M$ lost in the SCOTUS because they took a devious route to escaping the i4i lawsuit. Instead of trying to invalidate the patent on grounds of obviousness and prior art, they fell back on something they could not prove, that i4i had sold the product more than a year before the patent was applied for in the USA. The programme M$ suspected had the patented software no longer exists in source code so they could not prove that directly.
M$ did not want to invalidate all software patents, including their own (thousands) so they have to pay the piper. The concept of keeping data and metadata separate has been around for ages in IT and M$ could have gone after that ground for prior art but it would have defeated a bunch of their own patents. M$ could have taken the obvious path that all software is obvious given an instruction set or programming language and basic elements of IT: storage, logic and operations. Give a programmer a goal and he will write a programme to accomplish it without knowing anything about the “prior art”. That’s computer science 101. Given some prior art related to the art, the programmer will do the job sooner. How can the resulting programme be patentable?
So, M$ pays a huge settlement and the whole world of software remains burdened with frivolous lawsuits about software patents until the SCOTUS gets a case where they cannot wriggle out of it like they did for Bilski.