Plus some of the best, most interesting questions on there are "community wikis" which often tell you they should have been flagged as opinion based but are preserved for historical significance.
Because of that I really don't see the point in being so dogmatic about closing opinion based questions. Even something like "emacs or vim" I can imaging being a useful question, with a highly upvoted, thought provoking answer arguing each side.
The accepted tick on the answer says that "here's how" might have worked for the question author.
The most upvoted/popular answer works as a rule (there are exceptions). Popularity contest is not the best way to get the truth on a technical subject but it works more often than not (you can read the answer in the chronological order if you like).
The problem with this is that there are opinions and preferences and biases (and so on) hiding all throughout even the most (seemingly) objective questions and answers. The SO folks think they're making a more meaningful distinction here than they actually are.
Things like "lib A requires excessive boilerplate compared to lib B, see example", or "lib A can't be installed under the following platforms", etc. What should be voted down on those types of questions isn't the question, but poor answers.
It's really frustrating, because getting the opinion of experts who have tried a bunch of alternatives, and know the pros and cons that matter, is exactly why I go to SO frequently.