What can be done about repetitive questions?

Question

Some questions tend to get asked in thousands and thousands of permutations.

One such type of question is questions on Apache's mod_rewrite. This list of questions tagged .htaccess catches most (not all) of them. There are three "please write my rewrite rule for me" type questions today alone. My feeling is that one such question ("I want this and that kind of URL to be redirected to this and that file") comes up every day at least once.

Other examples are:

• "Headers already sent" (PHP)

• How do I let a container wrap around floated child elements? (HTML)

• I get garbled special characters on my web page (HTML / (my)SQL / Unicode)

• And hundreds, if not thousands more.

They are always very, very specific, so they can't really be closed as duplicates of anything else. Still, essentially, one quick run-down with a few examples would do for many of these. Instead, they usually get answered after some time. Because of their specificity, and their vast numbers, it is my opinion that many of these questions have little value for future generations.

What do you think?

Is this being perceived as a problem?

Does this warrant a whole new kind of close reason with a link to a FAQ page or reference question or something? That's what floats in my mind as the right solution.

Do we need a master question for these to close the other questions as a duplicate of?

Bounty

starting a bounty to incite some more discussion.

Answer 1

1. Make one generic question on the topic.
2. Answer with a concise, clear and awesome guide, with headers, examples and anything.
3. Accept the bestest answer
4. Put a link to it into the relevant tag wikis, for what is worth.
5. Turn it community wiki and bless it with a red "featured" tag. (← mod assistance reqd.)

Now answer the other questions by pointing to the relevant bit of that one "wiki."

---

Why bother?

If people come up with a question all the time, it means there's no one authoritative, clear enough source of information on the topic.

Just make one.

Answer 2

I'd like to exemplify with:
PHP Warning Cannot modify header information

The meta topic here would be that it's caused by language misdesign, but anyway it's a real issue. The web and most bulletin boards are polluted with question like this, and despite better base design even Stackoverflow is affected. Two solutions:

Systemic Prohibition

Let's be honest. The influx of these questions is caused by laziness. Posting a new question on Stackoverflow always brings up matching duplicates for "Warning Cannot modify header information". Some askers just refuse to use that feature and the search function. Or expect other people to debug for them, explain basic language features or want prosaic recitations of the manual.

Anyway, filtering might help here. A simple regex check on question and question title should suffice to detect the majority of these excessive duplicates. (We'll have to make a list though.) And if the new question form detects it, it should just bring up an extra confirmation box:

The system detected that your question is most likely a duplicate. If you have used the search function before and checked out some of the suggested links, then check this box [x]. Otherwise please do so now. Your question would likely be closed without answers if it's a repetitive duplicate.

Well, it works for spammers. And I assume it's at least likely to reduce the amount of annoying duplicates. So, let's not postpone the list making.

[
   "/Warning:?\sCannot\smodify\sheader\sinformation/": "https://stackoverflow.com/search?q=%5Bphp%5D+Warning%3A+Cannot+modify+header+information",
]

.

Evil Plan B

This is not an proposal, nor an announcement, but it's not a fun answer either.

Be realistic. Excessive duplicate questions are actively encouraged by the SO commmunity. Whenever one of those boring, stupid or annoying duplicate questions is posted, 10 people jump immediately to the rescue and answer it. (While the more interesting questions seemingly go ignored on weeks.)
But that's only natural. Gaining reputation is fun, and that's why people flock to simple (and sometimes dull) questions.

However this very human behaviour draws boring askers to Stackoverflow. If you reward them, they will come. And closevoting questions is no serious discouragement. Likewise are admonishing comments no functional repellent. Therefore I charily suggest "censoring" the overhelpful answers on useless duplicates. If a duplicate is closevoted quickly enough it's redundant. But if someone answers it, it's imperative to kill off those answers on the spot. I'd use something along a standard replacement template like:

Search for [php] Warning Cannot modify header information
The question is a duplicate and has been asked too many times before. +1 for effort on this answer, but links should be sufficient here.

Needs rewording. It partially punishes the answerer, but should make it clear that the purpose is censoring the stupid question. And one needs to take care not to wipe the sporadic extraordinarily good answers, and whatnot. If the answerer is appalled by this method, a discussion may ensue, or he can just revert it back if it seems unjust.

Yet, I think this is the only realistic deterrant against <PROBLEM OF THE WEEK>. A real stop to answering stupid duplicates will send a clear message. A blood font might also help.

Answer 3

I'm going to take a pessimistic view on this topic, seeing as how it has been covered many times in one form or another and never truly resolved...

Q: What can be done about repetitive questions?

A: Absolutely nothing.

I think there are too many "human factors" involved for this to be dealt with effectively. Here are some:

• Users that are lazy: Even if you made the most comprehensive FAQ to deal with an oft-duplicated question or set of questions, the user still has to find and read it! This requires them to search first (good luck with that) and even if they did that they would find a mammoth FAQ, and they would probably look at it and say "Geez, TL;DR. I'm just going to ask my question." What's worse, if there are already a slew of duplicate questions present, and a user does a search and gets a huge list of them, then that list could also probably elicit a "Geez, TL;DR." since they might find it easier to ask a question than read through a whole bunch for their answer.

• Users that think they (and their problems) are unique: When a user asks a duplicate question, you can often see the comments devolve into hair-splitting discussions like "I used the word "thinga" and the other question used the word "dongdong", so my question is different." Even if the core problem (and it's solution) are the same, they still take offense to the idea of their question being closed. Even if the answers they've received thus far are just rehashes of the answers to the other question, they still think their question is unique and deserves to stay open.

• Lower-Rep users (Rep < 3000) that don't help: Obviously, if a user doesn't have the ability to close yet, they are limited in their ability to deal with duplicates. They could post links to duplicates in comments or upvote comments listing duplicates, but they're new enough to SO that they probably aren't very familiar with what the most pervasive duplicate questions are anyway. They may eagerly answer thinking that it's (somehow) the first time that an easy question like that was ever asked. Oh how wrong they are.

• Mid-range users (3000 < Rep < 10000) that don't help: They have the ability to close duplicates now, but why waste time on this when they can easily gain Rep by answering the question, bringing them that much closer to partial mod privileges and nerd fame? They've probably had enough experience on SO to know what the oft-duplicated questions are, and they probably know how to find them easily. But hell, they could just copy from answers on the duplicates. Easy Rep is easy!

• High-Rep (Rep > 10000) users that don't help: You've got a lot of experience on the site. You know what the common duplicates are, and you can find them, but you answer anyway. Sometimes you even answer the duplicate by brazenly posting a link to your own previous answer on the older question! And you don't vote to close duplicates much either. How is this helping the site? As a high-Rep user, you are more visible than others. New users will often pay more attention to your behaviors and accept them as the norm (e.g. "If high-Rep user X can ask a question like Y, why was mine closed?"). If you answer duplicates, not only are you setting a bad example to the greater community, but you're reinforcing the idea in the head of the question asker that it doesn't matter if they search first or not, they'll still get an answer from someone. Then again, maybe you just don't care about this or think it's a big deal. And that's your prerogative, but I don't personally find it all that helpful.

So, there are a lot of things that may lead users to disregard duplicates. Laziness, pride, ignorance, greed, apathy... this is why questions will keep getting duplicated with minor variations time and again.

TL;DR...

Honestly, I don't think there's much more we can do about it, since we're kinda fighting against human nature here. Just keep trying to close them when you find them (and when you still have votes left), ask the mods for merges, and if you want to get into a fight with those who think that every duplicate is special, you can try to thin the herd of duplicates by voting to delete some.

...and yes, I see the irony in answering a question that is itself quite repetitive. ;)

Answer 4

They are always very, very specific

So close them as "too localized".

Answer 5

The suggestion I have (well wish) is being able to close a question with "too many similar permutations, figure it yourself". As I feel providing answers fo these type of questions (and a signficant amount of questions on SO) are rewarding bad behaviour, how are people meant to learn to problem solve if they keep being provided with the answer.

Answer 6

How about allow users (with appropriate rep level) to flag as similar (instead of dupe) and post a link to 'propose' as the original question asked that it is a dupe of. Then if approved by say 3 users (again of appropriate rep level) and/or a moderator, the question gets closed as a duplicate and a link to the question that it is a dupe of.

Important to note would be comments and answers that were posted to the dupe question should be listed under the question that is the approved original with a link to the dupe question. That way when users view the original question they can see all answers there including the ones from the dupe but can still look at the duplicate answers in context by going to the dupe question.

Of course still allow votes to restore the dupe question back to non-dupe if the appropriate number of votes by appropriate rep level users and approval accordingly. However, I think the questions should remain linked since they are obviously similar.

I think it is important to preserve the original context of the dupe especially for restoral purposes. Furthermore, this process would allow linking of questions as similar, but not exact dupes (unless they get approved as dupes).

Answer 7

If a question is closed as a duplicate, given no reputation bonus, or a reduced reputation bonus, to those who answered the question and got their answer up-voted or accepted.

This will reduce the perverse incentive of the current system, which rewards people who fire off a quick answer to the all-too-familiar question, rather than put in the greater effort to find the question it is a duplicate of and to vote to close.

Answer 8

Honestly, I don't see a problem with duplicate questions. Seriously.

Every question is likely to have slightly different context, and that context can help people that other similar questions might not. In addition to that, SO is about helping people get their questions answered, not simply building a wiki-like knowledge base of only one answer to any given kind of question.

Finally, there may be a good answer that comes out of a later duplicate question that is better than all the previous ones.