4

Discussed here in 2015, universally panned by the community: "Identify this song / genre / artist" - on topic?

[These questions] create a large amount of noise, with little value to anyone except the OP.

On Movies & TV site, identify-this-* questions make up 30% of the questions on the site, but 40% of those questions have a score of 1 or less.

On top of this, these questions tend to attract one-time users who don't give back to the community. 52% of the identify-this-* questions do not have an accepted answer. Part of this is because answers can only be guessed at based on vague descriptions, and part of it is that once a one-time user gets their answer, they move on without ever accepting.

Almost 11% of these questions never get an answer at all.

Data:

  • 7,240 total questions
  • 2,165 identify-this-* questions
  • 1,293 questions with score > 1
  • 1,144 with accepted answer
  • 234 with 0 answers

At the time of writing there are 1213 questions on musicfans.stackexchange.com, from which 380 (31%) are "identify something" questions.

  • 3
    And out of the identify-this-* questions, about a third of them are unanswered and make up the largest portion of the unanswered questions. – Dom Dec 30 '16 at 16:58
  • These questions are quite popular as you can see in the upvote ranking. For example, this question seems more liked than most of my questions about music history. – Bebs Dec 30 '16 at 17:04
  • @Bebs: At least he provides audio. I don't find the question interesting in the slightest, please note that "meme" images tend to get upvoted most on aggregators such as Reddit. – user2838 Dec 30 '16 at 18:59
  • @user598527, I don't thnik ID questions intersting either, but they often receive decent upvotes... more than history questions sometimes so we are in minority ;-) – Bebs Dec 30 '16 at 19:40
  • 3
    I can't call this an answer [so I won't] but 'identify this' questions are to music fans as 'what speakers/headphones/home hifi?' are to sound design. They are taking over, yet provide zero value to the site. – Tetsujin Dec 30 '16 at 19:44
  • I've updated the question to which you linked to clarify that these questions are currently on-topic. – user16 Jan 1 '17 at 18:29
  • 1
    @Bebs they are not very popular when you look at given votes. The most popular is identify-this-song which is the 5th most upvoted tag. – Dom Jan 1 '17 at 20:19
  • @Dom, I didn't say the tag is popular, but I can see that some questions (that I find useless) had been more rewarded that history questions. So I think some people might like ID questions... – Bebs Jan 2 '17 at 7:41
  • 1
    @Bebs it hasn't. The History tag has the most up votes of any other tags on our site. – Dom Jan 3 '17 at 6:18
6

It has been my experience that there is no single SE wide topic more controversial than Identification questions. Many hate them and many love them.

When the site first started, there was a loud majority of people who were dead set against them and a meta question to ban them was met with strong upvotes - and there was much rejoicing.

Then, inevitably, many of the initial enthusiasts left.

So, later on, a new meta question to allow them was asked and met with strong upvotes - and there was much rejoicing.

So that is the state of things - they are currently allowed. However, there is a rule in place that requires ID questions to have enough objective information to make it possible for someone to actually make the identification.

We tend to be soft on enforcing the requirement as ID answers sometimes come out of left field with surprisingly little information, but members of the community are encouraged to vote-to-close questions that they believe to have insufficient information. That's how a "community standard" works - the community sets the standards. Moderators are here only to enforce the established standard and resolve issues and disagreements.

  • "a new meta question to allow them was asked and met with strong upvotes" - do you mean meta.musicfans.stackexchange.com/questions/243/… ? – user16 Jan 1 '17 at 18:45
  • 4
    IMO, we need to start getting a better handle of this as they are swallowing the main page and not doing much good for the site in return. – Dom Jan 2 '17 at 17:04
3

I personally enjoy "identify" questions, and think they can be a value-add for the site as a whole, if they have enough information, and aren't simply links.

As things currently stand, they are a large percentage of our SE.

If we want to change that, what we need are more good questions on other topics. Banning "identify" doesn't make those magically appear.

0

The trivial / facile answer is that the tag isn't banned because the questions aren't currently banned.

As to why the questions aren't currently banned: I can't remember there ever being a strong desire from "the community" to have identify-this-song questions. My memory is that they were allowed as part of a more general 'relaxing' of on-topic rules at a time when the site was getting very little traffic indeed, and the active community had shrunk to a handful of users.

Since then,

  • participation and on-topic questions per day have increased...
  • ... but a lot of the incoming questions are identify-this-*

So if we ban identify-this-* or even a large subset of them, we're likely to fairly instantly go back to the 'not much meaningful participation' state, at least in the short term.

I don't think a constant trickle of identify-this-* questions gets us anywhere fast in terms of adding valuable content and building the community - but we didn't come up with a recipe for success without them either. I think before deciding whether to lock down these questions more, we'd need to have a look at ourselves and decide what we actually do want to be, if not a place for these kind of questions.

  • 1
    "So if we ban identify-this-* or even a large subset of them, we're likely to fairly instantly go back to the 'not much meaningful participation' state, at least in the short term." You're confusing correlation with causality here. That's not how it works. – Mast Jan 4 '17 at 20:45
  • 1
    @Mast i'm just saying something really simple... we're not getting many questions other than 'identify this", so if we don't allow those, we won't be getting many (allowed) questions at all for a while (unless we unexpectedly get a new stream of other types of question). No room for confusion there, i think... or am I misunderstanding you? – user16 Jan 4 '17 at 21:00
  • 2
    We are getting other questions, however they are quite hard to find due to the number of identify-this-* we get and keep having bumped to home page (we've had 6 id questions bumped the last week). The only people who will stop coming are the people who just want id questions which we already have a very low retention rate of. – Dom Jan 5 '17 at 20:37
  • @Dom If your perception is that there's significant non-identify-this-*activity that will be given a bit of room to breathe by getting rid of some or all of the id questions, I'm all for giving it a go. – user16 Jan 5 '17 at 20:55
  • 1
    BTW @Mast I'm still genuinely intrigued as to what you meant by your comment! – user16 Jan 5 '17 at 20:56
  • There's a difference between cause and effect. If X and Y both went up, you can't say X will go down if Y goes down. They aren't necessarily coupled the way you think they are. For all we know the participation will increase even further. – Mast Jan 5 '17 at 21:02
  • 1
    @Mast that's a reasonable thing to say if X and Y are distinct things. However, in my description above, I've indicated that 'Y' (identity-question-related-participation) is something I see as a component part of 'X' (participation and on-topic questions)... so obviously by that logic, if you remove Y from X, X will be smaller (at least for a time) unless there's some sudden other contributor to X. – user16 Jan 5 '17 at 21:26
  • 1
    By that logic, yes. But you have nothing to prove your inference with. Data which looks connected is not necessarily connected. – Mast Jan 5 '17 at 23:06
  • 1
    @Mast Data which looks connected is not necessarily connected. Sure, as a general point. But here, we are talking about data that is defined as a subset of a larger set of data, and is thus connected to that larger set by definition. It's not a question of it looking connected. This is sufficiently obvious that I think you may be winding me up TBH. – user16 Jan 5 '17 at 23:19
  • I only responded because you were genuinely intrigued. If your intrigue has been sufficiently satisfied I'll refrain from further comments. – Mast Jan 6 '17 at 0:00

You must log in to answer this question.