An empirical study of obsolete answers on Stack Overflow [pdf]

UX for fact checking, arbitration, correctness + staleness will be critical to the survival of information societies<p>SO not having a good way to retire these is the same issue as twitter&#x27;s &#x27;get the facts&#x27; banner<p>IMO wikipedia is the only organization that&#x27;s solved it, and my sense from the outside (I&#x27;ve only written 1 article) is that it&#x27;s with editor-gatekeepers, not with tools (though there are some bots)

What are the issues with having a &quot;mark obsolete&quot; flag that users can check? (with an optional comment)<p>At a minimum, that would be an input to the presentation ranking -- old, flagged items would drift to the bottom.<p>Long-tail &quot;Floatsam and jetsam&quot; content is a huge problem, generally, not just for software development information.

There doesn&#x27;t seem to be any remedy to the problem of poor, obsolete, or outright wrong answers being selected as the checked answer when the asker disappears. I frequently have to scroll past the approved one because there&#x27;s often gold hiding below it.

that&#x27;s something was long discussed on Meta, specially to answer the question: &quot;Is it fair to down-vote answers which were right in 2009, but aren&#x27;t right anymore?&quot; and the answer was &quot;Yes, if you want to keep your points, you should maintain your answer as long as it exists&quot; which is a kind of no-go for somebody like me with more than 14k points and more than 500 answers...

The call to action, section 5, is exceptional for an academic paper. I&#x27;m surprised the first suggestion doesn&#x27;t already exist on S.O.<p>Honestly this is my biggest complaint with the web in general: immortal anti-information. Proposing analysis and strategies for combatting it on curated platforms is a great first step.<p>I think also implicit in this discussion is the role of the readers to vote with their mouses, so to speak. Without feedback from users, the mechanisms can&#x27;t work effectively. Which is why I try to upvote as much as reasonable on HN and SO.

I kind of feel the same issue is being expressed in search engines as well. As time progresses, more relevant answers are moving down the list. Using Google for example, I&#x27;m finding that I have to employ the &quot;Tools -&gt; Time Range&quot; filter to get better, more relevant results.

It needs to be versioned. Maybe you are working in an environment with an old C++ compiler that can do the latest C++ tricks. You want best practices from state of the art 10 years ago.

It might be helpful to have the age of comments listed as part of their metadata, ie alongside the date the comment was posted. Some formatting could be added (eg red highlight for comments &gt; x years).<p>I know this sort of feature is useful on newspaper websites - The Guardian will flag stories older than some limit as being potentially out-of-date.

I’m definitely in the minority here but when I’m learning a new language I kind of prefer to use a slightly outdated resource. When you get stuck and look for help you get the gist of the answer you’re looking for but not the solution and then you have to figure out the rest. It’s like a hint that gets you 50-90% of the way there.<p>When you’re in the middle of work and not actively trying to learn new material I’m sure the obsolete answers are frustrating but I don’t think the outdated information is entirely useless.

I agree with the conclusion of the study (and payne92&#x27;s suggestion here) but I haven&#x27;t found obsolescence to be a huge issue because I generally take the time to review all the answers provided and the comments on them. Often times obsolete answers are noted in the comments.<p>But there is room for improvement.

This seems to imply that answers are binary, either obsolete or not, while in my experience they are often are only partially outdated. How is that treated in this study? Additionally, I find that for high traffic answers someone will often have posted an update to the obsolete response.

Per the paper, they are using an SO archive dump from <i>2017</i>, which is ages ago in internet time, although admittingly the problems with SO comments extend even before that.<p>It looks like the latest archive dump (March 2020) is available in BigQuery, e.g.: <a href="https:&#x2F;&#x2F;console.cloud.google.com&#x2F;bigquery?p=fh-bigquery&amp;d=stackoverflow_archive&amp;t=202003_comments&amp;page=table" rel="nofollow">https:&#x2F;&#x2F;console.cloud.google.com&#x2F;bigquery?p=fh-bigquery&amp;d=st...</a>

How could SO incentivize the community to clean that up?<p>Given that SO&#x27;s content is basically 100% volunteer&#x2F;user generated and free to access, it seems like the first step would be to allow users to flag obsolete answers with a very visible and obvious UI element.<p>Maybe second would be SO fielding a team of experts as &quot;pruners&quot; that would delete&#x2F;update the flagged answers.

This of course(!) sounds as if like it&#x27;ll like be quite good.<p>But. I am more interested in the inevitable corollarial follow-up:<p><i>An</i> Thorough &amp; Ignominious <i>Probing Investigation Into The Problematic Presentation Of Pseudo-Intellectual Wankery On Hacker News</i>

It would be nice to have answers marked as obsoleted and then a little historical breakdown of the greater changes that have come about to make that answer obsolete