All of those projects which you listed are more or less programming or web-content related - for which the domain knowledge is readily available among the people who enjoy working in open source. One needs domain experts in the particular domain to implement successful software.
Domain experts are motivated by money, peer reputation, and improvement in ones domain. Two of these are readily available in FOSS projects, and becoming a developer in a reputable FOSS project also improves ones career prospects financially quite considerably.
So, for a programmer, participating in a FOSS project is quite appealing.
Now take some other domain, like music. First, if ones peers are likely to be more or less ignorant of software development, they really cannot appreciate the effort an expert places on developing a software. They may find it interesting, but not as admirable as progressing in ones own domain. Secondly, while developing software may give insights in ones problem area, it's not directly related to her craft. Take a hypothetical composer - which is a more appreciated thing, him doing a composition, or testing/speccing some new compositional software. Thirdly, domain experts seldom gain any obvious financial benefit or participating in open source software (I'm talking about the dominating dynamics, not special cases).
The fact that it's neither of those would suggest that people don't want it open badly enough. Listing off projects that are open to show that people like open things is absolutely irrelevant.
You're forgetting the third and much more realistic option: that people will come together and create a substitute that far surpasses the original, because of the naturally collaborative nature of free software.
Wolfram the computational engine is a hosted solution, so there's very little point to "opening" that, business continuity reasons aside.
Mathematica I don't know much about, but afaik serious engineers, mathematicians and physicists use MATLAB generally. There's more of a shift to using R and Julia nowadays but academics are notoriously stubborn when it comes to using software, regardless of its quality.
That seems a bit like saying that, if people demanded that North Korea were open, then it would be either (a) open or (b) gone. Of course, the lock-in to a software package is less than the lock-in to a country, but at a certain point it is not totally absurd to say that it is effectively impossible to switch.
I think your point is missing the point. The market of people who just want a solution to work and be easy to use is much, much larger than the market that explicitly demands BSD or GPL-style freedom. Wolfram serves the former. Makes more money that way, too.
People don't "like open things" because of marketing hype, they just tend use the most useful things. Collaboration in programming is a feature. If no other projects on github are written in "wolfram language", then learning to program doesn't really enable you to do as much.