I strongly disagree. CC-BY and CC-BY-SA are extremely useful if someone wants to cover your work (e.g., figures) in a textbook or review article and needs to make some editorial changes. Extremely annoying for everyone if non-free licenses are being used and a lot of paperwork has to be done. Same goes for the case when people deem your work so important or interesting that they want to put it into Wikipedia. Great for the scientists, but a real PITA if the license of the paper is incompatible with the one used by Wikipedia.

Contrary to your claims, the seminal paper on decoherence by Joos and Zeh is from 1985, so this even predates Weinberg's nonlinear QM paper by several years. I'm not claiming that we understand everything about decoherence and the quantum-to-classical transition, but it is extremely unlikely that the gaps in our knowledge can be filled by looking at exotic extensions to QM not backed by any experimental findings, no matter how mathematically appealing they might look in the first place.

1. Last May, this guy announced he would GPL his stuff once he gets $4,000 in monthly donations.
2. Eight days later, he received a total of $4,000 in one-time donations and released his code under the GPL.
3. About a month later, he discovered that one-time donations and recurring donations are not the same thing.
4. Apparently until today, he is whining around how bad this all is and that open source is evil.

Start with any introductory material on the theory of decoherence, such as the book by Joos and Zeh.

There is nothing about the correspondence principle that needs to be resolved. This is hardly noteworthy science.

I like Wikipedia-bashing as much as everyone else, but in this case there is hardly anything to complain about. The Weinberg paper talks about nonlinear extensions to QM, which are widely believed to be nonexistent. So this guy found a statement on an obscure theory almost everyone believes to be wrong anyway, and you expect this to be notable enough for a Wikipedia article?

Semiclassical gravity, i.e., couple the metric to the expectation values of the energy-momentum tensor. Granted, it's not pretty, but it contains all the physics we know and is not refuted by a single experiment.

Apparently, the WePad was doing just that, no idea how well it worked though.

Access to the ~47,000 packages, including thousands of libraries that you do not have to painstakingly port to Android (good luck with that on a non-POSIX system)?

Wrong.

No. Any system that can be botched more or less accidentally is a complete failure. While GNU/Linux and to a lesser extent OS X are far from perfect, they make it considerably harder to run untrusted code, simply because this is an operation typically not needed during daily use.

I think you have a hard time proving in court that your product is not a derivative work when it's just a piece of binary waste once you take away the GPLed portions.

$ cat /proc/sys/kernel/tainted
0

Not very much, apparently.