5
\$\begingroup\$

Sometimes a question "A" will be closed as a duplicate of question "B", where "B" is closed as a duplicate of a third question, "C". Thus the relationship looks like:

A -> B -> C

For a concrete example, consider:
A: Can I use a single resistor for multiple LED with different +ve sources?
B: Why do different colored LEDs interfere with each other when connected in parallel?
C: Why exactly can't a single resistor be used for many parallel LEDs?

Is this the commonly accepted way to close duplicates? I understand that Q&A quality, not age, of a question should determine which one is a duplicate of the other. In this case, C clearly has the best Q&A, judging by votes.

I can make an argument for any of the following strategies. Perhaps the correct strategy is one of the following, or perhaps there is no one-size-fits-all approach.

1. All other questions (A, B) should be closed as a duplicate of the highest-quality question, C.

This allows other users to clearly tell which question has the best answers at a glance. Users are not directed through a chain of duplicates, where such a chain could be confusing to new users. Finally, this helps collect all the good answers in C.

2. Newest questions (A) should be closed as a duplicate of both B and C.

Navigating to question A gives a link to both B and C. Why wierd capacitance values is an example of such a question with multiple duplicate targets. A user is given access to a (potentially) large variety of answers.

3. Keep as-is.

New questions (A) can be marked as duplicate of another duplicate (B). Similarly to argument 2. above, this gives a user access to a variety of potential answers, but in a different manner.

Is there a canonical best approach to these duplicates? Or, is there guidance in choosing which one of these approaches fits a given situation?

\$\endgroup\$
  • \$\begingroup\$ I don't have an example at hand, but sometimes A is quite specific and best answered by answers to B, while B itself is generic enough to be closed as a duplicate of C. Indeed, such chains of duplicate questions should be as short as possible, but not shorter. \$\endgroup\$ – Dmitry Grigoryev Mar 20 at 15:24
4
\$\begingroup\$

Option 1 should apply to any new questions. But we're not going to go back and re-re-reference old ones.

Option 2 would only be used if two questions that are still open have relevant material.

Option 3 applies to any current questions.

If more than one question has useful answers, they can be merged together. Flag it for a moderator to handle.

\$\endgroup\$
  • 2
    \$\begingroup\$ I take it back. I didn't realize that was possible, but yes, a question can have any number of duplicate links. \$\endgroup\$ – Dave Tweed Mar 14 at 19:17
4
\$\begingroup\$

Never close duplicates as dupe to a duplicate. Find the highest quality post on the topic and use that one as "canonical duplicate", meaning all dupes should point to that one.

Technical quality of the question and answer is all that matters. Keep the highest quality post, which is not necessarily the oldest one.

( As a side note, "dupe chains" can be a major pain when (user) moderating. If you have A -> B -> C and then need to delete B for whatever reason, it isn't possible as long as A or other posts link to it as dupe. In order to get rid of B, you would have to find all posts linking to it, then either get rid of those or close them as dupes to some other post (like C). )

\$\endgroup\$

You must log in to answer this question.

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