25

Is there anything that can be done to answers already provided some time ago, that are currently out of date?

I mean: I may have a problem, and find on TeX.SE the answer to that. Only that the answers uses inferior packages which were perfectly fine then but now they aren't anymore.

Since I find my question answered, I proceed to implement the possibly ugly solution.

Is there anything that can be done to mark off some questions as "not-applicable anymore"?

Is adding one such feature viable?

Thanks

6
  • 3
    See also How can I know whether an answer is obsolete?. In general it is hard to know whether an answer is outdated or uses deprecated packages. Even if you know that it is outdated, it can be (but by no means always is) hard to get it updated. See also What shall we do with future out-of-date answers?
    – moewe
    Mar 30 '17 at 12:18
  • Ok, but it can always be flagged, right? That way everybody knows that they're dealing with obsolete code...
    – Moriambar
    Mar 30 '17 at 12:19
  • 3
    It depends on what exactly you mean by 'flagged'. If I know something is outdated, I will certainly leave a comment and hope that the answerer is still active and decides to edit their answer either with a better/newer solution or a warning. I am normally not keen to edit other peoples' answers myself even if I know what needs to be changed (I have done it a few times, but try not to).
    – moewe
    Mar 30 '17 at 12:23
  • 3
    are there really so many obsolete packages? there are some new ones but the bulk of commonly used packages would be recognisable to a tex user from the 1990s.
    – David Carlisle
    Mar 30 '17 at 18:54
  • 1
    @DavidCarlisle well, I don't know. Sometimes there are some deprecated packages options like utf8x for inputenc, that still are in some well-received posts
    – Moriambar
    Mar 30 '17 at 19:00
  • 2
    utf8x isn't deprecated in the usual way that is meant, deprecated usually means something that was once standard that over time is being removed from the standard or replaced by something new. neither of those apply.
    – David Carlisle
    Mar 30 '17 at 19:03
17

The StackExchange model is that over time the 'best' answers get upvoted whilst less good ones float downward. When an answer becomes outdated, one can therefore

  • Post a comment asking the original author to adjust it
  • Suggest an edit to the original
  • Post a new answer which draws attention to the reasons that it is preferable to the older one (perhaps linking to the other answer)

The last approach is likely the best where the answers are long-standing: this will move the question to the top of the front page.

19
  • 2
    Well it depends: it's not always the case that the topic has some new answers; going into obsolescence it may be that not that many new questions arise about it (especially if it's some kind of obscure topic). A newbie could read that answer and persevere in obsolete usages... But I understand what you mean
    – Moriambar
    Mar 30 '17 at 12:29
  • 2
    @ilFuria if there is an obsolete answer and no hint that the answer indeed is obsolete then there is no way to prevent that someone may use it. But if you notice that there is an obsolete answer then you can follow Joseph's suggestion. People may still use the obsolete answer but then you can at least be sure that you've done what you could to prevent it. :)
    – cgnieder
    Mar 30 '17 at 14:13
  • 3
    Fourth option: offer a bounty for a better, up-to-date solution.
    – Andrew Swann
    Mar 30 '17 at 15:42
  • @AndrewSwann From the question I'm assuming that we are dealing with the case where the starting point is 'I know this answer is out of date because ...'
    – Joseph Wright Mod
    Mar 30 '17 at 15:50
  • 12
    In some cases, this is strikingly ineffective. People read the top answer. If that answer is accepted, it is always top. The answers below are unlikely to garner more votes than the top one and cannot float to the top even if they do. People only bother reading the top answer, even if other answers make clear that they are updates. Even if an answer is not accepted, if it has enough of a head-start, the idea that the 'best' will float to the top is optimistic, at best.
    – cfr
    Mar 31 '17 at 1:01
  • 2
    @cfr's comment is especially true for topics that are not as popular with voters here. Especially if a new answer comes in way later than the original one it is going to have a hard time convincing people to use it.
    – moewe
    Mar 31 '17 at 8:12
  • @cfr That's a general criticism of the StackExchange model but I'm not sure what we can do: we have a relatively small number of regular voters.
    – Joseph Wright Mod
    Mar 31 '17 at 8:18
  • @JosephWright I know. I don't really know how it works in practice on other sites, but I'm aware it is a function of the SE model.
    – cfr
    Mar 31 '17 at 14:37
  • 2
    @moewe Indeed, I was thinking of certain Forest answers, in particular, where answers explicitly updating an earlier one have been posted, but follow-up questions almost always rely on the older code from the earlier, more popular and/or accepted answer.
    – cfr
    Mar 31 '17 at 14:39
  • 6
    The problem with the last approach is that an earlier (now obsolete) respose may have many upvotes. Meanwhile, the newer (and better) response lurks at the bottom. Since many users would have to reason to re-visit the question and vote, the better answer may permanently stay at the bottom. Suggestion: Can answers be sorted by "newest first"?
    – user103221
    Mar 31 '17 at 22:10
  • @RobtA Yes, it's one of the options you have when visiting any of the pages (active, oldest, votes): down to the person viewing the page.
    – Joseph Wright Mod
    Apr 1 '17 at 7:56
  • 2
    i tried out the options. an accepted answer is always listed first, regardless. "oldest" (apparently the default) does what it says, as does "votes", both subject to "accepted first". "active" seems to rearrange with newest first, with the same exception. so even an answer with many more votes than the accepted answer will never be listed first. a real pity. maybe we could request, and justify, another option?
    – barbara beeton
    Apr 1 '17 at 12:22
  • 2
    @clemens Yes, but this is a bit different. It is like handing somebody a manual in which the first chapters are all about some old, deprecated version of the package and all the currently recommended solutions are at the back. Moreover, the manual probably doesn't indicate this in anyway. And there are a bunch of awards and gold stars and user endorsements attached to the code at the front of the manual, but few or none attached to that at the back.
    – cfr
    Apr 3 '17 at 22:21
  • 1
    @alephzero IMHO if they want to “get their problem solved” but aren't willing to at least put some effort in reading what other people have (and already had in case of existing answers) to say to solve their problem and as a consequence end up with a non-state-of-the-art solution it really isn't my and shouldn't be our problem. All it takes is scrolling a bit down and invest some time! »Why would an efficient person even think about wasting their time reading the unaccepted ones« because efficient people know that not always the first solution they find is the best :)
    – cgnieder
    Apr 4 '17 at 14:37
  • 3
    @alephzero Always choosing the next best solution doesn't pay in the long run. It's inefficient. ;)
    – cgnieder
    Apr 4 '17 at 14:40

You must log in to answer this question.

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