Return to Answer

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.

Experience on Stack Exchange sites about computer-related and creative content-related topics shows that while some recommendation questions can work, it's only a small subset of them that do work. (I speak from experience on Science Fiction and Fantasy, the defunct Literature site, Stack Overflow, Super User, Unix & Linux, Software Recommendations where I'm a moderator, and more.)

Recommendation questions tend to devolve into list questions, i.e. questions where answers are items rather than full answers. The Stack Exchange format is not well-suited to this kind of questions. Voting on answers doesn't work: answers float up not because they are good, but primarily because they came first, secondarily because the item they cite is popular, and only barely based on the suitability of the item for the requirements of the question.

A good recommendation must encourage answers that strive to be complete. The community must hunt and downvote or delete answers that cite a single item, no matter how good that item is, unless the answer validly claims that one should listen to such-or-such and nothing else.

Bad: I like X and Y, what other songs should I listen to?

Fine: I want to introduce people to this genre, this is their background culture, I'm looking for about a hour's worth of listening. What should I play for them? [And do make sure to delete answers that recommend a single 5-minute song as not answering the question.]

Moderating recommendation questions can be difficult, so my advice is to allow them with a stringent policy. (You can find the policy of the defunct Literature site here; the few threads where questions and answers met the policy brought out some excellent material, but it wasn't really enforced and the threads that failed the policy were mediocre.) If the community doesn't manage to consistently enforce the policy, then recommendation questions will be more trouble than they're worth, and banning them outright is better than letting all the bad ones stand for a 1% gem rate.