Why do we close duplicate questions?

Question

I know that on AskUbuntu, we close duplicate questions. What is the reason that we do that? Is it to narrow the amount of questions that exist on the site that way users don't need to sift through 20 of the same question to find an answer?

If that is the case, why do we close? Why do not simply have these duplicate questions? That way, they won't be there to sift through.

So, my question is multi-fold:

What is the reason that we close duplicates? Why do we not delete duplicates?

Answer 1

My opinion only:

Duplicates get closed instead of deleted for several reasons.

• seo
• adding search terms without modifying the meaning of the original question
• having duplicates closed allows a central way to put all the answers in one place
  • (read as merge)

Here's a specific example where something got closed and merged to put an answer that merited being seen into the original question.

It was originally intended to be an ISO mount/burn question for Ubuntu only. A later question was asked about burning an Ubuntu ISO for Windows.

It makes more sense to have a single ISO burning question. It keeps the site clean and centralises canonical answers. This way we can have one place to point a user with that type of problem.

Answer 2

Closing duplicates also helps the following things:

Maintenance

We don't abandon or "archive" wrong information, we strive to always have that page be the relevant awesome answer for that particular question. In October Ubuntu had to change how we shipped Java.

It took us about 2 hours to update all the Java questions here, because it's centralized into a few questions instead of 15 variations of "How to install Java?". We were able to remove almost all reference to the wrong information referring the partner repository nearly instantly.

This is way more powerful and useful to users than blogs, forums, and mailing lists. We can make wrong information just go away and not pollute the internet. Search around other places for posts recommending the (now wrong) partner archive for Java, we can minimize problems like this.

Finding stuff

Check out this question:

• Desktop forgets theme?

And then check out the "Linked" on the right side. Those are all duplicates that were linked to this question. We're nerds, so we know the problem is gnome-settings-daemon crashing. Look at all the useful user-centric terms they used to describe the problem that now are nice landing pages that redirect here. "gray and boxy", "sometimes ugly when I log in", "looking like Windows 98". Each one of those is a sign post to this master question.

One thing that is not obvious to people is when a user is asking a question the site is automagically filling out the sidebar with questions it thinks are relevant. The more links and duplicates a question has the more information is available. The more links a question has also makes it show up in the autogenerate FAQ for the tag and the site, etc. and drives traffic to the one canonical answer and question.

Consolidation

I originally posted this on the G+ page but it's relevant here:

One of the common responses to having a question closed as a duplicate is "But the original question is unanswered!"

This seems callous, you're asking for help, and you see that the other person didn't get help, so we seemingly just shove you both into the same sinking boat.

It's not immediately obvious, but what we're doing is consolidating people with the same problem into one place. Imagine if everyone had the same problem and posted another question, we'd have 15 unanswered questions! Now we're worse off than before, now people who might know the answer are lost in clutter.

What we do instead is rally every one around the same question. This way it's in one place, it makes it easier to share, to tweet, and more importantly, to find when people are searching.

• Do you have any information you could add to the question?
• Does the question itself need improvements?
• Does the question have links to logs?

Is there more information missing? By having everyone improving one question instead of a bunch of half finished ones, it increases your chances of getting a question answered!

Deletion

I don't vote to delete it as it helps people find things unless it's a worthless bug report or something too localized that would just clutter the site.

Answer 3

If you delete a post that is duplicate of an original any information in it will be lost forever.

Ideally all duplicates should be merged in to single one and we should be left without answers in the closed post.

The site has recently enabled something called automatic redirection, that means that if an user arrives at a page via any search engine that is a duplicate of an more complete with better answers post, he/she will be automatically redirected from the closed post to the complete one. But for that the closed post has to have no answers.

Closing and merging allows the worthy answers from the closed post to be transported in to the more complete one without intervention from the users involved in posting them.

We normally try to save all good answers that give extra information, if we can't there is a chance that we delete a few so that automatic redirection works.

In the end it is just a way or directing the user to the right information without him worrying if there are other better / different answers out there.

In simple terms:

• Deletion = gone, useless, waste of time to all that posted anything there;
• Closing (and eventually merging) = redirection to proper information, happy users.

It's a lot of work sometimes and a pain in the butt when something goes bad, but it works and its a useful way of making users happy and information centralized on the site.