I've just finished one exciting week at Samsung LSI (the group inside Samsung responsible for the System-on-a-Chip line like the s3c24xx, s3c64xx as used in Openmoko, TomTom and E-TEN glofiish devices). Specifically, I had one week of close contact with the team there developing the Linux kernel port and drivers for those products.
I do not want to go into too many details here in public, but it was a very productive week, and I think everyone involved is drawing a very positive conclusion. Let's hope the actual results from the now-improved understanding of the Linux and FOSS development model will be of mutual benefit to the community as well as Samsung. I'm also hoping for better and faster integration of Samsung's codebase into mainline Linux.
Parts of this Openmoko-triggered drive can be already seen on git.kernel.org, where Samsung LSI has started to push their current development trees to.
Since this is not the first time that I'm trying to lobby for more understanding of the Linux development process, the benefits of going mainline, etc. - I wonder if this kind of work should not be done in a much more scaled and proactive way by somebody like the Linux Foundation. Especially in the embedded world, there are many companies who are probably very interested, but just don't know where to start or whom to talk to. There just is no (or no easily identifiable) entity catering to their needs - and since they are always busy with developing new products and working on ports for other operating system, the initiative should probably come from the outside.
Right now this field is left to embedded distributors who have their own agenda and are always only representing a small fragment of the Linux stakeholders