That's what I remember from comp.lang.c -- lawyer-ly obsession with ANSI C, rather than C as it's being used. That is an important distinction which I respect, but they should have had some kind of sister forum that's more practical. They denied that hardware exists, etc.

Also, obsession with ANSI C, analogous to obsession with POSIX shell, is sort of "middlebrow" in the sense that the people who WRITE the spec need to go outside of it to create a new version of it. Good specs are derived from real usage.