Hello Ilya,
the discussion has largely deviated from the original topic, which is Genode's move from GPLv2 to AGPLv3. Please note that the relationship of proprietary software and Genode remains untouched by Genode's license change.
Even though I am little afraid of the discussion getting further derailed, I don't want to leave your remarks unanswered.
I am also sharing Nobody's POV. If proprietary vendors are unable to run their code. the genode has no interest to them. even if it would gain popularity and became most used OS we will not see native Skype or Photoshop.
You are mixing up cause and effect. As I noted in my previous posting, proprietary application vendors target their applications at consumer platforms, they don't create consumer platforms.
Since Genode is not a consumer platform, it is uninteresting for proprietary application vendors. The same is evident in the GNU/Linux world. In contrast to, let's say Android, iOS, or Windows, GNU/Linux is _not_ a consumer platform. So Adobe does not make Photoshop available on GNU/Linux.
Consumer platforms are generally not created by grassroots movements. They don't happen to exist by chance, but they are planned products and marketed as such. Unless a platform vendor pursues the creation of a new consumer platform based in Genode as a product, the discussion about accommodating proprietary software with the open-source Genode system remains moot. It does not solve any problem.
Everyone understands your point and respects your licensing choice. Would it be possible to have the kernel level components stay free software but standard library to become more proprietary friendly.
in Linux for example libc is LGPL not GPL.
This would be both impossible and wrong.
It would be impossible because there is no reasonable distinction between kernel-level components and other components on Genode. Given Genode's flexible architecture, even the attempt to draw a line would be futile. Capturing this attempt in the form of legal terms would result in a convoluted license full of heuristics (and thereby loopholes).
But even if it was possible, it would contradict with the position I stated earlier:
* It would dramatically weaken the effectiveness of our copyleft license because - in principle - *any* functionality could be implemented in the form of proprietary code that solely uses the POSIX API.
* It would benefit no one other than proprietary application vendors. Making such a concession would help them, but neither Genode nor the Free-Software community at large.
* It would undermine the dual-licensing business model of Genode Labs and thereby the funding source of Genode's development.
Still, the world is not black and white. There are certainly cases where proprietary software on top of Genode makes sense. Let me give you two examples:
A consumer platform vendor wants to create a platform based on the open-source Genode system. That is, the base system is AGPLv3 and all drivers and the software stack are open. Still the vendor wants to create a distribution channel for proprietary applications (an app store) or other digital goods (games, movies). In this case, the vendor could seek an exemption from the AGPLv3 for its app-store application in exchange for letting Genode Labs participate in the commercial success of the consumer platform. Under this model, Genode would indirectly benefit from the proprietary application vendors. So an exemption is granted for a tangible benefit for Genode.
As another example, for some software the source code may got lost. So it becomes technically impossible to comply with the AGPLv3. The ability to combine such software (e.g., "abandonware") with Genode should not pose a problem.
That said, both stories are very different. I think it is impossible to cover such a variety of corner cases in generically applicable legal terms like our linking-exception clause. They should instead be considered on a case-by-case basis. If confronted with such cases, I see it as the responsibility of Genode Labs to take even-handed decisions. Frankly speaking - looking at the past 8 years - I don't foresee that happening too much.
Regards Norman