Regarding: How to do a Full Outer Join using old style Oracle syntax

  1. For a high-quality question with high-quality answers, we want the answers that are correct for the question as it stands to float to the top. (Right?)

  2. However, according to https://stackoverflow.com/privileges/vote-down: "Use your downvotes whenever you encounter an egregiously sloppy, no-effort-expended post, or an answer that is clearly and perhaps dangerously incorrect."

Which of these concerns is more important?

In this case, the OP asked a question and did not initially give the most important piece of information: that he was on a very old database version (Oracle 8). So he got a number of good and correct answers that are relevant to people on Oracle 9, 10 and 11, which is reasonable because most customers will be on these versions now, and there is very little work on Oracle 8 anymore.

Then, the OP edited his question (as he should) to advise he's on Oracle 8. The answers are no longer correct, in fact are not helpful, because they don't work on Oracle 8 - and they got downvoted (with some complaints to boot).

I think the end result was satisfactory, however, as the answerers did edit their answers to take into account the database version, and some of the downvotes were reversed.

So: can we confirm that it is not desirable to downvote answers even if they are not correct for the answer?

Related: My answer down-voted because it had just (recently) become out of date where it seems the accepted opinion is that downvotes are, in fact, intended to reflect the accuracy of the answer - and not to be reserved only for the worst, sloppy, no-effort-expended posts.

If this is the case, I think the instructions at https://stackoverflow.com/privileges/vote-down are inconsistent with the opinion of the majority on the site.

Disclaimer: I was not a party to this question, merely looking in from the outside. I neither answered nor voted in it.

  • ok, who's the smart aleck who downvoted this? :) Commented Mar 28, 2013 at 6:19
  • 1
    Downvotes on meta are a way to signal disagreement. Commented Mar 28, 2013 at 6:23
  • oh I see. thanks Toon Commented Mar 28, 2013 at 6:25
  • 1
    Strongly related search on the chameleon phenomenon
    – user50049
    Commented Mar 28, 2013 at 7:25
  • 2
    I was involved with the OP and let's just say it was extremely frustrating to have the question change and then the downvotes across the board because the OP was altered so much.
    – Taryn
    Commented Mar 28, 2013 at 10:07
  • I feel your pain. I've had a similar discussion here, and apparently consensus is that the question the OP meant to ask is more important than the one they actually asked. I would probably not downvote an answer that has been invalidated due to question editing unless it had a non zero score. The reason is that you want the answer that solves the OP's question as it stands to be at the top. It sucks, but we've got to roll up our sleeves and deal with it.
    – user200500
    Commented Mar 30, 2013 at 18:34

3 Answers 3


What you encountered is known as a chameleon question - a question that drastically changes scope after answers have been posted. This is an exception that we hope never happens, but it does, and it needs some human attention in order to right. If it's handled correctly, the potential negativity surrounding it can be mostly avoided.

When the author of the question changes something that possibly invalidates the answers they have received, the authors of the answers should be notified by comment as soon as possible to let them know. The scope change in the question should also have been clear, for instance:

Edit: Sorry, I should have indicated that I was working with Oracle 8

As a responsible netizen, this is all stuff the question author should have done. But they don't always do it. That doesn't make them irresponsible, it usually means they didn't understand the ramifications of a major edit.

If the question author didn't notify anyone or make the drastic change clear, the first thing you can do is make sure any recent drastic change is clear and that authors of affected answers receive a comment indicating that their answer is no longer valid due to changes in the question. If the change is really drastic, it may need to be rolled back. Involve a moderator if unsure. Sometimes users that are barred from asking questions edit old questions to ask new ones - if you see that happen, flag immediately.

This provides a rather obvious beacon that the question did drastically change, and the now strange answers under it weren't written by sleep deprived monkeys that snort energy drinks. Generally, people will withhold a down vote if it's obvious that scope creep has recently come into play. That's not always the case, just generally.

In most cases, the authors of the answers will either edit them to be correct once notified, or employ an exit strategy from a chameleon question. After a short amount of time has passed, you'll see answers where someone was interested in making a correction has done so, or decided to just delete and exit the question completely.

If there are remnants left that just didn't get fixed, it's up to you what to do. Is the answer worth saving? Can you edit it to just be correct? While blatantly wrong answer should not be flagged - there's additional circumstances at play here, should a moderator be involved?

Voting is always done at your discretion, and you're perfectly correct to down vote an incorrect answer. However, in exceptional cases like this, discretion becomes the key - if you notice strange circumstances, use your best judgement. Votes are reversible after an edit has been made, but there's nothing wrong with holding off for a little while.

Finally, if changes, rollbacks, and what not have just left the entire thing as a colossal mess, you probably want to flag and involve a moderator. Flag the question, use 'other' as a reason and describe the situation as succinctly as you can.


This is one of the reasons a downvote can be reversed if the post is edited.

Ideally, with up and downvotes there will be an order of good to less optimal answers. And if the question is changed, existing answers can change from excelent to less optimal. This is annoying, but it happens in the real world.

If you get downvote, it is always wise to find out why. And in this case, the system works, because the answers are changed.

By the way, if the question is seriously changed (turned into a different question), it should be rolled back, or flagged.

So, I disagree with the confirmation that downvotes are not desirable. They are a very good tool to mark the lesser answers, and give the OP a chance to improve their posts.

  • Interesting. That's the interpretation I've had, until I searched and found the faq stackoverflow.com/privileges/vote-down. It's also the interpretation that some people had on the question I linked to - but they got battered down for it (one even ended up deleting his CORRECT answer as a result of all the criticism of his downvoting!!!) Commented Mar 28, 2013 at 6:18
  • The description in the FAQ still allows room for incorrect answers. They don't have to be dangerous. Commented Mar 28, 2013 at 6:30
  • you're right Toon, it does allow downvoting "clearly incorrect" answers. So in this case do you think it was fair for people to lambast individuals for downvoting answers (that were clearly incorrect in the light of the edited question)? Commented Mar 28, 2013 at 6:33
  • 1
    Yes, although if they know of the history of the question, an @username comment would have been a more polite way. But the problem is that the edit of the question invalidates their answer and we like future users to see the correct answer if possible. Commented Mar 28, 2013 at 6:37
  • 1
    @JeffreyKemp That answer was indeed correct but we don't know why it was deleted - unless you have some inside information. It could be for the reason you say, it could be because of the criticism for the accusations that user made (and then retracted), it could be because he had a bad day, whatever... Commented Mar 28, 2013 at 13:12

IMO the question deserves some down votes if the OP has added a fondamental detail just hours before submitting the question.

In this case an user was aware of the edit, but gave strategical down votes to other answers, that was in concurrence with his one. This is a very incorrect behavior.

IMO users shouldn't down vote an answer if an edit changed the question in a way that the answer results incorrect. Instead they should give the answerer a reasonable amount of time to edit the answer, after notifying him with a comment. In the case that who down votes was unaware of a major edit, a down vote isn't right but comprehensible, since looking for all edits on a question is a loss of time. Hopefully in that case who down votes will revert it after seeing the edit.


You must log in to answer this question.

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