I'm stepping into a hornet's nest here.
Popularity contests have been controversial. Some detractors believe that non-objective winning criteria are simply unacceptable. Others are OK with subjectivity in theory, but dislike that voters upvote answers based on coolness rather than programming merit or following the spec. My proposal is aimed at the second group.
I propose a new winning criterion voter-judged. The highest-voted answer wins, like pop cons, but the question spec must explicitly give (subjective) criteria for voters to decide what to upvote or downvote. There are many things that humans can judge much better than a hard algorithmic spec, like looking like a forest.
Will voters actually follow it? I don't know. But, I hope that explicit instructors will sway some voters and make it clear what a good answer is expected to be, a source of endless controversy in discussions to close.
The problem with the name popularity-contest is that it's a mildly disparaging phrase that evokes high-school elections and voting in a bandwagon on appearances rather than merit. Hence my proposal for a new tag that frames voters as judges. The tag popularity-contest would be reserved for questions where voters are in fact expected to vote for whatever they fancy.