And yet, their O/S remains dominant. As far as a graphics standard, would MS be any more likely to give up in favor of OpenGL than the supporters of Beta tapes were to give up to VHS. The latter fights over, but the current one continues -- and whether it's a bad decision or not, don't ever underestimate the ability of MS to come out on top, regardless of whether their product is better or not.
No question. Remember that other than CPM, which died a quiet death, there were no real competitors for DOS, then only expensive Apple machines versus Windows. Junk, like counterfeit currency, persists if there is enough of it around.
Linux, inspired by UNIX and Minix, was very much late out the gate (or should I say Gates?). Remember that for a long time you needed a fairly expensive machine to run UNIX, and the fore-runner (Multics) needs a huge main frame. UNIX would never have happened if Bell Labs hadn't pulled out of Project MAC at MIT. The guys at the Labs couldn't stand not having the Desk Side Computer Service, so: Invented C; Invented TLA; wrote UNIX as a replacement for Multics; and ran it in the skunk works for quite a while before letting this cat out of the bag. I am not sure that this same group produced the initial versions of YACC and LEX to help get the C compiler going.
In light of hardware developments these days, I wonder just how hard it would be to add the associative memories to the processors to produce a microcomputer that could run Multics? Of course, it is a time-sharing system, so wouldn't be much in use now unless there was a way to operate it as a single workstation. Somehow I think that if it were cheap enough, the U.S. DOD would be delighted. The whole source (PL/1) is public domain.