What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

Programming languages and environments change rapidly, and SO answers may quickly become obsolete and incorrect. Is there a clear policy on dealing with such answers by others, so that users landing on the page from a search engine aren't misled?

Prior reading:

The last bullet above suggests a policy that leads to stale, incorrect answers. I've been its victim[1], and got banned from editing because my edits to old answers radically changed them - which was exactly the point, as the answers had become incorrect.

But please, let's not discuss my specific case; rather, I'm interested in a policy for allowing contributors to edit obsolete answers left by other users, without being banned for changing the answer too much. On Quora, this wouldn't be a problem, as the edit suggestions are transmitted to the original answerer, who's usually happy to amend their answer. On SO, the edits are reviewed by users who, by policy, don't need to have specific niche knowledge of the topic the answer is in. This has resulted in the bans I mentioned above, and stale, incorrect answers, some with upvotes, left untouched.

To pre-empt some potential objections:

  • suggesting to comment instead of editing the answer only solves part of the issue. It's undeniable that many users will stop reading an answer after the first paragraph that gives them the wrong impression about the topic (e.g. "Does X run on platform Y -No, but ..."), and even more users will stop before reading the comments
  • tracking down the original user who answered is often impossible because SO has disabled messaging users
  • giving a new, correct, answer and hoping that it will bubble to the top, is... weak. I've been giving a lot of these "better answers after others gave less satisfactory ones", yet they haven't been voted to the top. I can dig up specific examples, but that's besides the point. Having correct information should not depend on whether the community gathers sufficient users to upvote the correct information above incorrect answers.
share|improve this question
    
Ideally version tags should handle this IMO, but they're heavily misused (not there when relevant, there when not particularly relevant). Two questions asking about two versions of a language with two answers seems perfectly to me –  Ben Brocka Oct 8 '12 at 15:02

2 Answers 2

Editing isn't (or shouldn't be) the prefered way of updating answers, when the update completely changes the answer. If you have a better answer to contribute to a question, then just post an answer. The community will review your answer, and if it's actually better it will soon enough float at or very near the top.

If the question is closed, and you can't post an answer, either try to improve it and get it re-opened (if you honestly feel it should be) or just let it go. A closed question is rarely seen as a source of useful information and it's on its way to deletion, if you absolutely must answer a closed & unfixable question then re-asking a better version of it and answering it would be better than trying to force radical changes to existing answers via edits.

share|improve this answer
    
Why should a method of conveying information (comments vs. edits) prevail against conveying truth? Giving a new, correct, answer and hoping that it will bubble to the top, is... weak. Check my acceptance rate, 33%. It's that low precisely because I've been giving these sort of "better answers after others gave less satisfactory ones", yet they haven't been voted to the top. I can dig up specific examples, but that's besides the point. Having correct information should not depend on whether the community gathers sufficient users to upvote the correct information above incorrect answers. –  Dan Dascalescu Oct 8 '12 at 10:43
2  
@DanDascalescu What does your accept rate has to do with your answer? Your accept rate is the percentage of answers on your questions you've accepted. –  Yannis Oct 8 '12 at 10:44
    
@DanDascalescu Answers prevail over edits when it comes to conveying correct information, because answers are peer reviewed on technical accuracy and piggybacking your answer on someone else's is unacceptable. –  Yannis Oct 8 '12 at 10:49
    
Heh. I though "acceptance rate" meant how many of my answers were accepted as the best one. –  Dan Dascalescu Oct 13 '12 at 4:57

Totally rewriting an answer is not a good idea, for several reasons. One being that the text will be attributed to the wrong person, another that the votes on the answer doesn't reflect the quality of the new answer.

It would be much better to add a comment saying "This was valid for version X, please see my answer for versions Y+". Then you can write a totally new answer that we can all vote on. That way, the community votes will reflect your answer.

In addition, you don't have to go looking for the original poster, because he will be notified that there is a proposed edit to his answer. However, unless he happens to be online at the time, it is likely that other people will review your edit instead.

share|improve this answer
    
This. Most answer votes are cast a long time after the answer was actually written, by users who found the answer genuinely useful. –  Emil Vikström Oct 9 '12 at 7:29
    
Also: editing the outdated answer to add version info is not bad in my opinion. Just add in a "In versions prior to 47.11 ..." to the beginning. –  Emil Vikström Oct 9 '12 at 7:31

You must log in to answer this question.

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