If AGPLv3 was slapped onto everything back then, the likelihood of linux/open source being where it is today would be an order of magnitude less. A good chunk of the original windows TCP/IP stack was ripped from BSD licensed code. Had that not have been "easy" for Microsoft to take, the internet may not have developed the way it did and we'd all maybe be on proprietary networks like AOL/MSN/etc.
The solution isn't always swinging super-far in the other direction.
That being said, commercially supported OS software has essentially become shareware - just enough to get you hooked, and then the price jump is enormous.
> The solution isn't always swinging super-far in the other direction.
I'm not convinced. Compromise is the root of all evil. It allows self-contradictory and mutually exclusive ideas to exist.
I make a conscious effort to develop my morality by eliminating contradictory ideas and forming self-consistent world views. If the conclusions are extreme, then so be it. I accept the consequences.
Before Microsoft added TCP/IP to Windows, there were third party ports of the BSD TCP/IP stack to Windows. The winner ended up being Trumpet Winsock, until Microsoft wisely used the same winsock API for their own implementation and embraced and extinguished.
Those ports would still have occurred even if the BSD TCP/IP stack was GPL'd.
The alternative history here would not have been that Wrumpet Winsock didn't get crushed by MS but that MS used another more proprietary technology to do it. They have done it in many other areas, e.g. DirectX vs. OpenGL. You cannot hope to compete against the platform you build your product on.
IIRC, Microsoft wrapped a good chunk of the BSD code with it's native implementation of the Winsock API (I may be misremembering and the BSD code may have only been in the NT or win32 code - it's been decades and I'm pretty sure the whole stack has been rewritten since).
But if there wasn't a bunch of people pushing various implementations based on a permissive licence, it may not have been clear that the demand was there. Winsock (the API) heavily leveraged BSD sockets in its implementation, too. V1 was for all intents and purposes designed to deal with unifying their BSD implementations that had problematic issues because Windows didn't have a lot of the system calls that UNIX/POSIX did (though POSIX was available later on in NT) and allowing inter compatibility. It's not clear that Peter Tattam would have written Trumpet if the API wasn't already published, which was driven by the BSD code.
Also, Trumpet was not free or open source. It was notoriously pirated and illegally distributed.
The solution isn't always swinging super-far in the other direction.
That being said, commercially supported OS software has essentially become shareware - just enough to get you hooked, and then the price jump is enormous.