On first read, I was going to ask to get a link to that thesis, thinking it was some Masters or PhD work. Every 10 years you're going to have the 'big-evil'. I'm with you 100% about it being standard practices in big companies, and 110% about users being complacent about accepting the walled garden.
Also re: patents- IBM has a huge patent portfolio. Very little of it is used to patent troll. Those boxed-office Texas shell LLP law-firms setup are nothing like what IBM is doing today (second only behind Red Hat in terms of core Linux kernel commits, IIRC[caveat: source was from a talk I attended in 2013 by one of the 3 upstream-controllers, including Linus]). "Evil 80s IBM" is now actually too stagnant to really exploit the consumer and barely breathing off of old banks and gov't systems that run System/36,360,38,390 on z/OS.
"Evil 90s MS" is now opening up basically everything for the consumer, minus the cash-cows (enterprises). MS Research funds tons of Haskell research, language development that gradually goes from PhD/post-doc work -> "obscure" languages like Haskell/F# -> C# (LINQ is effectively a monad). Python 3.5's new async/await is largely influenced by work Microsoft did (both in semantics and syntax). The CoreCLR is on Github. Inspect, modify, alter, run your F# code on a RS/6000 or 8 dollar ARM if you want to. Visual Studio Community Edition is free, a la the IntelliJ model. I argue they're now a net positive on the community. Just like IBM, they've realized the enterprise is where their money is, so charge for SQL Server licenses (which are really reasonably priced compared to Oracle), Sharepoint, Server, and Dynamics/Biztalk type stuff.
In the 80s it was IBM. In the 90s it was MS. Today, Apple performing so much walled-gardened anti-competitive practices (I can't install my legally purchased copy of OS X on my Intel machine? Really?) that it makes the MS DOJ trial look completely insignificant.