The ```reference-request``` tag is currently described as: “this tag is used if a reference is needed in a paper or textbook on a specific result.”

I have found current usage of this tag too broad to be meaningful (and was therefore motivated by Francois Dorais’s [question][1] about tag issues). In particular,
 - if this tag applies to any post that says “references welcome”…
 - if this tag applies whenever someone does not know if good references exist…
 - if this tag applies to any question which might be well answered by an existing reference…

then it applies to almost all questions, and is not useful. Nonetheless the tag is often used for these reasons.

I would propose to see this tag redefined to be more specifically about requests for *references* as opposed to requests for *mathematical results*:

> Requests for references on results which are already known to the
> person asking the question, e.g.:
> - results which have been used elsewhere without comment,
> - results which seem like simpler variants of results elsewhere,
> - results which you have proved and suspect have been proved elsewhere.

Would other people here want this tag to be described like this? And if so, how might we go about removing incompatible prior uses of the tag? Or would others have comments in favor of the current description, or other proposals?


  [1]: https://meta.mathoverflow.net/a/5442/44143