18

Community manager Nicholas Chabanovsky has posted this on Meta Stack Exchange:

Unpinning the accepted answer from the top of the list of answers.

On Stack Overflow, apparently, answers are now by default sorted by score, with the accepted answer appearing wherever its score places it, instead of being pinned to the top.

The linked question asks whether we would like to have the same behaviour on our site, or stick with the old way of having the accepted answer appear at the top of the list regardless of its score (or, I guess, something else entirely).

What are your thoughts on this?

This question is about the main site, Ask Ubuntu. For the same question with regard to meta, see my other question.


I have posted feedback on this discussion in answer to Nicholas Chabanovsky's question

4
  • 1
    Why the confusion? "(or, I guess, something else entirely)" Sep 8 at 18:31
  • 2
    @RandomPerson I can only imagine two possibilities, but I am not so arrogant as to think my imagination has covered everything :)
    – Zanna Mod
    Sep 9 at 4:57
  • Here is (another) perfect example of why this is bad idea (from the SO change). I just noticed this today, since the OP finally accepted my answer, but someone posted a wrong answer a month ago that somehow got 2 votes (perhaps during "late answer" review) and is now the top answer by sorting. Low "viewer" questions (like WSL) are often easily susceptible to just a few votes pushing the wrong thing to the top. Sep 9 at 20:17
  • @Zanna Since Nicolas Chabanovsky has said that they ".... have adjusted the site setting ...", I request you to edit the post informing that the change is live now. On a side note, I guess AU is the only site which discussed about accepted answer pinning in meta (๐Ÿ˜›). I hope the accepted answers are still pinned in AU meta and no changes are made in AU meta. Please check if that's the case. Sep 29 at 18:04
19

Accepts are a way to indicate what worked best for the OP or which answer they personally liked. Sometimes the OP's choice isn't the best one - after all, many people ask questions because they don't know much about the topic, so they can't necessarily be expected to be a good judge. Also, sometimes better answers are posted later after the OP has stopped interacting with the question or no longer has the ability to test solutions.

The score of a post gives a more reliable indication of its quality, as potentially more (and possibly more knowledgeable) people are involved in determining it.

Therefore, I think we should have the accepted answer sort according to score, rather than being pinned to the top.

For anyone interested, here's an SEDE query for posts where the accepted answer is outscored by another answer (according to this, on Ask Ubuntu there are around 9000 such posts).

2
  • 2
    Nice explanation. It would be nice if you can mention few examples of questions where there are better answer than compared to the accepted one. Sep 8 at 18:33
  • 2
    @RandomPerson see my edit - seems better than individual examples
    – Zanna Mod
    Sep 9 at 5:19
12

Obsolete Answer Flag

In many cases the answer with the most upvotes is simply the one that has been here longest. Often these high vote answers are totally obsolete and no longer work.

It was much easier to get votes in the old days, when Ask Ubuntu was new and unique. Answers did not need to be very good to get lots of upvotes.

There should be an obsolete flag so that users can flag these old and obsolete answers to drop them from the top of the list. A flagged answer does not need to be totally removed, just dropped to the bottom of the page where it won't hurt anyone. It should probably take a vote of five flags to demote an answer.

Closing a duplicate question and then linking an old question with 20 obsolete answers wastes the users time and does not give users much confidence in Ask Ubuntu.

Some people talk about historical value of these answers but people come here to get help making their Ubuntu work, not history lessons.

Lets keep Ask Ubuntu relevant and up to date.

7
  • 4
    "Closing a duplicate question and then linking an old question with 20 obsolete answers wastes the users time and does not give users much confidence in Ask Ubuntu." -- Hear, hear! And this applies to all of Stack Exchange. Sep 9 at 4:33
  • 3
    +1, although this isn't on the table in this particular round of changes, it's good to keep it on the radar - also it's relevant to NotTheDr01ds' argument' - I'll be sure to mention it when I post feedback from our site on Nicholas' post.
    – Zanna Mod
    Sep 9 at 4:58
  • 3
    This is an excellent point and is a +1 by me. A drawback, however, is that this would introduce another queue in the review queues, and I don't think that we currently have the number of active reviewers needed for this. I would like to make clear that I'm not against this suggestion, I'm just mentioning a potential problem. Sep 9 at 15:10
  • 1
    @BeastOfCaerbannog I guess this will only be a problem for the first few months when this new flag gets launched (if it gets launched). I guess a new review queue named Obsolete Answers would be created. Once the review queue gets cleared, then new posts will be rarely added in the queue as answers don't become obsolete so soon. But anyway, I am just assuming things, so I could be 100% wrong. [Let's assume that the 14-day invalidation rule doesn't exist for this review queue๐Ÿ˜….] Sep 9 at 18:07
  • This seems like a very good idea, but posting it here, as an answer to an unrelated question makes it sure to be ignored. This isn't something that can be done by the AU community, it would require SE developers to add new functionality, so this really should be posted as a feature request on Meta Stack Exchange where it can get the attention it deserves.
    – terdon
    Sep 11 at 12:02
  • 1
    @terdon I don't think this is an unrelated question. Also, please check these links: meta.askubuntu.com/a/19732 and meta.askubuntu.com/posts/comments/42909 Sep 11 at 18:22
  • 1
    @terdon: Both question and answer are about unpinning obsolete answers from the top of the list. I am advocating sorting answers by value rather points or acceptance. I think acceptance should be at the top at least until it becomes obsolete. Sep 12 at 8:12
8

From my limited experience looking for solutions (especially programming related on Stack Overflow), I have noticed that the accepted answers are not generally useful for my case. This could be due to different versions of software used by the original questioner and newer features have been built over the time, which makes the accepted answer less useful for my situation.

Fortunately, there are multiple other people having similar questions and some of them have posted their own working solutions which have been helpful for me. It was interesting to note that these newer updated answers have way more upvotes than the accepted answer. I reckon unpinning the accepted answer will help us find our solution faster -- a few minutes earlier at least.

The argument that OP will be responsible enough to come back and check which answer is more useful and change their accepted answer does not seem to happen often. We have had multiple flags asking if mods can reassign accepted answers, because people rarely change their accepts even if a better answer has appeared.

5
  • 1
    I mostly agree. If the answer worked for the OP and they accepted it, why would they come back and change the accepted answer? They are likely not even using 10.04 or WUBI or whatever obsolete OS or app the question was about any longer. The people closing questions should be encouraged to read the proposed duplicated question before voting to close the duplicate. Giving the OP a similar question with 25 obsolete answers is not doing anyone any favours. The new updated answers are usually down at the bottom. Sep 18 at 8:50
  • wait.. what? Mods can change accepts? I didn't know this. ๐Ÿ‘€ BTW, have you (any moderator in AU) accepted such flag requests? Sep 18 at 19:29
  • 1
    @RandomPerson no, mods can't change accepted answer. They only received flags from uninformed users who don't know about this restriction.
    – Andrew T.
    Sep 19 at 3:40
  • But AU is not SO. Sep 21 at 23:43
  • @AndrewT. Thanks for the information. I am glad to know that the OP is powerful than a diamond moderator at least in one aspect (i.e., accepting answers of their questions) ๐Ÿ˜‚ Sep 22 at 17:20
5

Neither system is perfect. The problem with vote totals deciding the "best answer" (by sort order, at least) is that they are momentum that is hard to overcome in the event that things change. And with the age of some questions (on most of the Stack sites), a new answer can come along that makes much more sense today than the top voted answer of yesteryear. I've seen this happen multiple times in my relatively short time answering on Stack sites.

The ability of the OP to change the accepted answer based on new information is useful, and allows the new information to automatically get to the top, regardless of how many votes the old, outdated (often even now wrong) answer might have. Without this, the momentum is for readers to continue to read the highest voted answer, and often even upvote it again since it might seem to work for them.

Sure, it's a lot of "power" to give the OP, but someone needs to have the ability to "pin" an answer, IMHO, and who better than the person who asked it originally?

Vote totals, which are difficult to change after months or years, aren't always the best metric for the "best answer" -- Especially on older questions.

4
  • 4
    Very good point, and I have seen a few accepts change, but I feel it rarely happens. Also, just pinning the accepted answer doesn't seem like the best way to address the problem of accumulating obsolete answers at the top - I'd prefer a new feature like C.S.Cameron proposes here to deal more effectively with that.
    – Zanna Mod
    Sep 9 at 5:02
  • 1
    @Zanna Absolutely agree that an Obsolete Flag would be better, but in lieu of that, all we have is the current "Accepted Answer" pin. So I'd personally prefer to see it kept in place. I say that as a relatively new poster who has provided updated answers to older questions, and then had it changed to the Accepted answer. I'm not one who (like some I've seen) resort to putting something obnoxious like a Heading1 at the beginning saying "2021 Answer", so I tend to rely on the good sense of the OP (when they are still around) to pick the new answer. Sep 9 at 19:42
  • "is that they are momentum that is hard to overcome in the event that things change" this answer of mine would like to disagree
    – Braiam
    Sep 15 at 14:38
  • @Braiam I'm confused - it looks like that answer agrees with my point. The top answer (now pinned, but also accepted) has 1395 votes (currently). Yours has 76. Even if the OP were to accept your answer as correct, it would still be buried beneath 4 other answers (and about 15 page scrolls) for the years it would likely take it to achieve another 1000+ votes. Sep 15 at 15:02
1

I am late as usual. But, Nicolas Chabanovsky, a community manager, said "before the end of September 19th". It is still 18th September according to UTC ๐Ÿ˜œ.

I am not gonna discuss about the merits or demerits of unpinning the accepted answer. Other answers do a good job about them.

I am proposing to have a trial before coming to any conclusions. Since many people seem to support unpinning the accepted answer, let's have a trial of that for, let's say, 1 month or 3 months (or whatever time period which would be good). This will help us get a better picture.

If the impact of unpinning the accepted answer will be good, we can continue having accepted answers unpinned. If the impact of unpinning the accepted answer will be negative, we can change back to pinning the accepted answer.

But, I am not sure how we can assess the impact. (Maybe the moderators or community managers can assess the impact...)


Catija, a community manager, said this:

.... This should be a site setting that the CMs have access to and can change in a minute. Provided the request was status-reviewed on the child meta, I don't think this would take more than a day or two to change.

But since she mentioned "should be", I presume that this "setting" has not been created yet.

Meanwhile Anita Taylor, Product manager for the Public Platform team, said this:

The Public Platform team will go through these requests and determine what to prioritize, defer and decline. We ask for your patience -- key members of the team are on vacation over the next few weeks, so we won't be able to triage these requests until mid-to-late October.

Considering the fact that the triaging of requests will take time until mid-to-late October, I request the moderators to not remove the tag from the two posts such that more people will have look at these posts.

Also, our meta is not much active. So having a trial (and then coming to a decision) is better than directly coming to a final decision based on votes in meta.

3
  • 1
    I can't think of any metric to gauge the effect (let alone over such a short time period) other than how people feel about it
    – Zanna Mod
    Sep 19 at 6:12
  • @Zanna Sorry for nagging. Can you please confirm with a CM? Also, I wasn't aware that the time periods which I mentioned were short (anyway, I also said "or whatever time period which would be good" ๐Ÿ˜œ). Also, I feel people's feelings too matter in addition to statistics. Sep 22 at 17:18
  • 1
    People's feelings most certainly matter - I just mean that I'm not aware of any statistics that would be informative in assessing the impact
    – Zanna Mod
    Sep 23 at 6:50

You must log in to answer this question.

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