Return to Answer

"Duplicate of" should be a transitive property (if A is a dupe of B, and B is a dupe of C, then A must be a dupe of C)

Assumption of transitivity looks wrong; per my reading of the definition in duplicate-questions tag wiki, it is not required:

we want to tell the user ...something like, “Somebody already asked this. If that other question doesn't solve your problem, please clarify your question to explain how it's different.” Perfect: if the other question helps them, they're happy because they got an answer. If the other question doesn't help them, they know exactly what to do. No argument about how exact an "exact duplicate" needs to be...

Question A having an answer in B does not automatically guarantee that it also has an answer in C, even though question B may have its answer in C.

For a simplified example, imagine "chain of duplicates" like as follows:

• A: How do I see result of multiplying 2 by 2?

  • B: How do I multiply?
    answer:
    You invoke multiply method of arith package. In order to view result, you invoke method of display package.

    • C: How do I do arithmetic operations?
      answer:
      You invoke appropriate method of arith package, like multiply, divide etc.

It's easy to imagine readers being satisfied by closures A -> B and B -> C since answers given solve the problems asked about.

However, closure A -> C would be problematic: first question asks about how to display results, while answer in the second question says nothing about this.

^{Note how question in B does not ask about displaying results, making dupe closure look reasonable. Also note how answer in B provides guidance beyond its scope, along with addressing the question asked.}

For a more realistic example, refer to the non-transitive chain of duplicates at Meta Stack Overflow:

• A': Spiteful downvoting? Two fast DVs to my unrelated posts

  • B': What can I do if I'm the victim of serial downvoting?

    • C': What is serial voting and how does it affect me?

A' is closed as a duplicate of B', which is in turn closed as a duplicate of C'.

If you take a closer look at revisions history and comments, you will notice that initial closure has been changed from A' -> C' to narrower one, A' -> B' because readers disagreed that widest dupe target has answers relevant to the question asked in A'.

"Duplicate of" should be a transitive property (if A is a dupe of B, and B is a dupe of C, then A must be a dupe of C)

Assumption of transitivity looks wrong; per my reading of the definition in duplicate-questions tag wiki, it is not required:

we want to tell the user ...something like, “Somebody already asked this. If that other question doesn't solve your problem, please clarify your question to explain how it's different.” Perfect: if the other question helps them, they're happy because they got an answer. If the other question doesn't help them, they know exactly what to do. No argument about how exact an "exact duplicate" needs to be...

Question A having an answer in B does not automatically guarantee that it also has an answer in C, even though question B may have its answer in C.

For a simplified example, imagine "chain of duplicates" like as follows:

• A: How do I see result of multiplying 2 by 2?

  • B: How do I multiply?
    answer:
    You invoke multiply method of arith package. In order to view result, you invoke method of display package.

    • C: How do I do arithmetic operations?
      answer:
      You invoke appropriate method of arith package, like multiply, divide etc.

It's easy to imagine readers being satisfied by closures A -> B and B -> C since answers given solve the problems asked about.

However, closure A -> C would be problematic: first question asks about how to display results, while answer in the second question says nothing about this.

^{Note how question in B does not ask about displaying results, making dupe closure look reasonable. Also note how answer in B provides guidance beyond its scope, along with addressing the question asked.}

For a more realistic example, refer to the non-transitive chain of duplicates at Meta Stack Overflow:

• A': Spiteful downvoting? Two fast DVs to my unrelated posts

  • B': What can I do if I'm the victim of serial downvoting?

    • C': What is serial voting and how does it affect me?

A' is closed as a duplicate of B', which is in turn closed as a duplicate of C'.

If you take a closer look at revisions history and comments, you will notice that initial closure has been changed from A' -> C' to narrower one, A' -> B' because readers disagreed that widest dupe target has answers relevant to the question asked in A'.

"Duplicate of" should be a transitive property (if A is a dupe of B, and B is a dupe of C, then A must be a dupe of C)

Assumption of transitivity looks wrong; per my reading of the definition in duplicate-questions tag wiki, it is not required:

we want to tell the user ...something like, “Somebody already asked this. If that other question doesn't solve your problem, please clarify your question to explain how it's different.” Perfect: if the other question helps them, they're happy because they got an answer. If the other question doesn't help them, they know exactly what to do. No argument about how exact an "exact duplicate" needs to be...

Question A having an answer in B does not automatically guarantee that it also has an answer in C, even though question B may have its answer in C.

For a simplified example, imagine "chain of duplicates" like as follows:

• A: How do I see result of multiplying 2 by 2?

  • B: How do I multiply?
    answer:
    You invoke multiply method of arith package. In order to view result, you invoke method of display package.

    • C: How do I do arithmetic operations?
      answer:
      You invoke appropriate method of arith package, like multiply, divide etc.

It's easy to imagine readers being satisfied by closures A -> B and B -> C since answers given solve the problems asked about.

However, closure A -> C would be problematic: first question asks about how to display results, while answer in the second question says nothing about this.

^{Note how question in B does not ask about displaying results, making dupe closure look reasonable. Also note how answer in B provides guidance beyond its scope, along with addressing the question asked.}

For a more realistic example, refer to the non-transitive chain of duplicates at Meta Stack Overflow:

• A': Spiteful downvoting? Two fast DVs to my unrelated posts

  • B': What can I do if I'm the victim of serial downvoting?

    • C': What is serial voting and how does it affect me?

A' is closed as a duplicate of B', which is in turn closed as a duplicate of C'.

If you take a closer look at revisions history and comments, you will notice that initial closure has been changed from A' -> C' to narrower one, A' -> B' because readers disagreed that widest dupe target has answers relevant to the question asked in A'.

"Duplicate of" should be a transitive property (if A is a dupe of B, and B is a dupe of C, then A must be a dupe of C)

Assumption of transitivity looks wrong; per my reading of the definition in duplicate-questions tag wiki, it is not required:

we want to tell the user ...something like, “Somebody already asked this. If that other question doesn't solve your problem, please clarify your question to explain how it's different.” Perfect: if the other question helps them, they're happy because they got an answer. If the other question doesn't help them, they know exactly what to do. No argument about how exact an "exact duplicate" needs to be...

Question A having an answer in B does not automatically guarantee that it also has an answer in C, even though question B may have its answer in C.

For a simplified example, imagine "chain of duplicates" like as follows:

• A: How do I see result of multiplying 2 by 2?

  • B: How do I multiply?
    answer:
    You invoke multiply method of arith package. In order to view result, you invoke method of display package.

    • C: How do I do arithmetic operations?
      answer:
      You invoke appropriate method of arith package, like multiply, divide etc.

It's easy to imagine readers being satisfied by closures A -> B and B -> C since answers given solve the problems asked about.

However, closure A -> C would be problematic: first question asks about how to display results, while answer in the second question says nothing about this.

^{Note how question in B does not ask about displaying results, making dupe closure look reasonable. Also note how answer in B provides guidance beyond its scope, along with addressing the question asked.}

For a more realistic example, refer to the non-transitive chain of duplicates at Meta Stack Overflow:

• A': Spiteful downvoting? Two fast DVs to my unrelated posts

  • B': What can I do if I'm the victim of serial downvoting?

    • C': What is serial voting and how does it affect me?

A' is closed as a duplicate of B', which is in turn closed as a duplicate of C'.

If you take a closer look at revisions history and comments, you will notice that initial closure has been changed from A' -> C' to narrower one, A' -> B' because readers disagreed that widest dupe target has answers relevant to the question asked in A'.