This question is solely about (editorial) style.
Consider the following (made up) examples:
- Title: Is foo safe?
In the body: Is foo unsafe? - Title: Can baz affect qux?
In the body: Can I assume qux is constant even if baz happens?
When you see an answer starting with "yes" or "no", you don't know if it answers the title or the body, until you read more.
An actual example is this question (revision 1) on Unix & Linux SE. The title is:
do device numbers persist after reboot?
Then there are details and an explicit question in the body. Personally I totally support the concept of asking the actual question in the body, even if it's in the title; and I'm totally against treating the title as the question (example discussion: here). In my opinion any post should be like a neat article. In an article that asks a question, explicitly asking the question at the end is something that makes it neat.
The linked question does explicitly ask at the end. The last sentence is:
In other words, can the device number for a partition change […]?
The question in the title and the explicit question in the body are opposite to each other. When I'm writing this, there are two answers published almost simultaneously. Answer A (revision 1) explicitly says (starts with) "yes", answer B (revision 1) explicitly says (starts with) "no"; as if they disagreed with each other.
In fact they agree with each other. You need to read more to see "yes" answers the body and "no" answers the title.
I note the problem in the question is decently clear; each answer is also decently clear. If only you are able to understand the technicalities then you will understand what each person wanted to say.
The issues:
If you read the question from top (the title) to bottom (the explicit question) and then you proceed to any answer, you expect the answer to address the question you have just read in the body. The answer that answers the title feels it doesn't belong, there's a dissonance.
It's worse if you're a user or a stranger with a similar question and at first you want to know if the thread looks promising. You read the title, you skim through the question body only to make sure the question fits your problem, you probably miss the question in the body, you concentrate on answers. And then:
- any answer that clearly states "yes" or "no" catches your eye because you hope to find a definitive answer quickly;
- but you notice one answer says "yes", the other says "no"; you think they disagree (especially if the answers are relatively short, so you can see them one above the other on your screen);
- and if they are both upvoted, it feels strange; even more strange if they are roughly equally upvoted;
- then you start reading and the answer that answers the explicit question feels it doesn't belong, because you expected it to answer the title;
- so you go back to the question body, discover the explicit question and everything starts to make sense.
One way or another, almost always you read any answer with some question in mind. If the answer answers the opposite question then a dissonance will appear.
Don't get me wrong. I think it's the reader's responsibility to understand the question and the answers before he or she applies any of them to his or her specific problem. Still if contributors (writers and editors) are able to make it easier and get rid of possible dissonance then I think they should.
My personal opinion is in principle we should edit for consistency. On the other hand I'm not eager to edit such questions (yet) because:
In general the question and some answer(s) need to be edited, so the thread becomes consistent as a whole.
An edit that changes "is Q false?" to "is Q true?" (or vice versa) or an edit that changes "yes, Q is false" to "no, Q is not true" (or similarly) may be considered trivial.
An edit that changes "yes, Q is false" to "no, Q is not true" (or similarly) may seem to contradict the answerer's intent ("yes, …" changed to "no, …"). It would be an alleged contradiction, not actual; still some misunderstandings may arise.
Because of (1), if (2) and/or (3) result in some rejection(s), rollback(s) or edit war(s) then the whole operation will fail. If it fails then it will possibly do more harm than good. Maybe it's not worth the hassle. OTOH if we agree (in answers below) it is worth the hassle then anyone will be able to link to this very thread while editing, exactly to explain the problem and mitigate the risk of (2) or (3).
If the title and the body are opposite questions, should we leave them be? edit for consistency?
