What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 122 Stack Exchange communities.

Currently, the auto-accept rules state that in order to be eligible for auto-accept, the answer must be given after the bounty was started, and receive 2+ votes. I think that pre-bounty answers with substantial post-bounty edits should be eligible as well.

Only the votes received after the first post-bounty edit should be counted to decide what answer gets auto-assigned the bounty. For tie-breaking purposes (i.e. when the oldest answer among the post-bounty ones with the same vote count wins) the time of the first post-bounty edit should be considered.

The rationale behind this change is that it gives improving an answer an equal treatment with giving a new answer for the purposes of bounty hunting.

share|improve this question
    
Oh and if you're wondering about the bounty on the question of your top answer. Somebody recently ragequit and drained all their rep on bounties - across the entire network. I flagged it and the user has been removed, but the bounties are still there. –  Mysticial Aug 21 '12 at 18:23
    
@Mysticial Yes, that question and the other one were indeed the inspiration for this question. But I think the change request is good on its own, without regard to an auto-assign by the Community user. It it's a rage-quit bounty, though, I think it should be rolled back to the poster. –  dasblinkenlight Aug 21 '12 at 18:36
    
3 of them actually (just on SO, with more on other sites). But apparently there's a bug that prevented the 3rd one from appearing on the user's profile. (before the user was deleted) –  Mysticial Aug 21 '12 at 18:39
add comment

2 Answers

If you have posted an answer and later want to radically change it, you should consider giving a new answer. Then you would automatically be eligible for the auto-accept.

That way, the need to determine what a "substantial edit" might be disappears, and you don't run the risk of handing out half-bounties for answers that obviously didn't answer a particular question (otherwise there would've been no bounty).

share|improve this answer
2  
Although I completely agree with the approach of giving an alternative answer when you wan a "radical change", there are situations when you want to improve the answer without providing an alternative solution. For example, you may want to discuss your assumptions, explain the limitations of your approach, provide references to standards, or give an extended quantitative analysis of your solution. All of the above adds substance to your existing answer, but it does not give a good reason to start a brand-new answer. Unfortunately, the current rules push users to providing a second answer. –  dasblinkenlight Aug 21 '12 at 17:21
    
@dasblinkenlight: That is perfectly fine, and the OP still has the option of manually accepting your answer. However, if we agree that the core of the answer remains unchanged, the default should be to assume that it solves the OP's problem just as much as it used to before the edit -- even if you add sources and quotes. –  bitmask Aug 21 '12 at 17:27
    
When OPs assign the bounty manually, there is no problem: they could give it to old answers as well as to new ones. The situation in question is when the bounty is assigned automatically using the default rules (see "What is Automatic Awarding" here). –  dasblinkenlight Aug 21 '12 at 17:31
    
@dasblinkenlight: Yes, exactly! That's just what I'm saying. If -- after your edit -- the bounty is not awarded manually, it's probably because the OP is not convinced the post answers their question. You're making a point about auto-accept; That shouldn't award bounties to answers (edited or not) that were given before the bounty was awarded, because the OP wouldn't have awarded a bounty if they felt their questions was answered. –  bitmask Aug 21 '12 at 17:38
2  
The same exact argument can be made about the new answers posted after the bounty period has started: since the OP did not award them a bounty, he or she did not feel that the new answer solves their issue. Yet these new answers are eligible for auto-award, while the edited old ones are not. This lack of symmetry is wrong, and should be fixed. –  dasblinkenlight Aug 21 '12 at 17:42
    
@dasblinkenlight: The point is, the OP offered a bounty despite the fact that there were already answers. If the content of the answers doesn't change they shouldn't receive a bounty, because their value to the OP hasn't changed. Frankly I'm not a big fan of the auto-award altogether, but at least it shouldn't be applied to old answers. –  bitmask Aug 21 '12 at 17:46
    
A situation comes up periodically when a user who offered a bounty gets deleted before assigning the bunty (here is an example of one such question; this bounty will be auto-awarded in six days if there is an eligible answer). An attempt to capture this bounty "on the cheap" is very likely. I think it would be better if some of the original answers could participate as well, perhaps after an edit, because none of them are incorrect. –  dasblinkenlight Aug 21 '12 at 18:19
add comment

I could not agree more with the point of allowing existing answers to be eligible if they have had substantial edits made to them, or edits of any nature.

Why should we have to spam answers in order to be eligible for the community bounty? That is clearly a workaround / hackish approach.

"Only answers submitted after the bounty is started (with a score of at least 2) are eligible for auto-acceptance." is the current rule, and needs to be changed as it is not conducive to "getting great answers to questions" which is the goal of the bounty system.

As it stands, the auto accept feature of the bounty system hinders great answers because it discourages people from improving their answer.

share|improve this answer
add comment

You must log in to answer this question.

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