During the last year I’ve focused a significant part of my effort on driving the GENIVI Development Platform, together with some Codethink colleagues and other GENIVI professionals and community members.
What is GENIVI Development Platform?
GENIVI Development Platform (GDP) is a project and an outcome.
As a project, it can be defined as the delivery side of the GENIVI Alliance. Today is a fairly standard Open Source project, done in the open following many of the most common practices any FLOSS developer would expect.
As an outcome, GDP is a Linux base distribution (Poky) derivative, built with Yocto, that integrates the software that GENIVI community (automotive professionals) develop as Open Source software.
It still a small project but the quality of the platform and the number of people involved has grown this last year significantly.
GENIVI Alliance is a consortium of +140 companies so obviously most of the overall effort is done by paid developers. Changhyeok Bae (community member) together with Tom Pollard and Robert Marshall (Jonathan Maw before him) from Codethink Ltd, constitute the maintainers team, who are responsible for the integration, testing and release of GDP.
These guys are supported by people like myself, doing coordination, marketing, documentation, IT services, infrastructure, testing and many other key tasks which provides the project a level of robustness and scalability that any serious attempt of this nature requires nowadays.
I am interested, where can I get more?
You can find more general information about the project in the following resources:
- GDP Project wiki
- Communication channels:
- Mailing list
- IRC channel: #automotive in irc.freenode.net
- Slides and videos from talks about the project.
GDP-ivi9 is the current stable version although we have moved so fast this last year in terms of the software shipped, that I recommend you to try the following:
- If you are interested in a solid base system, try GDP 11 RC2
- If you are interested in checking the latest UI and demo applications, try GDP 11 RC3
- If you are interested in building from scratch you own images with the latest software, do it directly to GENIVI’s rolling release, called Master for now. You can get the latest software there. It should most ly work since we put stabilization effort on it, following the openSUSE Tumbleweed mindset in this regard.
Currently GDP supports RPi2 & 3, Minnowboard MAX and Turbot, Renesas Porter and Silk and Qualcomm Dragonboard 410c. GDP ships Qt 5.6 at the moment, since it is based in Yocto 2.1…
…which makes GDP a great target for KDE software, specially for Plasma.
GDP and KDE
Putting the effort on having KDE well supported in Yocto would provide the project a third life, landing on an industry that is heavily investing in Open Source with a key piece of software, with no clear competitor today in the open.It would revamp the interest of many KDE developers in porting their apps to embedded/mobile environments and would bring attention to the project from Qt professionals all over the world. Currently KDE is significantly better than anything else that is open in automotive. It would just require the effort to include it and maintain it in Yocto, which is not small, and adapting Plasma a little initially, not much.GENIVI launched a Challenge Grant Program that might help to put some funding in the equation 😉
Whatever effort done to put Plasma on Yocto (so GDP) would also be picked up by GDP’s competitor, AGL UCB (Auto Grade linux Unified Code Base), the Linux Foundation automotive group Linux (again, Yocto) based distribution. So there would be at least two players for the cost of one.
It wouldn’t surprise me if Qt companies would jump in on this effort too. In order to play in the open filed today, they need to Open Source their products, which is a big risk for most of them. Playing with KDE, which is based in the technologies they are familiar with, would be simpler for them. I bet the Qt company would be heavily interested in promoting this effort. It would help to dissipate all the pushing back from the automotive industry to the current Qt license model, GPLv3 based. And it would do it in the best possible way, by providing great ready-to-use software with no competitor. I have been one year preaching about how big the opportunity currently is for KDE, but this is not the kind of challenge that can be sustained on volunteer basis, sadly, since keeping KDE up to date in the Yocto project would require a high level of commitment from KDE as a whole.
The community probably needs first a small success story and some company/corporate push before really jumping on it, I think. The support of a couple of KDE or Qt companies would catalyze the effort.GENIVI and AGL make a significant promotion effort around the world within the automotive industry, participating in forums where KDE is unknown. Many companies that currently develop close source Qt applications for automotive would be interested in KDE which would increase our potential targets for our Corporate program.
Having KDE on GDP and AGL UCB would increase the incentive of developing new applications for many of our young developers who currently do not have automotive as a “professional target“.
Companies like LG, Renesas, Bosch, Hartman, Intel, Jaguar Land Rover, Toyota, Visteon, Fujitsu, Mitsubishi, Volvo, among many others…. are key stakeholders of GENIVI and AGL. Isn’t it this attractive enough?A success story like the one I am proposing would be yet another example of how KDE can play a key role for the Qt ecosystem. Sadly not everybody in this ecosystem understand what a great “tool” KDE can be for them. After a hit like this one, it would be undeniable.Think about the exposure, think about where we are today… Something like this would place KDE where Unity, Android (AOSP) or GNOME are not… yet. I believe this is the kind of strategic decision that would change KDE future. But also the business perspective of those companies (specially Qt ones) who would get involved.
Let’s do it now… or somebody else will.
Update(29/10/16): to find out about the state of KDE in Yocto please read the comment to this article from Samuel Stirtzel. Thanks Samuel.