I wish people would take the time to realise that the Linux kernel would not be anywhere near as reliable or widely used if Torvalds had used a more permissive license. The GPL means that all users of the software have freedom, no matter who packages the software. As a result, you end up having a single point of reference for Linux rather than 1000 proprietary forks that are all equally shit.
... although there are still proprietary forks, especially in embedded systems, because of limited resources (and apparently sometimes appetite) for GPL enforcement.
(Conservancy, one of a tiny number of organizations that does this work, has been having a lot of financial troubles lately: https://sfconservancy.org/)
> ... although there are still proprietary forks, especially in embedded systems, because of limited resources (and apparently sometimes appetite) for GPL enforcement.
As you said the software freedom conservancy is having financial troubles (why the Linux foundation doesn't just cut them a cheque is beyond me). However, a lot of embedded systems use BSD and other such systems. And ultimately, that's not the point I was making. The point is that the GPL not allowing proprietary forks is a benefit (both from a freedom aspect, but also from a technical aspect) and has given us all much better technologies as a result. The "freedom" to make proprietary software (which is a misuse of the term) is a tiny "benefit" (I don't think it is a benefit) when compared to the end results of GPL software like Linux.
I agree with you about this, just noting that the benefits aren't as automatic or as universal as we might expect due to compliance problems. But it seems that more often than not it works as designed.
(I guess another question is to what extent there are people actively scouting for useful stuff to merge upstream; are there kernel trees with useful fixes or enhancements that are published never merged into the mainline kernel because nobody notices?)
