18

Every once in a while, I stumble upon posts where comments ask users to accept an answer. In my opinion, it generates a lot of noise for something that is no longer needed.

An example of such noise: Why doesn't using two cd commands in bash script execute the second command?

Does it really matter that much to have that tick on an answer that we still need to ask for this?

I know in the old days of SE, we used to ask this so users have a higher accept rate, but that is not the case anymore.

  • 8
    If an answer is not accepted or upvoted, we can't close similar questions as dupes. – Pilot6 Mar 16 at 11:00
  • 1
    @Pilot6 But that's by design for a good reason, in my opinion. The system won't allow that for a good reason, as it considers the other question as answered. And there is an easy solution to overcome if the answer to that question is valid in one's opinion. They just upvote the answer. If they don't think it's worthy of an upvote, then it shouldn't be worthy to be a valid candidate for a dupe. – Dan Mar 18 at 12:32
  • 1
    @Dan Is that always the case? Here's a hypothetical example: suppose there is a question related to a very specific hardware. There is an identical question about the same hardware with an answer. Now I wouldn't upvote the answer even though it "looks valid" to me as I am unable to explicitly check the validity of the answer (don't have the specific hardware). In that case the fact that the answer is accepted is good enough for me to consider it as a valid candidate for a dupe. It's quite hard to always have 5 reviewers/close-voters with firsthand experience, especially for a niche question. – pomsky Mar 19 at 6:21
25

Is accepting an answer useless?

I know some people think differently, but I do believe accepting an answer is functional. Skimming for answers, I always check for the accepted answer as a first pick to check, simply because the answer is proven to work at least for OP. This is especially the case if I am not too familiar with the subject.

Is an accepted answer always a good one, let alone the best? Absolutely not, but the same counts for voting. I've seen quite a few (up)votes over time on utterly incorrect answers, still the voting system is a good indication in general on the quality of an answer.

Do we ask people to accept?

It totally depends on the case and, even more importantly, in what way. I hate to see someone pushing OP to accept his or her answer, even before anyone has had a fair chance to post a second/third answer, or the quality of an answer is all but a proven fact.

I don't see anything wrong however in a polite request to consider accepting if OP tells you excited "Yay, this works fabulously, have been looking for this for ages", but obviously is not familiar with the fact that he can accept an answer. Nothing wrong with a polite "(only) IF the answer works for you and solves your issue, would you consider accepting..." in such a case.

If someone's profile shows he or she is familiar with accepting an answer though, I'd never ask for acceptance.

  • I feel you have a very good basis for making a decision: to prompt - or not to prompt? – Seamus Mar 26 at 1:30
16

I advocate the practice of accepting answers (mainly by new users) for two reasons:

  1. It shows that the answer worked for at least OP.
  2. It is the only way for a new user to explicitly show that they found one working solution for their problem (note that a user needs to reach some fixed, albeit very small reputation before their upvotes are visible and counted). As @Pilot6 pointed out in comments, "if an answer is not accepted or upvoted, we can't close similar questions as dupes."

I have a canned comment:

If you find this answer useful, please consider "accepting" it (by clicking the tick (✓) next to it) to indicate you've found a working solution and also so that others may more easily find it in the future.

I post this comment (or a slight variant) only whenever I see a "thank you, it worked" comment by a newbie OP (and the associated answer is not accepted).

Newcomers must be taught IMO.

  • I like your 'canned comment', and your basis for making the comment. If I use it, will I be guilty of "cultural appropriation"? :) – Seamus Mar 26 at 1:32
  • @Seamus Not at all! Please feel free to use it :) – pomsky Mar 26 at 1:38
1

Does it really matter that much to have that tick on an answer that we still need to ask for this?

It depends. The original point of accepted answer is to signal that the solution works best for OP's question, as per Stack Overflow's help page:

As the asker, you have a special privilege: you may accept the answer that you believe is the best solution to your problem.

Now, of course this implies the answer is subjectively best, not objectively.

  • The OP can be completely unaware that the seemingly working answer can be actually wrong, malicious, unintentionally buggy/faulty, advocating bad practices, or creating new problems on their system. A small example I've once found is an accepted answer that advocated allowing SSH login for root on an embedded system running Ubuntu.
  • The OP may have left the site/forgotten the password/has no time to accept an answer. There may be an answer that addresses OP's case, maybe even objectively better, but they might not have returned to the site to even see there's a better answer available

That also implies that the solution works for OP's specific case and doesn't imply quality of the answer. Depending on the question it can be applicable elsewhere; shell scripting, basic GNU/POSIX utilities, text processing - these topics often have basic sets of practices and conceptual foundations behind them that fit a broad range of topics, so an accepted answer can work in theory for someone else's question. Questions about broken package system, compiling kernel modules, getting specific hardware working - those depend on a lot of OP-specific factors, so accepted answer really means just that it works for OP alone. The answerer also grows and changes over time, so their accepted answer posted 5 years ago may not be the best solution they could come up with today.

Accepted rate often is used as a metric of how good the answerer is or the extent of their knowledge, often this is viewed as a metric that could potentially help one find a job, even though it is not an effective metric for that. However, there's no consistent co-relation between accepted rate and answerer's knowledge or quality of answers. There are users who have decades of professional experience as developers or sysadmins, and yet their reputation points are fairly low, plus accepted rate may be low.

IMHO accepted tick mark should matter only to OP themselves. Canned comments are OK, they give OP incentive to think about the best solution they've received thus far (which may not be the best solution 2-3 years down the road). So long as "Please accept an answer" doesn't become unnecessary noise and unnecessary pressure for OP, that's all fine and dandy.

And the same Stack Overflow post has expressed this the best:

Accepting an answer is not mandatory; do not feel compelled to accept the first answer you receive. Wait until you receive an answer that answers your question well.

Even though back in the day I would rush to ask OPs to accept my answers, by now I sort of figured out it doesn't particularly matter, and objectively means very little. Others can go ahead and try to win OP's favoritism, but I'll just focus on quality of my own answers here, and if OP's unavailable or cannot figure out my own answer might be better - well, that's their freedom and I'll respect that.

  • 2
    I ask to accept an answer when I see that an answer can be used as a duplicate in the future. Driver fixes tend not to get upvotes. – Pilot6 Mar 16 at 10:59

You must log in to answer this question.

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