No, Apple Is Not Trying to Patent LLVM. That's a Lie.

There's been some nonsense being floated around on identi.ca (as well as less savory locales) the past week, to the effect that "Apple has patented LLVM". It turns out to be based, as near as I can tell, on someone's having done a simple string search through Apple's patents for the string LLVM, finding two instances, assuming the worst, and pouring out their darkest, most ignorant, most ill-informed fears onto the "End Software Patents" site, which Ciaran Riordan runs, without ever actually having read the patents (actually one patent, one application) that had gotten them so upset.

This was more or less the state in which I found the Apple page, a couple of days ago, over there. Ciaran and I have been discussing the matter for the past couple of days, pretty constructively for the most part, while I corrected the page in various ways. It's in a state at this point where we can both tolerate it, at least.

If anyone tries to tell you that "Apple has patented LLVM", it's a load of FUD and nonsense. All the details are over here. I've summarized the five biggest arguments the opponents of this situation—which turns out to involve a single patent application which actually has nothing to do with LLVM per se—have brought forward as to why this particular patent application is so problematical in their view here.