Foreword
This is the fourth in a series of five daily blog entries covering the various tracks in the Plasma Active initiative. Today we'll be looking at how we plan to distribute the fruits of our labor for use on devices. As we will discover, this is rather new ground for a KDE initiative. It will bring many challenges, but also open new opportunities.

A Missing Link
KDE produces a large quantity of software which traditionally have been aimed at use on laptops, desktops and workstations. Getting our software out to users was fairly straightforward: we released bundles of source code every so often, operating system projects and vendors came along and turned those into binary packages and make them available to their users. KDE has been a wonderful toolkit from which people putting together operating systems can assemble the desktop shell that suits their target audience.
A few years back people started looking to new fields of opportunity: Windows, MacOS and mobile. We don't have participation from the vendors behind either Windows or MacOS so KDE people have had to integrate, build and distribute binaries themselves. This has had mixed results, and really highlights the challenges a community like ours must climb over when the OS vendor is missing from the picture.
Mobile has been another kind of story. At first, we essentially waited on OS groups to do as they did with the desktop: pick up KDE software themselves. This "build it and they will come" approach gave sporadic successes, but it became evident this wasn't going to deliver the results we wanted.
We experimented with pro-actively engaging with operating system vendors. It was difficult to get some vendors to even look up from their own navel long enough to consider the benefits of an active partnership. When we did manage to clear that hurdle the results tended to fall on hard times quickly. (The netbook image, for instance, simply stopped building and there was essentially nothing we ended up being able to do about it.)
We also worked directly with operating systems with a clear mobile focus, including the companies behind MeeGo. The process and politics were often a difficult match for the distributed, agile, community-centric nature of KDE. Where we did find promising footholds, we were still firmly at the mercy of the decision making of others.
It became evident that it was time to be better masters of our own destinies and simply do it the KDE way: community and engineering efforts. With the operating systems track, Plasma Active aims to build an active, direct path to getting our software on an operating system that can be used on consumer devices.
The Wheel That Exists
The last thing we wanted to do was to reinvent wheels and delude ourselves into thinking we could become a full operating system provider ourselves. That wheel has already been invented many times over, particularly in the Linux world. We want to create a diverse ecosystem that retains choice while bringing us a moderate level of predictability to the results. We need to have a clear and powerful voice in how the technology we work so hard on gets put together, and we're going to "stand on the shoulders of giants" to get there.
We have started by working with OpenSLX who produces a (physical) box release of OpenSuse. They have experience in rolling out operating systems to consumers, including providing access to telephone support and creating first-class documentation. Just in time for the Plasma Active wheels to start turning, OpenSLX has rolled out a new product, Balsam Professional, and packages for Plasma Active will be delivered on top of it.
With the packaging and build services provided by OpenSLX, we can provide live, installable disk images for users and developers to download, put on a USB stick and boot on their device from.
By using an existing operating system stack we will avoid taking on that whole set of tasks ourselves. By becoming directly involved, we will be able to ensure it goes in the direction we need it to progress in.
Embracing Existing Initiatives
While this may sound grand, we are also realistic and humble: in a world of robots and 2-D block people, we don't believe there is need or room for more entries to the Linux-for-devices party. We'd also like to take advantage of investments, communities and momentum that is already there.
To that end, we intend to be MeeGo compliant. While the user interface and application selection will be very much driven by Plasma Active, we wish to keep the OS familiar to those working with MeeGo now and compatible with their software.
We currently do not have plans for Android, but that's only because we don't have people yet in the Plasma Active community to work on that. We'd be very happy to see an Android based release as well, so if you are interested in making that happen, please join us and let's get that happening.
MeeGo gives us a great starting point, however: it's the most friendly stack technology-wise for our needs, our friends at Qt and several other groups within the KDE ecosystem are already involved with it and it is openly developed. These are all strong arguments in MeeGo's favor for our needs.
What About Other Operating System Choices?
This does not in any way impact how we approach the traditional desktop or netbook form factors. We are quite happy with how our existing relationships and partnerships work there, and the desktop is not the focus of Plasma Active.
That said, just as we hope that Active Apps will bring increased continuity to the world KDE applications aiming for use on devices, we hope that having a role model operating system that showcases how Plasma Active can be delivered on a device will inspire other operating system projects.
The Plan
Every two weeks, an updated image for download and use will be posted for your enjoyment and use. The frequency of releases should help explain all the dots on the OS Platforms track image at the beginning of this entry. The basic idea is that in each two week cycle a Contour milestone pre-release will be tagged, Plasma Quick will progress and a new Active App (or set of them) will be announced. A few days later, a new OS image will be made available containing all the new hotness.
These bi-weekly images will be pre-releases that bridge the chasm between "the software is out there somewhere" and "it is now running on my device". As such, the target audiences for these builds are:
- early adopter users wanting to watch it unfold and help us with testing and feedback
- developers looking for an easy path to getting Plasma Active on their device
Initially, the images will be for Intel architecture devices and we will be actively testing on the following devices:
- ViewSonic Viewpad 10
- WeTab
- Lenovo Ideapad 10
We have plans to broaden the catalog of devices over the next six months and will be announcing those additions as they come on-line. As the list of recommended devices grows and changes, we will be keeping track of them on the Plasma Active Devices wiki page.
We have a target date in September 2011 to make a first release of the full stack. This will trail the release of KDE Platform and Plasma Workspaces 4.7 by a few weeks, giving us a stable but fresh foundation to launch with.
By bringing together the strands of Plasma Quick, Contour and the catalog of Active Apps together on top of Balsam, we begin to close the loop between developers, users and devices. Ultimately, we wish to provide something to both show to and offer vendors who wish to build exciting, modern devices around. That's a topic for tomorrow's blog entry, however.
How To Get Involved
There are a lot of tasks that remain open here, even with the concerted efforts of the OpenSLX team already being applied. There are packaging tasks, configuration validation (oh, multitouch, thou art a heartless mistress!), hardware support issues to identify and find solutions for, testing to be done, etc. There is a serious amount of work here and we hope to grow this aspect of the Plasma Active community over the next six months.
To follow our progress across Plasma Active, get one of the recommended devices (or if you're more adventurous, something else) and follow the installation instructions.
Be sure to grab new releases as they roll out and give us feedback! You can find us on irc in #active on irc.freenode.net and on the Active mailing list.
If you'd wish to get involved with Plasma Active by packaging software, bringing Plasma Active to new devices or platforms, testing, etc. raise your voice and we'll find a home for your efforts in Plasma Active. We will continue to keep track of open tasks on the wiki in case you are looking for further inspiration.

8 comments:
I have been struggling building KDE from src for a few months, but never really got most parts done, maybe because I am using debian/ubuntu or I don't have much time to fix problems at home. As a result, I haven't got a chance to play with code yet.
So aside from providing pre-built pkgs for some distro, it will be great to have detailed instructions (even a script) to build plasma-active from src so that it can be built easily in "unsupported" distro, more important, making developement involvement easier. btw, I was using kdesrc-build script but it did not work very well.
I would suggest to add ARM support as early as possible, not only because it plays much more important role in mobile world, but because it is almost impossible to build for ARM from src directly, thus, I have to rely on distro's pre-built pkg to try out new things instead of developing.
All these are obstacles for developer to participate since they can't work with HEAD and are distracted by building issues.
And if people would just remember (or know) that Linux kernel is the operating system.... (No, this is not trolling or flamebate starting)
@Fri13 If the Linux kernel is the OS, then I drive to work in a combustion engine.
@ezjd you should use Project Neon. It was purpose-built for you. https://wiki.kubuntu.org/Kubuntu/ProjectNeon
It's been a while since i've heard anything about Linaro, but I would be very interested if you have any plans on interfacing with that as it seems like exactly the purpose linaro is trying to provide.
@ezjd: the packages for plasma Active are build on an OpenSUSE Build Service. I don't know if it's done 'in-house' or on the public buildservice at build.opensuse.org but if it is on the public one you can easily clone the project and build it for other OS'es like Debian, Ubuntu, Fedora, CentOS and more. build.opensuse.org supports 21 operating systems and 6 architectures so there should be something of your liking :D
If you make it build on other operating systems and esp architectures, you can contribute the changes by doing merge requests to the original project (think of OBS as a "github for building packages for a variety of distro's").
If the packages are build on a private Build Service instance, you can ask on the plasma-active ML if they can grant you access.
@IanMonroe
> @Fri13 If the Linux kernel is the OS,
> then I drive to work in a combustion
> engine.
i'd say that yes that is an accurate analogy, and i'm sorry to read that you don't have a GNU-toolkit to sit in, with great KDE dashboard styling, to comfortably steer that engine.
@Ian Monroe OS's are car engines, you dont use car engine to drive. You use wheel, pedals and other parts to drive the car and those are totally different parts.
Engine is just there to make the car move and give electricity (gasoline/diesel) and make the whole car work as a car.
You dont use OS but other softwares to solve your problem. OS just makes all other software to work on the computer.
There are two kinds operating system architectures, monolithics and server-client
Operating system original name was "kernel" when there were only monolithic architecture. Decades later the Server-Client architecture was designed and the name "kernel" did not anymore apply as new name "microkernel" was in place and microkernel did not have all functions but just the core functions of the kernel, where rest were on servers. So new name got popular. Other names were "Master program", "Core", "Nucleus", "Supervisor", "Controller" and so on.
The operating system is that part of the software what exist in supervisor and/or kernel space.
OS does not offer any controllable UI for the user. Without OS any system program or system library would not work, like glib or bash.
Technology has evolved, but it has not changed that the mathematic and software science would not apply todays OS's. The marketing has changed a lot and they lies just to sell their own product as better than others. Lies and partial information to abuse users knowledge to make best choises for them selfs.
Thanks for the reply.
@Ian Monroe: I think project Neon is dead. Last update was in 2009 as I checked again today.
@jospoortvliet: I believe using build service to build KDE pkg for some distro is in pretty good shape now, even with Debian/Ubuntu. However it is more about packaging KDE for a distro than developing KDE itself. For example, I think all the KDE pkgs from OBS is not directly from HEAD of trunk branch and I can't randomly pick a revision to build, also I think Aaron or other KDE developer mainly uses cmake instead of OBS. Unfortunately cmake can't support building for ARM easily.
So it is different than what I need: I want to keep up with latest changes in GIT/SVN (so that I can contribute easily) and directly build from source (git/svn) for ARM.
Post a Comment