How do we approach closing dupes of dupes?

Question

Every so often, as I'm going through the review queue, I notice that there are close votes to question as duplicates of other questions which are already closed as a duplicated. I remember one time that I had to click through two differently closed links to get the 'parent' question and solution. The closure was something like this:

A (closed) --> B (closed) --> C (open)

However, I personally have always stuck to the notion that duplicate questions should be closed like this:

A (closed) --> C (open) < -- B (closed)

The reasoning behind this is that it's more compact and easier for inside/outside traffic to navigate to the correct solution.

Which of these approaches should be taken, when closing duplicate questions?

Answer 1

Why not have the best of both worlds!

With one of the newer releases of Stack Exchange, if multiple questions are linked in the VTC process, then all links are displayed:

This allows for users to multiple questions that may be related to the closed one, even ones that are already closed.

Answer 2

Note this has been brought up over at MSO already, and they advocate closing with linking to the original

I think there are a few things here to be considered.

What is the definition of "Duplicate Questions"?

From the close menu:

This question has been asked before, and already has an answer

In other words a similar question has already been asked, AND the solution found to the 'parent' question ALSO solves the 'child' question. This means, that if problems A & C have the same solution even if problem A seems completely unrelated to problem C then they are duplicates.

Ultimately the solution should always be found within the parent question. If not then:

• They should be merged if the questions are indeed asking the same thing, and viable solutions are found between the two, or
• The aren't really duplicates.

Which is better?

With this in mind, I believe that the second option is optimal because it becomes more concise to traffic and finding the solution. For example, if user lands on A, then there is only one more 'click' until he reaches the solution found within C. Same thing with B.

But what if there is some useful info in the child duplicates, that can be found by traversing the "chain"?

Handling this situation should be done with discretion. Very rarely are there times, where pieces of information found in question A, B & C lead to the solution. In this case, there are two thing that I see:

1. Again a merge of information would be best if the information to be gleaned is large enough
2. There is the "linked" questions that users can see where the parent question is linked from