23

Does damaging an enemy ship reduce the amount of scrap found? had been closed as a duplicate of How can I maximise the amount of scrap I collect?.

While it's been reopened now, I'd still like to get the general stance on this - as per the Dr. Strangedupe blog post and probably many other blog and meta.SO posts, questions are only duplicates if they ask for precisely the same thing, not for specific concise subsets, correct?

  • We were just saying we needed a meta for this in chat. It's an issue that's come up on multiple occasions, but I don't think it's ever really been hashed out on meta. – Wipqozn Jul 2 '13 at 20:29
  • Because I am confident it is correct, I have updated the existing answer to address the new question. – user9983 Jul 2 '13 at 21:23
  • 2
    @OrigamiRobot, what makes you confident that it's correct? A number of other FTL questions refer to experiments done in the game or to examinations of the game's data files. What have you done to determine the correctness of the sentence you added? – Kyralessa Jul 2 '13 at 22:32
  • @Kyralessa I use this strategy all the time through different means (teleporting, fire beam, etc) and I have not seen a difference between one health left and max health left. – user9983 Jul 3 '13 at 2:20
  • 3
    @OrigamiRobot, and how would you see such a difference, given that the amounts are always random? – Kyralessa Jul 3 '13 at 11:07
  • @Kyralessa By observing said randomness. There is no pattern to indicate a correlation between ship health and reward beyond intact > destroyed. – user9983 Jul 3 '13 at 11:38
  • 2
    @OrigamiRobot What I am hearing is that the answer is more than one sentence, and is furthermore a matter of personal judgement and experience (which is valid; no backhand there). But if so, doesn't that undermine the argument that it's simple and too trivial to be a separate question? And don't we need users to vote on that answer and its methodology? – SevenSidedDie Jul 3 '13 at 15:52
31

One question is about general strategy. The other is about a specific mechanic that happens to be something that could be taken into account in an answer about general strategy.

These are not duplicates. We don't serve anyone well by trying to force them to be. The reality is that asking for updates to an old question for new info most often fails. As the answerer of that other question with 20+ up votes, I can say that I do NOT know the answer to the specific mechanics question, but that did not stop me from being able to leave a useful answer about general strategy. I will not be able to update my other answer to include the info he wants because it is much more narrow and focused.

Who, exactly, does it help to force this to be a duplicate? Who does it hurt if we don't? Lets not forget our goal is to help people. Yes, obvious duplicates should be closed. This is not an obvious duplicate. It is simply a related question. Lets not doom it to never receiving answer simply because a broader question could maybe kind of sorta cover it.

Duplicate information in answers (in part or in whole) does not automatically make questions duplicates.

  • No, duplicate answers do not, the scope of one question completely covers the other. – user9983 Jul 3 '13 at 2:19
  • 2
    @OrigamiRobot With that argument, just as my hyperbolic comment suggested, you might just post one sufficiently broad question that is not yet too broad to eliminate ~20 more precise questions, with the result that in order to actually solve them all the broad question will have to have a gigantic all-in-one answer - tl;dr... – Zommuter Jul 3 '13 at 6:15
  • @Zommuter And just as I have said, there is such a thing as too broad. If you're going to base your arguments on not reading anything I say, there's not much I can do to defend my point is there? – user9983 Jul 3 '13 at 11:40
  • 3
    @OrigamiRobot I read your statements, but you seem to have missed my exaggerating on purpose... – Zommuter Jul 3 '13 at 12:13
  • 5
    I like to emphasize this point by using math as an example. If someone asks 'what is 1+1' it gets the answer '2', and someone else asks 'what is 3-1' the answer is also '2', but just because the answers are the same, it doesn't automatically make the questions the same – RhysW Jul 5 '13 at 9:49
  • Since when do sane approaches actually get upvoted on meta, am I reading a parallel world's Arqade :-) I've given up on pointing out incorrect dup closes of this sort with how often I kept seeing it ... – Alok Aug 29 '13 at 22:10
11

There are two problematic issues when closing a specific question as a duplicate of a broader one:

  • If the broader one does not already answer the specific issue, even though it would be part of a complete answer, this looks like a pretty hostile move to the user asking the new question. We're essentially telling him that his question should be already answered in that other post, but it isn't and there is nothing except adding a bounty he can do to encourage a better answer. We're applying an ideal standard, where the broader question has a comprehensive answer that would help the asker of the newer, specific question, while in practical terms we're rather unhelpful and pretty much appear to refuse to answer the question.

  • If the answer is contained in the broader question, there is still the issue that it might be somewhat hidden among other information. In the worst case we're telling the use that his question is answered over there, but he has to search for himself where in those several pages the actual answer is. This issue applies also to all people that find the question via search engines, a specific answer to a specific question can be more useful than a comprehensive guide. If we hide the answers to specific questions in large guides, we're reintroducing the same problem of having to wade through a whole forum thread that Q&A sites were meant to solve.

Closing as a duplicate is probably a good idea if the specific question is already answered in the broad one, and if it isn't too hidden among other stuff. If it is unanswered or too hard to find in the more comprehensive answer, we should not close. The drawback of the duplication is minimal, we're causing more harm than we benefit in such cases.

  • 5
    Updating answers is an essential part of how SE works and should be encouraged at every available opportunity. Both of your issues can be solved by this. The first with updating content, the second by updating formatting. – user9983 Jul 2 '13 at 20:37
  • What's considered "too hard to find" though? I think that if someone fails to find their answer in an existing comprehensive answer, it's kind of more of a failure of basic reading comprehension than failure on our part to answer. "If we hide the answers to specific questions in large guides, we're reintroducing the same problem of having to wade through a whole forum thread that Q&A sites were meant to solve." I disagree with this. Our Q&A>forum system is better in that it's easier for us to filter out the wheat from the chaff. I don't think that one comprehensive guide answer is worse than > – FAE Jul 2 '13 at 20:40
  • 4
    @OrigamiRobot The problem is that the broader question has far less incentive to update and receives less attention than a newer, specific question. We're making it harder to answer the specific question by closing it. And in the cases I've seen this, while the specific question was a subset of the broader one, it was easily possible to write a good answer to the broad one without actually answering the specific one. Ther is possible overlap, but often not necessary overlap. – Mad Scientist Jul 2 '13 at 20:41
  • having to filter through "Me too!" "Thanks!" "Wtf this didn't work for me!" posts. At "worst", there are a few good answers to a an umbrella question, all of which will have more upvotes and be floating higher. In most other cases, there will be a larger umbrella answer in a more "guide" like form. I don't think that it should be too much to ask for anyone coming in to actually read the existing answers properly to determine if their question is answered there or not. – FAE Jul 2 '13 at 20:42
  • @MadScientist That is the case for any old question. We don't ask new questions about the same mechanics when they are changed by a patch just because the question is old. The new information being requested is a single sentence. – user9983 Jul 2 '13 at 20:42
  • 1
    @FAE I vaguely remember a D3 question where this happened, the specific question was answered in a single sentence of a full page of text. Someone arriving from Google won't be sure that the answer is there at all, a specific question is more useful to those users than a broad one where you have to search for the answer. – Mad Scientist Jul 2 '13 at 20:44
  • 3
    @MadScientist I'd expect anyone arriving from Google to still read the answer to see if what they're looking for is there. Not doing so and asking another question that falls under that umbrella seems like a very clear case of "poor research effort" to me, especially when the research is right there, they just decided it was too much effort to read. – FAE Jul 2 '13 at 20:53
  • 4
    @OrigamiRobot, you tried to close my question as a duplicate of another, even though that question's answer didn't answer what I asked. So you updated that answer with one sentence, but you didn't give any indication of why you believe it to be correct or what evidence there is for it, which makes it not much of an answer. While you're standing firmly and vehemently on some principle or other, the fact remains that my question hasn't been answered yet. – Kyralessa Jul 2 '13 at 22:40
  • 2
    @FAE The problem is not lazy readers, the problem is the more general an answer, the less utility it has. We optimise for utility. When someone says users will have a hard time finding the answer to the question, that is a statement about utility, clarity, and ease of finding answers quickly. Motive for wanting an easy, clear, and quick answer is irrelevant—it's just what we do. Yes, that helps lazy people, but it helps everyone else too. – SevenSidedDie Jul 2 '13 at 22:45
  • 1
    @Kyralessa As Sterno's answer says, answers do not determine whether it is a dupe or not. The question itself does. IMO, your question falls 100% within the scope of the prior question. – user9983 Jul 3 '13 at 2:23
  • 4
    @OrigamiRobot "Duplicate" means =, not >. Nowhere that I'm aware of does the word "scope" appear in any policy statement about duplicates. What I am aware of says that duplicates are the same question, nothing about general and specific scopes. Where does this "scope" idea come from? – SevenSidedDie Jul 3 '13 at 16:00
  • @SevenSidedDie I simply could not disagree more. – user9983 Jul 3 '13 at 17:07
  • 2
    @OrigamiRobot You disagree with the question I asked too? To repeat: where does "scope" enter into any policy or guidance about the software's duplicate voting functionality? – SevenSidedDie Jul 3 '13 at 17:31
10

We should not be closing as duplicate anything except exact duplicates. Closing as a "duplicate" of a different question that would not naturally attract the same information as the new question is actually harder, and less likely to create quality answers, than just answering the new question and linking the other as "Related."

It's also not prompt, and requires a bunch of editorial work that circumvents the answer-voting system to make sure the new question is properly answered. Using duplicate closing for only near-duplicates means embarking on this process:

  1. Collect close votes, possibly involving a lot of comment discussions or a meta post.

  2. Select an answer on the old question to be the new home of the correct answer to the new question. This introduces significant issues:

    1. Which answer needs to be edited? The accepted one? (What if the checkmark is later moved to a different answer? What if there is no accepted answer?) All answers? A new answer?

    2. The editing user is effectively deciding that this new info being edited in is correct. That's not their job; that's the voting system's job. This eliminates competition to submit the best answer to the new question.

  3. (Optional and usually not done) Edit the old question to specify that the info asked for in the new question is required in new answers, so that future answers on the old question properly cover the new question. This is necessary because open questions are never "dead" and can always attract and accept better answers than what's already there. Not doing this, the new question's needs are invisible and don't leverage our core system (voting) to attract the best possible answer.

    Of course, this step wouldn't be necessary if it was an exact dup. You can tell when it's not an exact dup when this step is needed to ensure that future answers to the old question properly cover the new question.

Compare that with this process for near-dups:

  1. Answer the question with what you think is correct and let the votes decide.
  2. Add a comment saying, "Related: [url]"

The latter is clearly faster, but more importantly, lets the voters decide what is a correct and good answer. The more convoluted our human processes are, the more prone they are to judgement errors and quality issues.

  • Exact duplicates -- the same question from the same user -- are merged (when both garner separate good answers). The duplicate functionality is different, and admittedly has been a bit fast and loose with the term "exact". If you're really arguing against an established network-wide mechanic then I think you need a bit better argument than "it requires some effort and thought". – Matthew Read Jul 2 '13 at 21:45
  • 4
    I agree with this answer 1,000% and will personally knife anyone who disagrees with me. I have knives; you've been warned. – agent86 Jul 2 '13 at 22:29
  • 3
    @Matthew This is the only site in the network I've seen "exact" played fast and loose. Elsewhere (admittedly not a large sample) I've seen "exact" used to mean "the same question", and I believe that is the original design intent. It's not "same person, reposting", nor is it "could be (but isn't) solved by a more general question". The latter seems to be how it's used here, and it seems to be the exception on the network, not the other way around. – SevenSidedDie Jul 2 '13 at 22:35
  • 2
    @agent86 Good thing I'm immune to knives! Because I disagree most vehemently with this. – Frank Jul 2 '13 at 22:38
  • @Matthew Oh, and regarding "effort and thought": the objection isn't "wah hard", the objection is "that's lots of additional opportunity for exactly the sort of human error the site is designed to eliminate". Does that make more sense? – SevenSidedDie Jul 2 '13 at 22:39
  • 8
    I think any time we edit the scope of a question to suit our own organizational desires we create more problems than we solve. If you want organized, concise data about a game mechanic, go to a wiki. If you want to ask a specific or general, messy or clean question about a game, ask. – agent86 Jul 3 '13 at 0:48
  • 1
    I actually tried to enact the suggested policy of editing the answers to a linked dupe question, and got chewed out by the person whose answer I edited, and the edits were rolled back. I then had the option of completely answering the question again to handle the previously unhandled subcase, or just giving up. I invested the extra effort, but I would much rather have done what this answer suggests - relate the two questions, and answer the specific case asked about in the new one. – agent86 Jul 3 '13 at 0:50
  • I agree that questions' scopes should not be edited, but this did not need to be done in this case. – user9983 Jul 3 '13 at 3:14
  • 1
    @OrigamiRobot I should be actually, else it will never attract new answers to the narrower question that may be better than the one sentence you believe is sufficient. In the current scope, a correct answer can (and did) entirely omit mention of this. – SevenSidedDie Jul 3 '13 at 4:03
  • @SevenSidedDie No, just because the question did not explicitly ask for this information doesn't mean it isn't already with its scope. The amount of damage doesn't matter. There is no need for a longer addition than that. – user9983 Jul 3 '13 at 4:16
  • 3
    @OrigamiRobot We don't deal with scopes, we deal with problems and solutions. The problem in question A was already solved adequately without solving the problem in question B. Therefore a solution to A is not a solution to B unless you change A to be A+B. – SevenSidedDie Jul 3 '13 at 4:39
  • @SevenSidedDie In that case, you better go edit every question who's answer give any more information than the question explicitly asks for. Or edit the answers so that they only answer what was explicitly asked. I could not disagree with that more. There is no reason to have this information separate. – user9983 Jul 3 '13 at 4:42
  • 1
    @Origami That is either an uncharitable reading of what I wrote, or a simple misunderstanding of the argument I put forward in my last comment. – SevenSidedDie Jul 3 '13 at 5:29
  • @OrigamiRobot What about this: Edit the broader answer and post a linked quote of that in the more specific question. That way both the short question is answered and anyone who wants to know more about the broader scope directly knows where to find it. – Zommuter Jul 3 '13 at 6:27
0

There tends to be at least two viewpoints when it comes to dupes like this. There's the specific question side, and there's the game mechanic side. This is what's coming into play here. We have the new question being specific, and we have the game mechanic question designed to handle the entire scrap mechanic.

The specific question type is generally what we get here from new users, as they have their specific problem with a game, and they're just looking for an answer. This is the narrower viewpoint, and is scoped precisely for the user. This is their exact scenario, and they're looking for a answer on how the game plays out in a specific use-case.

In the best case scenario, this hasn't been asked at all in either form, we can let it stand, and everyone's happy. The wonky bits happen when we get specific questions, and we already have a general question about that mechanic.

The game mechanic side looks at the entire mechanic behind the question. How does X work? What can I do to influence this to my advantage? Are there related mechanics I can game to influence and gain an advantage? This has no actual scenario attached to it, because it's looking for everything that influences this mechanic. In a perfect world, we'll catch every variable attached to the mechanic, and note how it's influenced by it.

Therein lies the clash between the two viewpoints. We have a question about a specific scenario, and how it influences scrap. We also have a general question about scrap, and how to maximize your gain of such. The specific scenario being discussed is encompassed inside this question, we've just had the bad luck of this variable not being addressed in the general question. If it had been answered, and wasn't buried, this wouldn't even be a problem. We'd dupe it and leave it.

When we get such a clash, I think what we should be doing is figuring out how much effort it would take to update our existing answers to take that scenario into account. If it would take two or three paragraphs or more to fully answer that scenario, I say don't dupe it. That amount of information can be extremely useful, and will help anyone else who needs it. If adding a sentence or two can account for that entire scenario, though, that's a modicum of effort to encompass the extra scenario. For that, I say dupe the specific question to the general mechanic one, and update the answer to include it.

Now, the question is, how much effort would it take to update our general question with this specific information?

  • 3
    The problem that I have with this is that close voters have to make a call about questions, answers, and effort that they are likely uninvolved in. They're essentially making a decision: "This other question and its answers should cover this topic, but they don't, so someone should do something about that." When often, it isn't their question (in either case) isn't their answer(s), and isn't their responsibility to do the work. It's heaped on the collective "community" who isn't really incentivised to do it. – agent86 Jul 3 '13 at 1:07
  • 5
    If the original asker (whose question was the dupe target) has their question edited, we're saying "your problem would be better if it was scoped like this, so says the community." That's not particularly friendly. Then, the answers to the question are now (by definition) incomplete. If an answerer has their answer edited to include additional information, they're now essentially responsible for that information that they did not provide or fact check, and it assumes the vote tally it was prior to the edit. – agent86 Jul 3 '13 at 1:10
  • 2
    Is it not easier to leave the new question open, and say in an answer "in addition to the great, community filtered answers to this related question, which serve as a foundation to my answer, here is my answer that addresses your specific case?" – agent86 Jul 3 '13 at 1:11
  • 1
    @agent86 Dupes are closed to prevent the fragmentation of effort in the same scope. That's why we have it in the first place. Part of our job here (volunteer, yes, but our job nontheless), is to build a knowledge base about games. If a minor amount of effort is all that's required to expand the scope (or, even, if we're discussing mechanics in general, no change to the question required), and add to someone's answer that extra information, everybody wins. If we're going for definitive answers, doesn't it make sense to include everything we can about that? – Frank Jul 3 '13 at 3:40
  • 1
    Isn't that why we allow collaborative editing in the first place? – Frank Jul 3 '13 at 3:40
  • I beg to differ with your point of view claiming this is a "new user with a specific problem" VS "established user with a more generally applicable problem" - not entirely, but I suggest the difference is more of the kind "specific problem anyone should from now on be able to find on google and solve quickly" VS "elaborate strategy guide for those who want to learn more". The two kind of users are not newbies VS establishes users, but more "quick question" VS "I'm completely lost" – Zommuter Jul 3 '13 at 6:20
  • 5
    If the scope of the original question didn't include a specific corner case, then I disagree that it's fragmentation to address corner cases in another question. I think in these cases a link between the two questions (as part of an answer, even) can keep us from fragmenting the previously invested effort. Trying to optimize fragmentation completely out is too high-cost compared to what issues it causes in the rest of the system. – agent86 Jul 3 '13 at 14:54
  • I thought collaborative editing was so that its easier to update / proof read / format answers (& questions), not to write shared wikis where even the older questions morph into something unrecognizable to the original asker. – Alok Aug 29 '13 at 22:17

You must log in to answer this question.

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