11
$\begingroup$

As a new user on Worldbuilding SE, I have to question the paradigm that once a question has been asked and the OP has accepted an answer, any future examples of that question must be closed.

My reasoning is this - First, the correct answer may have changed. Consider the following example:

Q: Do Gravitational Waves really exist?

A c.2015: Gravitational Waves are predicted by General Relativity, but we are not certain that they actually exist because they have never been detected.

A c.2018: Yes! Gravitational Waves do exist! They were detected in 2017 by LIGO.

Second, if a question was first answered some years ago, there have been many new users who have joined the site since then. Those users are unlikely to go back and answer old questions, and some of them may be able to provide much higher quality answers. By eliminating duplicates, we deny new users the opportunity to provide better answers.

So the question I have is: Why do we assume that closing duplicates is the correct course of action?

$\endgroup$
  • 3
    $\begingroup$ Welcome to the site Ark. If no one has pointed you to them yet, check out the help center and tour to familiarize yourself with the site. $\endgroup$ – James Nov 12 '18 at 21:22
  • $\begingroup$ Ironically, this question is itself a duplicate multiple times over. $\endgroup$ – Nij Nov 14 '18 at 4:13
  • $\begingroup$ Is it really? I searched for similar questions and couldn't find anything. $\endgroup$ – Arkenstein XII Nov 14 '18 at 19:24
11
$\begingroup$

Accepting an answer doesn't prevent new answers to be posted, nor prevent the accepted answer to be changed, should better options be made available to the OP.

Therefore, to use your example, the 2018 answer can be posted and accepted, while keeping a single source of truth for anybody who happens to search that question.

If we instead would let duplicate questions be, whoever would search for "do gravitational waves exist?" would find two questions, with two different answers, both accepted. Confusing, isn't it?

$\endgroup$
  • 2
    $\begingroup$ Sure, but do users actually go back and add new answers to old questions? $\endgroup$ – Arkenstein XII Nov 12 '18 at 20:30
  • 5
    $\begingroup$ Yes, there are even badges for answering old questions. $\endgroup$ – L.Dutch Nov 12 '18 at 20:35
  • $\begingroup$ Ah, interesting. That I didn't know. Might go for a trawl through old questions to see if I can improve any old answers! :) $\endgroup$ – Arkenstein XII Nov 12 '18 at 20:45
  • 3
    $\begingroup$ @ArkensteinXII yes, please do! The idea is to collect all the relevant answers in one place, and people can and do post new answers, update old ones, and move the acceptance mark. Even if the original asker is gone (and thus won't accept a new answer), the community can still show through voting that it is (now) the better answer. $\endgroup$ – Monica Cellio Nov 13 '18 at 0:20
  • $\begingroup$ @L.Dutch Yes, some users do that; but a question on the main page will get more replies than one from 2 years ago will get from users decidig to answer old questions. So it seems to me that the most beneficial approach is: leave duplicate questions open for a couple of days, and then have them automatically merged with the old question. Are there any downsides to this? $\endgroup$ – Ovi Nov 13 '18 at 2:39
  • $\begingroup$ @Ovi Just because it's a duplicate doesn't necessarily mean it's a good candidate for a question merge. (There's a continuum starting at completely original questions and ending at character-by-character copies.) Keeping duplicates open results in duplication of effort in the community, answering (effectively or exactly) the same question multiple times. Duplicate closures can serve as signposts for people searching using different choices of keywords. Answering an old question bumps that question to the top of the front page, so it gets (should get) renewed attention from the community. $\endgroup$ – a CVn Nov 13 '18 at 9:52
  • $\begingroup$ @L.Dutch So, the fact old answers are almost always not read and the first ones are always at the top isn't a problem then ? That's something you can see after reading a lot of SE questions : when questions are subjectiv, people tend to read the first two, upvote if they like them (often they do) and the others are just trashed. This doesn't work properly if the idea behind it is to give any good answer in the same thread. Most won't be read, as far as I can tell. $\endgroup$ – Kaël Nov 14 '18 at 15:17
  • $\begingroup$ I googled 'single definition principle', by which I actually mean 'single definitional repository' principle, but couldn't find one. in CS there are singletons, single responsibility principle, and referential integrity, but none of them precisely apply. At any rate, the practice of keeping a single referent to 'the truth' is the one that is apt. Keeping the newer dups around prevents any information loss, so no harm. $\endgroup$ – theRiley Nov 14 '18 at 17:50
  • $\begingroup$ @Ovi If you change or add a new answer to an old question it goes back to the top of the front page. Additionally the duplicate question will still be on the main page for a while and will link to the one it is a dupe of adding traffic to that one. $\endgroup$ – Tim B Nov 19 '18 at 9:49
4
$\begingroup$

An answer will make more sense when you understand the history of Stack Exchange. This whole glorious conundrum began with one site: Stack Overflow.

Stack Overflow is specifically for programmers. Somewhate by definition, questions about how to program are (or can always be made to be) objective, clear, specific, etc. Further, the answer to a programming question in a specific language is somewhat locked in stone. It should never change unless the language changes. Thus, closing questions as duplicates was a perfectly sensible way of keeping the site clean by pointing all of the myriad of ways a question can be asked back to the single best answer. It reminds quesion posters that they should be doing a bit of research (not just blindly posting yet another version of the same question on the site) and helps search engines (a big bonus for SE) to focus on the best answer. Ideally, if that rare case happens that there is a better answer in the future, that better answer should be posted to the original question, not a new one, to preserve the encyclopedic nature of the site.

Except this isn't always true.

I happen to have an answer on Stack Overflow (self-answered) that was hated by the mods and closed as a duplicate. It exists because I was bone tired of sifting through all the solutions, each having their own merits, that had been developed for the question over time. The site desparately needs a way of "rolling up" information to bring everything current. That stunt actually cost me (among other things) the privilege of posting questions for a while on SO.

So, bring this all forward to your question. We do it as a house-cleaning measure to try and keep the site clean. Regrettably, we frequently can't close a question fast enough to force people to post their better answers on the original question. Worse, we'll occasionally mark as duplicate a question that points back to a closed original — which doesn't help anybody.

But while there are rare reasons why this doesn't make sense, the reasons it does make sense are overwhelming. I therefore agree and fully support the behavior (despite having once intentionally violated it myself).

Ours is not a perfect world, and SE has not yet devised a way of rolling-up the info that time, well-intentioned users, and the best laid plans of mice and men have created. It's part of what we live with and work around. But it does have its value.

$\endgroup$
  • 1
    $\begingroup$ It is times like these where I wish there was a "reason to close" labeled as "Outdated information/No longer accurate, but it's been many years now and the OP hasn't posted since this question." A bit long-winded, but it'd be helpful. $\endgroup$ – Sora Tamashii Nov 13 '18 at 2:34
  • 1
    $\begingroup$ @SoraTamashii, that actually would be useful, on nearly every SE site. Data does get old, but is frequently listed in search engines first quite literally because of its age. $\endgroup$ – JBH Nov 13 '18 at 5:54
  • 1
    $\begingroup$ @SoraTamashii A question is supposed to ask a question. Closure is about questions, not answers. There is no support in the SE framework for "closing answers". (Questions can be locked, which prevents any action on any post associated with the question including the question itself, but that's very heavy-handed and not to be used for such a purpose.) Outdated or inaccurate answers is a reason to add a new answer, update old answers, or downvote existing answers for being wrong; not to close the question. Also, closure is about the question, not about the user who posted it. $\endgroup$ – a CVn Nov 13 '18 at 9:48
  • $\begingroup$ Well, using the situation in the OP, can you say what you would recommend, either here (although that gets dangerously close to conversational) or in an answer? $\endgroup$ – Sora Tamashii Nov 13 '18 at 17:19
  • $\begingroup$ @SoraTamashii, as I said in my answer: I recommend closing duplicates. Keep in mind, we don't own this site. Stack Exchange does. They set basic rules that we can work with, but not ignore. If we decided here and now to never close as duplicates, new users (who may not and rarely are new to other SE sites...) will come and and start closing questions. We have that problem with "Primarily Opinion-Based" where we have a meta page that defines our version of it - but we can't change the basic rules SE set. We're constantly fighting new users who haven't read the page. $\endgroup$ – JBH Nov 13 '18 at 20:25
  • $\begingroup$ I should have tagged. I meant a CVn $\endgroup$ – Sora Tamashii Nov 14 '18 at 0:56

You must log in to answer this question.

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