I am aware of On editing questions and invalidating answers but this is different.

In my thread Would low-grade levitation be of any use? I gave basic rules. As time went by I received a flood of questions requesting further details of the physics.

I answered these but in doing so I may have invalidated previous answers. Note this is not because I contradicted myself - it's because the answerer just assumed their own version of the physics without checking with me.

So have I dealt with this correctly? Is there a better way?

  • 1
    $\begingroup$ Traditionally we make comment and roll the question back to a point before the answers were invalidated $\endgroup$ – Separatrix Nov 16 '18 at 11:58
  • 2
    $\begingroup$ @Separatrix Booo to that answer. That is only the way to go if the question is edited to change it. Assumptions that an answer-er makes, without getting clarification in comments, are the responsibility of the answer-er. A question asker cannot be expected to anticipate every assumption someone might make. $\endgroup$ – kingledion Nov 17 '18 at 4:26
  • $\begingroup$ To support @kingledion's perspective, when a question of mine changes to the point of invalidating answers, I post a comment on those answers pointing out the issue and inviting the answerer to update their answer. In some cases, this has resulted in amazing answers as the answerer crafts a good idea to fit the clarifications. Things on this site are only as static as we allow them to be. Nothing stops an answerer from keeping up with changes other than a lack of auto-informing them that a change to the Q has occured. If I recall correctly, an enhancement to this effect has been requested. $\endgroup$ – JBH Nov 23 '18 at 5:56

This, like most things, is a two part problem at least in my opinion.

1) The person posting the question should be clear. They should meet all the requirements from the how to ask page. Namely this bit:

General guidelines for all questions:

  • Must be specific and answerable: What problem are you trying to solve?
  • Must include context: What are you trying to accomplish? Context gives people writing answers an idea of what your end state will look like and why you want to get there.
  • Must include restrictions/requirements: What will make one answer better than another? If any answer is equally effective your question is not properly constrained. How can this be executed? What tech, timeline, magic or other criteria apply to the situation.
  • Should include research: What ideas have you considered, or what information have you already looked at or failed to find?

If you haven't met all these bullet points odds are you are forcing users writing answers to assume/fill in the blanks for you.

2) Those answering have a responsibility as well.

I see this problem pretty regularly. Many posted questions fail to meet the requirements listed above. Instead of asking for clarification in the comments people will rush to answer the question. It is one of the necessary drawbacks of the reputation system. The sooner you get an answer in the more likely you are to get votes and rep.

This is a problem though. If we rush to answer and don't try to help the querent first we aren't doing anyone any good. If the question is unclear and we answer we aren't really helping the person that was looking to get an answer.

The short version is both questioner and those answering need to take it upon themselves to make questions better before answers are posted.

The responsibility isn't on one or the other. It's both.

  • Use comments to help people posting questions
  • Do what's best for the site. In the end this will get you more rep, and definitely more respect than rushing to get imaginary internet points
  • Take the time to write good questions that meet the requirements on the how to ask page.
  • If necessary flag questions to put them on-hold. Explain, especially to new users, what that means.
  • 1
    $\begingroup$ This is certainly the answer that appeals to me most. If only everyone lived by these maxims! I'll certainly try to tighten up my own questions although specifying in great detail all the possible constraints would sometimes require writing a two page essay. I rather wish that people would answer in the spirit of the question rather than nitpicking the minutiae and trying to trip the asker or paint them into a corner. $\endgroup$ – chasly from UK Nov 20 '18 at 23:47
  • $\begingroup$ @chaslyfromUK You make a fair point. There's a fine line between asking for clarification and just being difficult. $\endgroup$ – James Nov 21 '18 at 17:31

When your "clarifications" invalidate answers, then it is unacceptable change in question.

Your edit in such circumstances shouldn't be an edit. It should be a follow-up question. In case it was already made, it should be rolled back.

Things you decide not to write in your question are the freedom you give to people answering it. Removing some of it is a change of game, just as if you'd contradict yourself.

  • 2
    $\begingroup$ The problem is that other people (and even the same people later) asked me questions so that they could answer more fully. Should I refuse to answer those questions? Should I just say, "Invent your own rules "? $\endgroup$ – chasly from UK Nov 16 '18 at 17:37
  • 2
    $\begingroup$ @chaslyfromUK if it would invalidate already posted answer? You should direct them to more detailed follow-up question. $\endgroup$ – Mołot Nov 16 '18 at 17:38
  • 1
    $\begingroup$ This would be my ideal scenario in future, but wouldn't a follow-up question with the same question; same basic rules and a few extras be flagged as a duplicate? $\endgroup$ – chasly from UK Nov 16 '18 at 17:38
  • 1
    $\begingroup$ If you will say explicitly that it's a follow up because you want to exclude some solution found in answers to previous one, then no. And if there is a lot of requests for clarifications, question should be put on hold in the first place - to prevent existence of answers you could possibly invalidate. $\endgroup$ – Mołot Nov 16 '18 at 17:39
  • $\begingroup$ The Answers and Comment-requests for information were very much interleaved. Is it possible to put a Question on hold to prevent answers but at the same time to allow Comment questions? $\endgroup$ – chasly from UK Nov 16 '18 at 17:42
  • 1
    $\begingroup$ @chaslyfromUK yes. That's how On Hold works. And you can vote to put your own question on hold. If you do, probably friendly mod will cast his vote as soon as he sees it, and reopens as soon as you'll flag for moderator attention after all is figured out. $\endgroup$ – Mołot Nov 16 '18 at 18:16

The question is the ground truth

If an answer-er makes assumptions, and those assumptions are wrong, the onus is on the answer-er to fix it.

Example: one of my own top answers. I made assumptions about the question, but those assumptions were wrong because I never bothered to confirm them in comments. When someone else did ask the right questions, my own was invalidated.

That is fine, I posted a comment to the top of my answer indicating that it was wrong, and why.

If someone made bad assumptions about your question, that is their problem. Let me know so I can down-vote that answer for being wrong.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .