10

Today, on apple mailing lists, someone said:

Keep in mind that out-of-date information never gets removed from StackOverflow. It's a treacherous place to use as your primary source of information.

I believe, we all come across such Q/A frequently. Updating them is a huge task. However, marking/signaling the answers as outdated could be useful.

It could be presented in the form of a visible message telling that the information might be outdated. The number of users that did report the post outdated could also appear. There may also be other ways to proceed. Just some thoughts.

For instance, for those who need to check Apple's documentation, APIs, sample code… they have such system. It is limited, but effective at telling some content is outdated.

15
  • 12
    But outdated does not necessarily equate to "no longer useful". What about those who have to maintain legacy code?
    – Bart
    Commented Mar 23, 2013 at 0:03
  • 3
    That's why we can edit posts, and if enough people edit it will become wikied and end up being maintained by the community. There are many posts that are already like that. A post if it is outdated just needs to get the ball rolling by having someone update it.
    – kittycat
    Commented Mar 23, 2013 at 0:04
  • Of course. That's why I wasn't talking about deleting them. Flagging these question as outdated may actually help users by telling them there are more up-to-date solutions.
    – Jean
    Commented Mar 23, 2013 at 0:05
  • Then you might want to clarify that in your request. Because flagging is usually meant to bring something to a moderator's attention.
    – Bart
    Commented Mar 23, 2013 at 0:07
  • @Bart Ok. Sorry then. Is "marking" a more suitable word? (I am not native speaker) – I edited my post. thx.
    – Jean
    Commented Mar 23, 2013 at 0:08
  • So, what? Would we just base this naively on the date/time stamps? Or would we have some of marking the question as "outdated"? If so, who would do that? Would it require a consensus vote? A moderator? How would those people make that decision? Would we require them to be experts in the field (e.g., based on tag reputation)? Would the decision to mark be reversible? Would anyone even read the message if we had one displayed? (Hint: no) Commented Mar 23, 2013 at 0:14
  • And wouldn't this take away from the primary goal/expectation that when experts come across an answer with outdated information, they edit the answer to update it? Commented Mar 23, 2013 at 0:14
  • No it wouldn't. Editing all of them is a huge task. So, pending editing, they could be marked/tagged/signaled… Wikipedia does that, so that while the article waits to be edited, others can know. And it can also help the editing process by identifying what needs to be done. It wouldn't be based on timestamps, but on "flags" by those who did recognize the out-of-date info on the text.
    – Jean
    Commented Mar 23, 2013 at 0:16
  • I guess, reaching a certain amount of reputation, users would gain the possibility too mark outdated content. The number of such "flags" would be displayed in the message. Readers could then make up their mind "careful, 8 users have marked it as outdated… this may no longer be the solution to this issue"...
    – Jean
    Commented Mar 23, 2013 at 0:20
  • @Jean: How would the removal of such flags be (efficiently) orchestrated following an update? Commented Mar 23, 2013 at 3:13
  • Once an edit is made, the users that did mark the post as outdated would be notified, so they come back and check the edit to remove their outdated vote, if they deem so. I believe those who did mark the post as outdated to be the more qualified to decide wether or not the update does indeed solve the issue. Some question, however will stay outdated forever.
    – Jean
    Commented Mar 23, 2013 at 11:37
  • 2
    A similar thing was asked in "Evolving Subject Q&As are outdated" with a more extreme suggested solution. As I argue there, I don't think the rate of obsolescence for Apple-related development technologies is as fast as people assume. The vast majority of answers in [ios] and [cocoa] left three or four years ago still seem to apply. The person you quote there may have a mistaken impression of the site based on one or two problematic questions or answers. Commented Mar 23, 2013 at 21:47
  • Highly possible, I agree. However, while the second sentence is purely subjective and may or may not be based on too few experiences, the first one is a fact. And as technology evolves, it can only get worse. The importance of the need to take action purely depends on the number of such outdated answers and their impact. Left to the community to decide what action should (or shouldn't) be taken.
    – Jean
    Commented Mar 23, 2013 at 22:26
  • By depends on the number of such outdated answers and their impact I mean that an outdated answer to an outdated question probably isn't an issue. The real problem is outdated answers to current questions. APIs get deprecated, replaced. Vulnerabilities discovered, and advices on best practices may evolve...
    – Jean
    Commented Mar 23, 2013 at 22:29
  • How ever, (I read all comments) I - still - think that a flag or mark for (possible) outdated threads, would be usefull; just a mark in "related questions" and/or a mark when opening the question, not delete or filter, just a hint for silly ones like me, reading a simillar question as I have, ending up in 'with IE5 it will be corrected" ...
    – halfbit
    Commented Nov 9, 2013 at 20:23

2 Answers 2

5

Practical case

In this thread Firefox can't establish a connection to the server at ws://sockets.ny.stackexchange.com/, there was an issue about real-time exchange method between browser (client) and sockets...stackexchange (server).

This kind of technology is something new and the question was about a specific browser: Firefox, but in a specific version (v12 on XP).

Without real answer to this question, by normal evolution, newer versions of Firefox (now I use version 22 of Firefox) has this bug corrected.

So I agree with that simply dropping the thread is not a good idea, but I think that adding a little header, like this:

Outdated - 2013-05-28

because there is new versions of tool/libraries where the bug don't exist anymore

I've already do this on my answer: I'm using Firefox 17.0 (french..., but I don't now if I could write this on the top of the question. Maybe in a SE standardized form.

I think that this could become a special SE feature.

3
  • +1 Maybe in a SE standardized form, a special SE feature → Just like Wikipedia does. You get standardized, highly relevant, warnings about all kind of things, showing up at the top of the article.
    – Jean
    Commented May 28, 2013 at 11:18
  • Another example: stackoverflow.com/q/4883803/1224741
    – psoft
    Commented Nov 28, 2013 at 17:48
  • Another example - it's outdated according to this comment.
    – BartoszKP
    Commented Feb 3, 2014 at 14:19
0

This would be a great feature for outdated answers too - perhaps requiring multiple votes (like close) or requiring moderator approval (like flag).

To mark an answer as outdated there could be additional constraints, such as:

  • There must be another answer to the same question
  • Other answer must have at least 5 votes
  • The 'outdated' flag must reference the newer/better answer

Once marked/approved as outdated, the presentation and other behaviour could include:

  • Similar to duplicates, a styled "this answer is outdated, see this other answer instead" box would appear at the top of the answer.
  • The answer would be sorted below the newer-referenced answer, even if it is accepted and/or has more votes.
  • If it is the accepted answer, the OP could be notified to suggest they change their accepted answer.

You must log in to answer this question.

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