10

This question already has an answer here:

Recommendations are a recurring topic on multiple Stack Exchange (SE) sites.

Some sites allow recommendations (ex: boardgames), some do not (ex: gaming). There are many reasons for that and I can still remember animated discussions on this subject on the Gaming meta.

Then there is that existential problem with Sci-Fi SE, "What problems does answering a question on this site solve?"

The only common problem I can think of is What to read/watch next?, which obviously led to a big number of recommendation questions, later slightly disguised as list question. And the main trouble with these questions isn't objectivity at all, some of them aren't even subjective. It's the fact that they don't have an answer - they are open ended.

-Goran Jovicon meta.scifi

Those questions are monstrously popular because:

  1. There's an actual problem to be solved, find the next book I should read or which technical book to buy to learn something
  2. This problem is quite common.

In view of the Real Questions Have Answers blog post, a recommendation question fails the "real questions have answers, not items or ideas or opinions." criteria.

But it's still a "practical, answerable questions based on actual problems that you face. " and it requires expertise to answer.

The problem is that there are multiple acceptable answers, and the asker is not the best person to evaluate the best answer because he doesn't know which one is the best compared to the others, and he cannot test it in short term. Even the need to accept an answer is questionable.

So, is a recommendation question an acceptable kind of list-type question or should it be closed just like any other list-type questions?

|improve this question

marked as duplicate by gnat, Martijn Pieters, Danubian Sailor, Hugo Dozois, ɥʇǝS Dec 12 '13 at 23:05

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • In the end, a recommendation is pretty much a poll where the only option is a "fill-in field". I'm inclined to side with Aarobot's closer in your earlier question: I don't think we can come up with some universal ruling on this. – Grace Note♦ Jan 19 '11 at 17:22
  • 1
    @Grace, NOOOO, lists are not polls. Polls are looking for 'top' or 'best' answers, lists are looking for all answers without any of them having to necessarily be better than the others. – Lance Roberts Jan 19 '11 at 17:26
  • 3
    @Lance Lists aren't polls, but this isn't about lists. Lists can work pretty fine, especially if you can confine the contents into a single answer. The behavior of a recommendation, however, mirrors that of an itemized poll. Few people visitting a recommendation are looking for all of the answers. – Grace Note♦ Jan 19 '11 at 17:28
  • @Grace Note I was asking this question in relation with the "Real Questions Have Answers" blog post and the new argument that it bring to the debate. – DavRob60 Jan 19 '11 at 17:39
  • @Grace, why they may not realistically be looking for all the answers, they're not trying to find just the 'best' option, they're looking for a mass amount of information, from which they will cull the information they want, based on their own criteria (which is alway best stated in the question they're asking). A poll, like a political straw poll, is looking to rank the options. – Lance Roberts Jan 19 '11 at 17:45
  • @Lance If you can boil down a recommendation to 'seeking a list of options without caring that it's used to find what the author should find as best', then a poll boils down to 'seeking an ordered list of options without caring that it's used to find what the author should find as best'. A poll is just a more organized method of culling a mass amount of information - it's not that organization that makes it unwieldy. – Grace Note♦ Jan 19 '11 at 17:57
  • @Grace, I agree that poll is a subset of list. – Lance Roberts Jan 19 '11 at 18:13
10

Handling game recommendations - how can we solve these two problems of quality?

Summary

  1. Recommendations tend to perform poorly because nobody seems to care about making a good quality repository that is founded on knowledge.
  2. Because recommendations tend to cause voting patterns not to the scope of the whole site but instead by in-question popularity, votes on Repositories do not reflect the usefulness, effort, or quality of the contributions to the list as they should be.
|improve this answer
  • I don't suppose I could get you to summarize that link in a few short chapters...? – Shog9♦ Jan 19 '11 at 17:59
  • 1
    @Shog9 For its length, that beast really does just boil down to two sentences in bold. It's not even a complete analysis since it's just talking about quality issues. It doesn't even touch into other issues like scale issues (which are site dependent) and maintenance issues. – Grace Note♦ Jan 19 '11 at 18:04
  • @Grace: very good, I agree completely. – Shog9♦ Jan 19 '11 at 18:08
  • Great article.. – Lance Roberts Jan 19 '11 at 18:20
  • "nobody seems to care about making a good quality repository that is founded on knowledge." What would happen with Wikipedia, if we lock and delete all pages, that are poorly written and not updated for years? If you are annoyed that something is not up to date or written poorly then take liberty to edit it. – Calmarius Sep 13 '13 at 15:46
  • "votes on Repositories do not reflect the usefulness, effort, or quality of the contributions to the list as they should be." I would argue with this. The useful or good items generally have the best +/- ratio. – Calmarius Sep 13 '13 at 15:51
11

Frankly, I think folks tend to overrate the usefulness of even the best of these questions. Lists of books are frequently cited, presumably because nowhere else on The Internet can you find lists of popular books, and the "For Dummies" series has taught us that popularity == quality...

...But I'm being mean. Ranked lists are fun. Lists of movies, lists of songs, actors, books, personality types, clichés... They give us the opportunity to see where our tastes and experiences fit into the zeitgeist. They stand little chance of being kept up-to-date, and so any utility as a learning or shopping aid diminishes over time, but the entertainment/nostalgia factor lasts a bit longer - do you remember the '08s? I do...

So the real question then becomes, do you hate fun?

Personally, I've come to feel these questions should almost always be closed and deleted. Want a book recommendation? Look on Amazon. Want a list of favorite books? Read some blogs / join a forum / ask on Chat.

But if you do want to make these useful, then curate them: compile the list into a big CW answer and revisit it now and then to keep it up-to-date. Note that this utterly kills the poll / populist participation factor, which is intentional - your goal is to make the question answerable by providing an answer. When you tire of maintaining the list, and no one else steps in to help, then it's time for that question to die...

|improve this answer
  • Good points, A solution may be to lock them with only one Answer and to set them as community wiki to make them easier to edit. The only missing part is to "Lock them with only one Answer" – DavRob60 Jan 19 '11 at 18:01
  • @Dav: the easier way is simply to accept one answer, make it CW, and edit it. You'll need a moderator's help (or a lot of edits) for the CW portion. Locking is troublesome, as it prevents comments or voting on the question. – Shog9♦ Jan 19 '11 at 18:06
  • That why I said the locking part was missing, I was not referring to a locked question; just a feature that lock the "Your Answer" fields of the question so any new item NEED to be added into the unique answer. Commenting should be authorized. As you said, "this utterly kills the poll / populist participation factor, which is intentional" – DavRob60 Jan 19 '11 at 18:41

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