If you do software development for profit it is very likely that you work on software you never really use. Isn't this a scary thing? I man how should you know what works and what not? Are you finished when you completed 100% of the spec?

In the free software world on the other hand developers are often driven by scratching own itches. Fixing things they found bad or broken by using the software.

Looking at my desk I can see 7 linux based mobile phones around. Most of the time I used non-Neos as my phone for the daily use. This does not make me better then others devs not using the code they produce. Holger does it, Daniel does it and many colleagues in Taiwan does it. It was time to feel the pain yourself.

Over a week now I'm using my Freerunner as primary phone. Loaded with the latest and hottest on our way to FSO milestone 4. It is an interesting journey. Showing me how much I already miss compared to other phones. Also showing me how much we need to care about stabilize better for our milestones. MS3 was bad in this regard. We hope to do better with MS4.

During this journey I started to write down what annoys me. Annoyes me as a user. Besides all kind of smaller and bigger issues I had to learn that ZenPhone is really not a UI for the daily use. It is a testing apps to expose our interfaces, not more. Really looking forward to the UIs the SHR guys and raster are working on. Perhaps even a nice paroli. We will see.

What I found is a 2.4 driver and a small patch to port it over to 2.6, some small work from my side to catch up with 2.6.13 pcmcia API changes. That's it. The last time I checked it was recognized, not more. I'm pretty sure it needs some real love to get in shape for mainline, but it could be a nice coding project even if the docs are not available as far as I know. At least some code as reference.

I already send a mail to the linux driver project, but gathered no feedback so far. If any of my readers like to give it a try or know somebody who likes, just drop me a mail and I'll send the device over.

The reason for being there was the TechWeek from Pengutronix, a company from my area doing a lot linux embedded projects for the industry. I already known some of the people working there privately. While being there I got known to the other ones. I must admit that it is a nice bunch of smart people loving what they are doing. What I actually appreciate a lot is their work to get their patches into mainline, even if it costs a lot of time and money. This is a not-so-common practice in the industry linux embedded world.

While hanging out there and having good talks about git, patch handling and submission workflows I spend most of my time working on geting some of the EZX patches mainline ready. We now have a svn branch that contains patches sitting directly on top of the arm git tree pxa branch. While working on this I also started to submit three one-line fixes upstream to get used to the arm-linux workflow. 2 Are already in the git tree, one is acked and waiting in incoming.

I enjoyed the week. Smart people, good food and hacking on stuff you like. Life could be that easy...

The great guys from the OLPC Deutschland e.V. long-term borrowed up to 70 XO laptops for people which have interesting projects with them.

I asked for one to work on OpenEmbedded integration for it. This divides into two parts:

1. Machine support. This one should not be to hard as it is a x86 device and the first rootfs it ever booted in the AMD labs was build with OE. :) Still it will help me to understand the deeper internals from OpenEmbedded better.

2. Writing recipes for the sugar applications and libraries. Having recipes for them in the OE metadata will make it easy for other distros using OE as their buildsystem to use them and put them in their feeds.

I will be busy with EZX the next two weeks. I plan to work on this afterwards.

Still I'm looking forward to it. This time I hopefully have some time to attend the technically talks. I look at you kernel track. And let Harald de-mystify the security of the micro waves around us.

http://svn.datenfreihafen.org/$PROJECT_NAME

http://git.datenfreihafen.org/

The first one was that I got an 18bpp patch for all the second generation devices working. At least pxafb and fbcon are working fine now. I still need to test X more. :) The patch was from the gumstix patchset. Thank you guys.

The second was the boot_usb 0.2.0 release. We use this little tool a lot and SVN is stable most of the time. Especially after Daniel Ribeiro added support for initrd, commandline and setting the machine ID a release was needed.

I switched to gnome because I liked to have a more integrated desktop. Having this smooth experience that all stuff just works. It turned out that it did not helped as much as I wanted it.

I don't use Evolution, I use mutt. I don't use the gnome-terminal, I use xterms. I think you get the picture. After looking more closely what stuff I really use from my gnome desktop there was not much left. Some app starter in the panel, 2-3 applets, well I think that's it.

As I have worked with the Enlightenment Foundation Libraries lately and I liked what I found I decided to give E17 a go. Fresh from CVS.

I now have just a small shelf, panel thingy, with 6 app starters, volume, time, date and battery gadgets / applets. Next to it a small stalonetray for systray icons. That's all.

I've won faster startup times, some small but nice eye-candy stuff. Not to much to distract me from work but enough to keep the joy of use.

Another big plus is the brand new tiling module. It makes window placement more efficient as it allows you to define how windows are placed on different virtual desktops. No more overlapping windows you have to place right yourself with a mice or similar. The whole space is used, nothing wasted. I use this for my hacking workspaces. Open more and more xterms and don't worry about a sane placement. Great. I wanted something like that since Harald showed me something similar with his ion3 setup on his laptop. :)

Ease the development of new applications and services. Build your kick ass stuff on top of a good fundament. And if it does not give you what you need, extend it. It's not like other commercial frameworks where you have to deal with what you get. It's open, take it, extend it, send patches. :)

Let's hope the framework team get the resources they need for getting it done. I also have some private ideas how to contribute here. Once I have something ready I let you know.

As code is better then words, take a look at their git repos.

I'm not leaving with bad feelings. Still over the last months I disagreed with decisions and directions, got frustrated with the work I had to do. Not the best conditions for a productive work environment.

However I still like to see OpenMoko success. As open as possible mobile phone hardware on which I can flash all kind of software I like. Not fighting against vendor policies but having almost all options I like. Still sounds great to me.

Over the last 13 months I learned a lot. For both, coding and business. Working with some great people from the FOSS community from all over the world and with great taiwanese engineers. (Watch out for these guys, they are awesome but shy) Made new friends and had to deal with an interesting different culture.

It was an awesome but exhausting trip. It's over for now.

I plan to do no paid work for the next 1-2 months, perhaps even longer. In this time-frame I need to come back to the study life I had before and take some rest. Read some good books, hang out with friends and don't worry about deadlines.

After that we will see what kind of work will come in. At least it must be not as much time consuming as OpenMoko was. Need to focus on my studies.