I failed a review audit for Cleanest way to report errors in Haskell. I responded by voting to close it as "primarily opinion based." I understand that answering an occasional review incorrectly will not hurt, that is not my concern.
Am I out of line for believing that question should be closed as being primarily opinion-based, despite the seven upvotes? The asker outlines three different ways to deal with errors, and basically asks "what would you do?"
No, you are not out of line, but I would respectfully disagree with you.
Yes, the asker outlines three possibilities, but that doesn't immediately make the question opinion based.
The question is almost equivalent to asking "Should I use return values, errno or longjmp to report an error in C?" and although multiple possibilities are mentioned in the question, the consensus among the C programmers leans strongly to one side (return values).
This is what, in my view, sets these questions apart from the opinion based ones: all subject matter experts will give an answer that points in the same direction.
The really opinion based questions are the ones where two experts will give you three opposite answers or where expertise in the subject matter isn't relevant to the question.
