I’m now on twitter. I’ll use that for small status updates on the various open source related work I’m doing, e.g. FSO, OpenEmbedded, Vala, and the like.

Follow me, if you can

After Trolltech dropping the ball with the community back in the old days of Opie, I pretty much gave up on Qt (and C++) apart from accepting some contract work, so my C++/Qt skills would not get too rusty. Since my nightmares with getting something fluid out of Gtk+ (back in the Openmoko days), I did not have the chance to do much UI work — the freesmartphone.org middleware kept me busy enough.

I have been watching Qt progressing though, and ever since they introduced Qt Kinetic and QML it became very interesting for me again. QML looks like EFL’s Edje been thought through — don’t get me wrong, Edje was groundbreaking (as most of Rasterman’s work) when it made its debut, however in my opinion it got stuck in the middle and never lived up to what I was expecting from it.

Once QML ships with Qt — hopefully in the next minor or at least major version of Qt, I will get back on doing some FOSS work on application level to complete creating a smart phone stack. That’s going to be fun!

Just came back from FOSDEM 2010, which — after skipping the last incarnation — was a great inspiring and productive event. The Openmoko devroom we originally requested was declined, however thanks to the initiative of Serdar Dere, it turned out we could snatch a last minute 3 hours timeslot that was left open by the Xorg guys. Very shortly we prepared a schedule and managed to get a nice program which was very well received.

Due to the short notice, we could not manage to create a video recording infrastructure, so I’m afraid this year we can only provide the slides — which are a notoriously bad substitute for real talks though. We try to improve for next year — if we can get a devroom again. The pictures you are seeing are courtesy Dr. Nikolaus Schaller from Goldelico, btw. — thanks!

The FOSDEM team did certainly improve its organization over the last years, I was very pleased to see some of my criticism being taken into account. Apart from the lack of good coffee in Brussels (which the FOSDEM team probably is unguilty for), I can’t complain about anything. Even WiFi worked tremendously well on saturday. I still think due to the size of the ever growing interest in this conference that the ULB as location should seriously be reconsidered though. The special service transport on sunday to the main station is a great idea, folks — thanks a lot! Funnily enough, half of the ICE that took me to/from Frankfurt/Main to Brussels Zuid was filled with hackers, btw.

I have met some interesting people working on mobile devices, such as dcordes, leviathan, GNUtoo, cr2, larsc, heinervdm, etc. It’s great to see there is still momentum in real mobile FOSS architectures (i.e. something besides the Android, Maemo, or WebOS systems). I’m glad to tell you that this year we will see an exciting breakthrough in freesmartphone.org middleware supporting new platforms, i.e. progress on the HTC Dream and the Palm Pre is looking _very_ well. Stay tuned for more details appearing here soon.

I wish every conference would be like that. The only slightly disappointing thing was the cross-buildsystem-session in the embedded room. Just when I was expecting the discussion about the problems and potential collaboration to start, the time for the session was over. Rather than wasting time watching Andy Green telling us that our projects will die soon and we should all start using Fedora/Embedded now, we could have had some progress… Oh well, perhaps next year.

I’m fed up with booting my Linux-based smartphones like desktop-systems. Two major developments will help me accomplish enormous improvements in boot speed:

• devtmpfs — kernel support for the /dev file system
• dbus system activation — on-demand launching of dbus-based services

I’m going to carry out the following two tasks in OE:

1. Writing fso-boot, a small executable written in C, which mounts the filesystems, brings up DBus and (optionally) launches X11
2. Setting fso-boot as new init process, that way you still have sysvinit and udev in your root file system, but they’re not active unless explicitly asked for

I’ll do that for the freesmartphone.org adaptation for the HTC Dream (T-Mobile G1, Google ADP-1), which I’m running on 2.6.32 (necessary for devtmpfs) — stay tuned for the first benchmarks.

I bought the HTC Touch Pro some months ago in order to port the freesmartphone.org middleware to it and help to raise an end-user distro. While things began very optimistic (i.e. the modem support was completed after just a few weeks), it came to a relative halt pretty soonish afterwards — because of missing kernel support. Google releasing the kernel source code for the HTC Dream has enabled the HTClinux folks to quickly come up with some very impressive results, but due to the heavy differences in the baseband firmware it did not spare them from carrying out an amazing pile of reverse engineeering — just like every other anti-vendor-port.

While these guys truly have done great work, I personally think it’s not going anywhere soon — at least not to a point where we have an open GNU/Linux on competitive hardware fully supporting all peripherals of the device. Showstoppers are always Bluetooth, Wifi, Sound, Suspend/Resume, and all the other things where Google didn’t care about standard mainline interfaces, but rather decided to put the meat into userland — stowed away behind a “safe” closed source license.

Anyways, this hope/dream/experiment has ended, hence I’m offering to donate my HTC Touch Pro (raph100) to one of the HTClinux kernel hackers as a last act of supporting this anti-vendor-port. If you think you are qualified, drop me a mail. Meanwhile, I’ll continue supporting the Openmoko devices and concentrate on the Palm Pre.

I just came back from the annual OpenEmbedded Developer Meeting (OEDEM) which happened to be in Cambridge, UK. It was a very productive meeting and we agreed on some important things to move OpenEmbedded forward as a whole. Please see the mailing lists for meeting minutes and summaries. We also elected a new board for the e.V. and despite the grief that led to me leaving the OE core team (which subsequently lead to the dissolving of it), I have volunteered (and been reelected) to serve a 2nd year as board member.

As written in a previous installment of this column, I have dedicated the lion’s share of 2009 to the reimplementation of the freesmartphone.org APIs in Vala. Please see the wiki for architectural details, as I don’t want to repeat this here. This is an overview of the current status:

fsousaged

fsousaged has been fully completed and is being used for quite a while now in distributions. All of the plugins are working:

• dbus_service: Implementation of resource handling as per org.freesmartphone.Usage.
• lowlevel_kernel26: Low level suspend/resume handling for Linux 2.6.
• lowlevel_openmoko: Low level suspend/resume handling for Openmoko Smartphones GTA01/GTA02.

fsodeviced

fsodeviced has been fully completed, but is not yet being used in any distributions. All of the plugins are working:

• accelerometer: generic accelerometer handling, needs one of the device-specific accelerometer plugins.
• accelerometer_lis302: lis 302 accelerometer support.
• alsa_audio: alsa audio PCM output and routing (scenario) support.
• kernel_idle: system idle notifications.
• kernel_input: system input handling.
• kernel_info: kernel information.
• kernel26_display: display class-device based brightness control.
• kernel26_rtc: realtime clock, wakeup alarm.
• kernel26_leds: LED class-device based brightness control.
• kernel26_powersupply: peripheral power supply control.
• openmoko_powercontrol: device-specific power supply controls for Openmoko devices.
• thinkpad_powercontrol: device-specific power supply controls for IBM Thinkpad devices.

fsotimed

fsotimed is about half-way complete compared to frameworkd. The working plugin is:

• alarm: DBus alarm service as per org.freesmartphone.Time.Alarm.

fsonetworkd

fsonetwork is done with the same level of functionality as in frameworkd. The working plugin is:

• sharing: internet connection sharing.

fsogsmd

fsogsmd has been on hold since end of April due to waiting for more Vala language features. When they finally appeared in September, I picked up where I left and furiosly worked on what i perceive as the prime subsystem of FSO

The basic infrastructure is more or less complete now and we cover about 50% of the DBus API as per org.freesmartphone.GSM.*, i.e. device info, sim access, network registration, sms, and call handling is working. All work has been done in a generic way, i.e. without taking any care of modem specifics yet — which is what will be my next task before I go on covering the missing API.

fsogpsd

I have added a skeleton of that to the repository and adapted some lower-level classes in libfsotransport to work both for fsogsmd and fsogpsd. I would have done more work, but I’m not keen on implementing the Gypsy API, since I think it’s not a particular good DBus API

fsopreferencesd / fsopimd / fsoeventsd

All these have not been started, not even been thinking much about ‘em. fsopreferencesd will probably have to wait until dconf / gvariant / gsettings have finally landed in glib. fsopimd is waiting for a redesign of the opimd API. fsoeventsd needs a new architecture, but I have to discuss this with the others before we can start cranking.

2010

will be a very interesting year for Linux on mobile devices, even more so for freesmartphone.org. Due to the lack of someone funding FSO, I will probably not find much time to work on FSO in 2010 — that’s why I’m so furiously working on getting most of it to a state where others can jump in before the end of this year.

Apart from that, I hope we can get FSOSHRCON’10 happening very early in 2010 and uplevel kernel support for some of the more interesting semi-open devices such as the Palm Pre, Nokia N900, and the HTC family. FSO would be more than happy to add device-specific support for this hardware once the kernel is up to par.

Cheers!

As mentioned, the freesmartphone.org team and community has taken the challenge to put the FSO stack on the Palm Pre which is out next month. The goal is to manage a voice call with the FSO stack within four weeks.

The idea behind this is a very important one. With only the Openmoko FreeRunner as a platform, the FSO stack is doomed into oblivion sooner or later, since its a very limited hardware platform — in quantity, but considering the closed alternatives also in quality. Hence, we need to proof that FSO can run on current, competitive hardware — to embrace companies that want to adopt FSO in their niche.

The Palm Pre is currently our major hope — all other hardware being either too closed (yes, this includes the Nokia N900) or already outdated.

We just released the following statement to various mailing lists:

Braunschweig, Germany, 2009-07-29. For immediate release.

The freesmartphone.org core-team founds a BGB company to facilitate the further development of free and open source middleware for Linux-based mobile systems: “Lauer, Lübbe, Schmidt, Willmann, freesmartphone.org GbR”.

The core-team members of the freesmartphone.org project today announced the founding of a legal entity offering consulting, training, and implementation services around the freesmartphone.org middleware platform, also known as FSO.

“We now have a single point of contact for both commercial and non-commercial parties who want to use our services to create compelling solutions. This is of interest for groups or individuals creating new devices or freeing existing devices (”anti-vendor-ports”) and who decided to incorporate the FSO middleware”, says Dr. Michael Lauer, founder of the FSO project. “If you care about the further development of this platform or if you need guidance for tailoring or customizing the FSO middleware, contact us via E-Mail at coreteam@freesmartphone.org”.

With todays’ smartphones evolving into ubiquituous companions, a gap has emerged between widely used FOSS components like the Linux kernel and core system libraries on one side, and end-user applications on the other side. The lack of a complete free mobile software stack hinders innovation and leads to reinventing proprietary solutions for services middleware.

FSO’s mission is to close this gap by designing and developing solid middleware for mobile systems in an open fashion; this refers to not only publishing source code under open source licenses, but also to sharing the whole design and development process with the community and giving both commercial and non-commercial entities a way to co-drive and steer the process.

Built on top of the Linux kernel, FSO implements high level services for mobile application development, accessible via the DBus interprocess communication standard. Leveraging the FSO APIs allows the developer to concentrate on solving application domain problems, such as business logic and presentation of data, without having to worry about the device specifics and low level details, such as how to access resources, telephony, location awareness, data storage, etc.

About freesmartphone.org: Previously funded by Openmoko Inc, freesmartphone.org is a collaboration platform for open source and open discussion software projects working on interoperability and shared technology for Linux-based smartphones. freesmartphone.org operates on the services layer (middleware) and offers APIs and reference implementations that support modern interconnected mobile devices. To provide reference solutions, freesmartphone.org works closely together with various device-specific communities such as the Openmoko, OpenEZX, and HTC-Linux groups. The FSO team honours and bases on specifications and software created by the freedesktop.org community.

This means you can hire us (or donate money), if you want to support the FSO middleware development.

After the sudden death of my Dreambox 7025, the new OE-based device in the living room is a Dreambox 8000 — simply the best set top box money can buy these days. Yes, it’s quite expensive, but the hardware is fully loaded (heck, there’s even WiFi) and the freedom to install what you want is invaluable.

I’m on my way to LinuxTag 2009. Instead of a “real booth” like last year, we settled on a developer table in the hacking area — there we can present our Linux on mobile projects such as

• GPE
• FSO
• OpenEZX
• OpenMoko
• Gnufiish

in a more relaxed way — giving room to dive into some technical issues, when interested folks come around.

Find me there, if you’re interested in any of the aforementioned projects. I’ll be there until Friday afternoon.