During 4 years on Stackoverflow I've noticed that over time I got far less answers to questions and instead more and more comments like "this shouldn't be here" or no answers at all. Now being compliant with the regulations has become far more important than helping people with their problems.

There's a huge class of problems that is known to be unsolvable by a simple code snippet so suggesting a library is the only possible answer. There are thousands of questions like this on SO. In such a case what is the difference in asking for a 'solution' or asking for a library?

Many questions like this have been closed even though they get lots of upvotes - which means this is something people want ask and get answer for.

An example is question from today: https://stackoverflow.com/questions/21028484/how-to-upload-file-to-string?noredirect=1

  • 3
  • 1
    So I went to the link you provided... SO users prodded you into putting more thought into your question by enforcing the site guidelines, and you responded by eventually editing the post. This will ultimately help you by focusing the scope of any answers that come by. What's the problem here? – Anthony Neace Jan 9 '14 at 21:10
  • 3
    A bunch of similar questions, or duplicates of those, come up in search. In short, closing questions is important to the health of the site. The fact that low quality questions aren't answered is the reason that answers on the site tend to be of such high quality. If we answered everything, the answer quality would devolve to that of the competition, which is often quite poor. – Servy Jan 9 '14 at 21:10
  • 8
    Fun fact: Stack Overflow gets 7,000+ new questions every day. It has 11 million answers in total. For the most part, you're seeing more "this should not be here" comments because there are so many questions that really shouldn't be here. – Pekka Jan 9 '14 at 21:17
  • The best method to measure quality of answers is upvoting/downvoting not by closing/holding questions. There are many questions that are more than OK with the rules, but still have very poor quality - what happens most often is that if the question is not a very common problem it gets poor answers. Still they pollute the search and nothing is being done with it. Rules are quite easy to apply, but I can't agree that applying them ultimately makes the site better - im many cases you need something more than rules to tell if the question/answers are good - your brain. – kubal5003 Jan 9 '14 at 21:17
  • 1
    2k rep help vampire ಠ_ಠ – user1228 Jan 9 '14 at 22:00
  • If asking a question once a month is being a help vampire then yes I proudly am one. – kubal5003 Jan 9 '14 at 22:30
  • 1
    @kubal5003 Being a help vampire is more about the quality of your questions than the frequency at which you ask them. It's possible to ask great questions frequently or very poor questions very rarely. (I say this without having looked through your questions, for the record.) – Servy Jan 9 '14 at 22:33

For what it's worth, I don't happen to think there's anything inherently philosophically wrong with a site allowing requests where the right response will be a library recommendation.

But on SO, they became problematic as the site grew.

They worked well, and helped a lot of people in the early days of SO, and similar questions work fine on some other, smaller sites.

And, to be fair, some of them aren't that opinionated - you're looking for any library that does X, and there are only 2 out there.

But some of them are VERY opinionated. There are tons of libraries that meet the criteria. Worse, a lot of people have highly vested interests in which one is the top answer. The problem today is that they became unmanageable as the site grew. They do often attract spam and cause other problems, and over time, the community felt like they distracted more people who had to deal with spam or sort through tons of "which is better" to find the kinds of questions they were looking for.

At some point, they failed the key question we want communities to ask:

Does allowing these net help more people in the long run than making them off topic will?

In younger communities, I encourage allowing them until they see an actual problem, but on SO, they seemed to be creating enough distraction per person helped that the net effect was negative.

  • I agree with your statement. The problem is that there are no younger communities right now or at least it's hard to find them. Stackoverflow has become the one and only site that programmers use. That is why limiting things here essentially means limiting them everywhere. It's true that there are many people very interested in getting their answer the most prominent just because they are running a company that sells that product. This however can get easily verified by voting. An example could be - Q: Is it possible to write for Android in C#? A: Xamarin. Good or bad answer? – kubal5003 Jan 9 '14 at 21:33
  • To be clear: I'm not related to and have never used Xamarin - only heard about it. I've done the search and there is in fact a question like this on SO: stackoverflow.com/questions/10962104/… It got perfectly verified by voting! – kubal5003 Jan 9 '14 at 21:38
  • For a good example of how tool / library recommendations tend to go bad here, see this question (deleted, visible to 10k SO users): stackoverflow.com/questions/3049868/convert-pdf-to-word-offline . 8 out of the 9 answers there were spam, all from different users spamming different solutions. – Brad Larson Jan 9 '14 at 22:25
  • Unfortunately I can't see this and I fully support closing & deleting questions like this if they go bad but not because they can go bad – kubal5003 Jan 9 '14 at 22:33
  • 3
    @kubal5003 And the past experience of this site has shown that the percentage of times they go bad is so high, and the value gained from those that don't is sufficiently low, that it's not worth the considerable effort to deal with them at all; it's much easier to just make them go away. Your proposal would mean 1) drawing a line for when a question is "sufficiently bad" 2) constantly checking back on questions to see how they're doing 3) dealing with the "but they asked the same type of question and it's not closed" response to closures, etc. – Servy Jan 9 '14 at 22:35
  • The line is already drawn - not neceserilly bad questions go away just because it's easier to make them go away and easier to draw the line here. That actually answers my original question very well. Now I know what went wrong. – kubal5003 Jan 9 '14 at 22:48

I think we've covered this ground quite a lot but let me reiterate:

Stack Overflow is not the site for tool/resource requests. I believe it states in the close reason some of the reasons why this is off-topic:

Questions asking us to recommend or find a tool, library or favorite off-site resource are off-topic for Stack Overflow as they tend to attract opinionated answers and spam. Instead, describe the problem and what has been done so far to solve it.

Spam answers and opinionated answers are not the type of answers we want here. Recommendation questions almost always end up with spam and opinionated answers. Therfore, we close these questions as to not attract bad answers. This is to make sure that we still provide high-quality questions and answers to the community.

  • That is very true in many cases, however there are many others for which finding a right solution/tool is more than not trivial. In those cases reasoning behind those rules does not apply. It may apply later - if more than one solution becomes available. That is why I would definitely go for the model - see what happens and eventually close and not close and forget. – kubal5003 Jan 9 '14 at 21:23

One of the reasons for the shift is that these question are almost always fully-answerable via google. When there are libraries out there for a job, there are web pages that describe them, web pages that decry them, and web pages that ... oh, you can think of another verb that starts with 'd'. So it can't be that people need to ask on stackoverflow.com to discover the existence of these resources.

That leaves the evaluation of the resources. Here we have a gigantic bubbling cauldron of subjectivity and debate. One man's fish is another man's poisson, especially in distributed systems. We're not really helping anyone to provide an, ahem, forum for everyone and their pet goldfish to post an opinion on the 'best' library for parsing phishing email responses, or even the library best suited to a particular message.

Add it all up, and the sum is 'off topic'.

You must log in to answer this question.

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