9

With the increased usage of technology and particularly with the internet, the world changes at an increasingly rapid rate more than ever before. We’re still relatively bad at creating with this in mind but that doesn’t mean that we can’t try to do just that.

One of, the biggest problems with the SE model is its outdated answers. These days I find myself answering old questions which have out of date or not complete answers more than I answer new questions. I also see some other higher rep users also posting new answers to old questions fairly often. However, there are many more questions with old, outdated, or partially complete answers that could be improved. They are sometimes not easy to come across for us because we, as professionals, have sometimes moved past the stage at which we’re looking up the information involved in the post.

Here on SE we want very high quality posts and content. This is the reason why the “too minor” rejected edit reason was removed. Leaving incorrect old posts (even partially incorrect ones) when lots of people are viewing them should be a big priority to fix.

My goal in this proposition isn’t primarily to address what to do with out of date answers on old questions (besides improve them), there are already several others on the subject, but to have an easier way for these questions and answers to be found, added to, and improved.


I’d like to have another review queue to help users find and appropriately deal with these dated answers.

The way I imagine the queue would be similar to just viewing an old question currently. The reviewer can vote on the question and answers, leave comments, edit questions and answers, close and flag when needed, and provide an answer them self if they’d like. The only thing required in addition is something like “Bump this question”, “No improvement needed”, and “Skip this post” buttons. If 3 (or however many) reviewers choose to bump the question, it would get bumped to the main page by the Community. If 5 (or however many) reviewers choose that no improvement is needed, then it is not bumped and nothing is harmed. If the reviewer can’t make a good decision, they can skip it.

There would have to be some question requirements for questions to be eligible to be in the review queue. Here are my initial thoughts for those requirements:

  • A view limit - If no one really looks at a question there’s no real point in improving it. This view limit would be nice to be able on a per-site basis as SO gets a lot more traffic than smaller sites. Allowing a variable one set by moderators would also allow the most viewed questions (there are currently 223,000 SO questions with over 10,000 views) to be fixed first before moving on to lesser viewed questions.
  • A date limit - A certain period of time since the question has been posted or until it was last reviewed. I haven’t given this requirement too much thought or research, but most out of date answers I see are at least two years old. so a timeframe of about that long seems to be appropriate.

There would also need to be some reviewer requirements in order to assure that an appropriate person is reviewing the old question and answers:

  • A reputation limit should be required to view the queue in order to prevent crappy answers to old questions. I think something as low as 1k should be allowed as reputation doesn’t necessarily determine the quality of answers given, but something higher may be more appropriate.
  • There also needs to be some way to make sure the reviewer is knowledgeable about the subject. The way I can see that being true is requiring a score of 40 (or some other amount, maybe a bronze badge) for one of the tags used. This will also help prevent robo reviewers because I feel like most people with >40 usually know the system somewhat well and care some for the subject they're knowledgeable in. While this doesn’t necessarily mean the reviewer will know enough to make a good judgement, it will help do so. Perhaps additionally allowing reviewers with a lower score (say 20 or so) in two of the tags could be useful.

I am unsure whether or not the post should remain in the queue if a user chooses to edit one or more of the posts (and therefore automatically bumps the question). On the one hand if it stays then more useful content could be added. On the other hand it has already been bumped, so voting to bump it wouldn’t have much of an effect at all in most cases.


Overall there are some things we need to consider in more detail, but I think this is a step in the right direction because it would help improve the quality of content across the board. By providing the community with better tools to both post new, quality content and to moderate older but still relevant content, it’d help fix a big problem of the model SE uses. As such, it’s a feature I’d really like to see implemented.

  • 3
    OMG this isn't a highschool essay, this is a feature request. No need to start with a thesis, use an attention getter, add fluff everywhere... Needs more brevity! – bjb568 May 25 '15 at 13:57
  • 4
    @bjb568 I believe that some people don't see the need in a request like this, so I wrote what I did to try and help users see that need – Zach Saucier May 25 '15 at 13:59
  • I'm also not convinced that this is a problem that can't be solved with existing tools; you describe the interface as being very similar to viewing a qurstion normally, do we really not have the search tools to find old stuff to improve? – bjb568 May 25 '15 at 14:03
  • 1
    @bjb568 I don't believe that we have the tools to find a large number of these posts very easily. If I did I wouldn't be making a feature request. But finding them is only part of the proposition. Implementing this would assure that old content is reviewed regularly (every 2 years or however long) and also provide users more reason to moderate relevant content. People simply don't do this sort of thing too much because it's relatively difficult to do so – Zach Saucier May 25 '15 at 14:13
  • It seems like this queue would have to be used a lot to be effective. We have a huge pile of content of varying qualities, do we have people willing to look thru it all? – bjb568 May 25 '15 at 14:14
  • 1
    @bjb568 Our community does it currently, but it takes longer than it should for all the relevant posts to be reached because the community simply don't come across relevant posts that need improvement much. Instead the people who come across them are searching for answers, seeing dated answers, and then leaving with outdated or not a complete answer. Rarely but sometimes these users will ask a new question asking essentially the same thing but seeking more modern, complete answers – Zach Saucier May 25 '15 at 14:17
  • 1
    To make sure I understand, the result of action in the queue is bumping the post to the front, correct? Then what? How is this different than the automated bumping Community already does? – Andy May 25 '15 at 14:34
  • @Andy It's the same type of bump, but applied to specific questions that need to be bumped, not randomly as it currently does – Zach Saucier May 25 '15 at 14:41
  • 1
    If implemented, I would think the community bumping as it is now wouldn't be necessary – Zach Saucier May 25 '15 at 14:51
  • 1
    People need an incentive to improve these old questions. If such a queue is made then those questions should offer some sort of fake internet points or shiny CSS badge for those who take the time to improve. I honestly wish the community could override the accepted answer for a question though. – cryptic ツ May 25 '15 at 20:29
4

Effective software changes are implemented to fix problems. Is there one?

Yes it seems, and as mentioned:

  1. Outdated answers. These answers may have been best practices then, but now using them is just outright stupid with the better alternatives.
  2. Obsolete answers. They no longer work. They might once had.
  3. Incomplete answers. Needs more work to become helpful.
  4. Old posts gets unnoticed and therefore the above issues worsen.

Please notice the bold parts.

These are not fixed by implementing a review queue!


Ch. 1: Counterarguments

    ▶   Reviewers would need the ability to unaccept / vote to detain these answers which were once decided to be helpful, and therefore with a lot of votes. At the current Stack Overflow policy, you're not supposed to do anything to them aside from downvoting, and moderators are reluctant to fixing them anyway - if you know good enough, you'll be able to fix the answer into shape.
    ▶   Review tasks run based on reviewer voting. Review queues are inherently problematic as it attracts robo-reviewers who click on buttons to earn a badge, not to help on site quality. The effectiveness of implementing a review queue for this is questionable.
    ▶   If there are indeed that many minds alike who are looking to clan up the obsolete content, they would already be doing it - By pointing out the issue on Meta, then going on an edit raid. Unfortunately, the concern has been addressed and never got enough attention to life off, see: Previously, previously, previously, previously...


Ch. 2: Alternatives, Existing functionality

In discourse of this feature in Tavern on the Meta, I said:

Can't you sort by [active] and start at page 5 (or other arbitrary number) and scan through them? source

How about searching for questions with >N views with search parameters and then sorting by active? source

Also, for any arbitrary criteria that defines "old", you can just query that on SEDE or as a SE search. source

To be simple, currently everyone can access:

  • Your nearby search engine, with advanced search options to filter per site. Comes with SEO sorting.
  • Site search feature. Comes with vote sorting, active date sorting, as well as search parameters for view or any kind of criteria.
  • Stack Exchange Data Explorer.

Any arbitrary review queue criteria would be fulfilled by using these tools.


Ch. 3: Target Audience

A reputation limit should be required to view the queue in order to prevent crappy answers to old questions.

1: It is expected that reviewers have great answering skills.

There also needs to be some way to make sure the reviewer is knowledgeable about the subject.

2: It is expected that reviewers have technical knowledge in the field of posts they're reviewing.

I feel like most people with >40 usually know the system somewhat well and care some for the subject they're knowledgeable in.

3: It is expected that reviewers care about the subject.

Matching all three fields, sure.

4: They would enjoy investing time in such a review queue where they are presented old questions.

... Nah.

  • 1
    Thanks for your input! If people end up not liking this feature request perhaps a custom (perhaps a non-official) queue can be created using the tools you mentioned. The problem is that I'm not sure how one could be made that is has similar capabilities of the feature request. If someone could demonstrate that I'd be very grateful! – Zach Saucier May 25 '15 at 15:22
  • @ZachSaucier are you talking about using the SE API or just the SE Data Explorer? The latter is much more powerful and I've used it with great success in finding dupe user accounts, spammers, etc. Not sure what criteria would be used to determine if a post fits the issue at hand though. – cryptic ツ May 25 '15 at 20:33
  • "It is expected that reviewers have technical knowledge in the field of posts they're reviewing." Really? The expectation I've always seen has been explicitly the opposite: that reviewers, in most cases, only need expertise in the general site topic, not the question domain as such. (And enough experience to recognize when domain-specific knowledge is needed, such as duplicate closing.) – Nathan Tuggy Oct 2 '15 at 2:05
  • @NathanTuggy: Exactly! If this suggestion gets implemented, then reviewers are expected to have technical knowledge in the field of posts they're reviewing, which is in no way true. – Unihedron Oct 2 '15 at 9:02
  • @AMUnihedron: Ah, OK. It wasn't clear that you were listing that as a problem with the suggestion. – Nathan Tuggy Oct 2 '15 at 16:31

You must log in to answer this question.

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