Xan opened the session by surveying the history of browsers in GNOME: Netscape, Mozilla, Firefox, Chrome. The status of these browsers in GNOME has varying levels of integration but--of those remaining--they have vastly more resources than Epiphany currently does.
|From GNOME Summit 2010|
Xan proposed that we change our approach: browser as a service.
- Integrated tab management
- Web application integration (GMail), better Shell treatment, icons
- Search results in the Shell
- Direct results for unit conversion in search results
At this point Xan opened the floor to brainstorming for more ideas.
Next, the discussion moved to how to more deeply integrate WebkitGTK in to the platform. There was clear agreement that there should be two embedding API's: one where you want a browser-like experience and the other where you don't need chrome, cookies, caches, etc.
Jon McCann said that a good tab-window manager integration will require a lot of work so I should probably be a little later. Search integration is a more short-term objective.
There was a lot of discussion about what kinds of results could be shown in the search window. For example what API's are out there, whether it makes sense allow the provider to show ads if their EULA requires it, etc.
Owen opened the session by saying that historically we have relied on the Board and Marketing Team to articulate our goals and that that hasn't been fair to either of them. The motivation for this session was to set goals as a community.
He set some guidelines for goals: motivational, realistic, determinative.
To start the conversation, he took a look back at some old GNOME goals:
- Build a Free Software replacement Windows
- Pretty good goal: motivational, effective
- Ten years later it turns out that Windows doesn't matter that much
- 10% Market Share by 2010
- Bad goal
- GNOME Online Desktop is to adapt is to adapt the desktop to become the perfect window to online service
- Not motivation because it makes the desktop irrelevant
|From GNOME Summit 2010|
Owen tried to capture some major themes that we can all agree on, at the moment:
- GNOME is a community of people building Free Software for users
- Computing space is mind-boggling big: billions of users, markets, dollars
- We don't need to dominate the market to be a successful project
- But we do need to provide something great to our users
- We can't be "for users" unless we meaningfully control the user experience
- A desktop OS is only a small piece of the computing experience
The question Owen proposed to stimulate the discussion:
Giving all your data to Facebook or Google provides a great user experience. How do we provide an experience in the control of the user that is as good? Better?
With that he opening the floor to brainstorming.
Where should GNOME be going?
Ideas thrown out:
- Work with devices better
- Sync contacts
- Work with all devices
- integrate with home media
- Don't have an agenda--support user's choice
- Concentrate on things that aren't natural web apps (content creating)
- Work with web apps
- Put user more in control of existing apps
- Extract/backup data
- Shotwell example
- Single setup of web services across desktop
- Proactively working with services to adapt to changes
- Proxy so we don't need SW update
- Provide long term support story
- Technology provider
- Align with Mozilla--consortium for user control
- Focus on offline experience
- Web apps as 1st class "apps" /Deep integration of web/desktop; create standard for web developers
- App menus, Jumplists
- A beautiful experience (revisit HIG, animations, polish)
- Run on other form factors
- App. development platform/developer experience
- Best environment for developers
- Not just desktop developers
- Work with hardware vendors (OEM's)
- A knowledge base with direct desktop apps. links
- Target an audience
Owen brought the discussion back together and asked for any ideas on how to boil all of this down in to more coherent, articulable goals.
An over-arching theme of the ensuing discussion was on how difficult it would be to boil this down but that--historically--people have come up with great ideas and sold it to the rest of the community. He hoped that would happen again. There was a lot of discussion about what the end result of all of this brainstorming would be.
The session was closed without any conclusions and with a plan to get wider community participation. Owen will be sending out an email or a blog post on the topic, soon.