Port linux kernel to a PDA device (2005)
Linux is gaining more and more respect in the world of "big iron" but it can't really compete in the area of embedded devices. Although the 2.6 kernel series is ready for embedded usage, no real (widely used) implementations have been created for the handheld computers (PDA). There are some minor successes but PocketPC systems (MS Windows based) are still far superior to the Linux alternatives. PalmOS is quite a different thing but it's still more popular than linux. The userland applications themselves are usually written to be architecture independent so the biggest barrier is the kernel itself. Hardware specifications are known so the only thing that stands in the way is the lack of developers (such developer has to be given a PDA to do the testing).
I do have skills in programming embedded devices and I do own one of those PDAs so if I get accepted, I will join the effort at handhelds.org to help finish the kernel support. Once we manage to fix kernel-level bugs (at least the most important ones), the door is open for the GUI developers (these don't need to own the device) to show their best. That should bring new (and much needed) alternatives to the world of PDAs.
So, to be more specific, I will do the following:
- Ensure the kernel itself runs on the h5400 handheld series (the one I own; pretty popular)
- Write device drivers for peripherals such as bluetooth, wi-fi, LCD, etc. either by porting existing implementations or creating new ones
- If time permits, write even more drivers for devices which I (unfortunately) can't test but other users will benefit
- Prepare the code to be accepted to the mainline kernel by Linus
Support and improve handhelds.org build infrastructure (OpenEmbedded) as a lower-priority task
It's quite tricky to predict anything when it comes to kernel development but looking at the current state of things, I believe it is achievable to have a running 2.6 kernel (with drivers) ready for a public release (day usage by non-developer users) by the end of summer. After talking to the current developers at handhelds.org, they agree this plan is realistic and they would appreciate my help in this area. Even if the deadline is not met, any major progress is a success by itself.
The waters of handheld computer are still ruled by proprietary technologies. Help us bring the freedom..