Should recommendation questions be on-topic?

Question

The previous incarnation of Literature SE included a lot of recommendation questions. I don't have any specific examples to hand from that site, but typical book recommendation questions might look like the following:

• What are some other good 19th-century romance novels, for someone who enjoyed Jane Austen?

• Recommendations for science fiction short stories set on an aquatic model of Venus?

• Which is the best for someone new to detective fiction: Sherlock Holmes, Hercule Poirot, or Inspector Morse?

Should questions like this be on-topic for Literature SE?

Answer 1

No.

Book recommendation questions have a tendency to invite multiple answers, none of which is more correct than any other. Both the answers themselves and the votes cast on them will likely be based more on subjective opinion than expert knowledge or verifiable correctness. Such questions would fall under the "primarily opinion-based" close reason which applies on all sites:

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.

Admittedly there are at least two Stack Exchange sites which thrive on recommendation questions. I don't know how they manage it or what measures they have in place to avoid overly subjective questions, but past experience suggests that Literature specifically will be ill-suited to recommendation questions. We can and should learn from the experience of previous incarnations of the Literature proposal. Quoting from this answer from 2011, relating to an older proposal then named Books and later renamed to Literature:

One thing that worries me about the book proposal is that three of the example questions are book recommendations. We tried that on SF&F, and after a while we weren't so hot about them, because most recommendation questions were pretty bad (the questions were usually too broad, and the answers weren't very useful because they rarely explained why the answerer thought the proposed book was a good fit). [Ra'T note: recommendation questions are now banned on SFF, and have been for many years.]

Before that, Gaming had issues with recommendations and ended up rejecting them. It's a trend.

And a comment on that same answer from a member of that failed incarnation of Literature:

I'm not a mod, just a member of the Literature site. Guess what: The recommendation requests ARE a nightmare. They are relentless, numerous, and often very obscure or specific. I have no solution to propose.

One of the answers to Why did Literature fail? on main meta even suggests that recommendation questions were one of the main reasons Literature failed last time. That should be enough of a warning sign to tell us to disallow them this time.

Answer 2

Absolutely not.

Rand's answer is good, but I thought I'd add something.

I moderate Hardware Recommendations, wherein we deal with recommending computer hardware to people who ask for it. The one thing I can say about that job is - it's a massive handful.

After a year and a half of beta our scope has only just solidified, but we haven't yet managed to communicate it effectively. There's a problem with off-topic questions; there's a problem with questions being too broad. And there's a problem of people taking it rather badly when their questions are closed because of those things. I've learned some colourful new insults, thanks to my diamond there.

I'm saying this because it illustrates what a difficult problem recommendations in general are to solve. If a full site with a team of veteran moderators and an awesome base community can't get it right after this long, the likelihood of getting it right on a site where there are other types of questions being posted for people to edit, answer, and moderate is... low. Far too low for comfort.

Yes, I know they'd be nice - but it's not going to work. Maybe it's one to revisit down the line when the site is more stable in public beta, but we should absolutely not risk destabilizing a baby, private-beta site for the sake of "Hey, I think this book is cool."