A large and growing set of math.SE questions are of the random task type:

1. Question is a task : "prove (a given mathematical statement)", or worse, "prove or disprove" or "find all $n$ (from an infinite set) such that (...)".

2. No source for the problem.

3. No information to indicate whether the problem has a solution or is an unsolved research question.

4. If problem is solvable, no information on the difficulty level.

5. Seemingly arbitrary problem statement: no obvious reason (and none given in the problem) to be interested in the particular equation, function, etc in the problem.

6. No given reason to believe that the solution of this random-task would be interesting to those who attempt it. e.g., we are not told that the problem is from a textbook (hence likely to be instructive), or a competition (hence likely to have a short, clever or nice solution).

7. No other motivation or rationale for the problem is provided.

For these the tag [random-task] or equivalent indicator would be very useful, as well as an FAQ item specifically addressing and discouraging random-tasks. Notice that 1-7 are objective indicators so that this classification does not rely much on personal judgements. A [random-task] tag would signal to users that a question may not (in its present form) be worth their time, and especially if codified as an FAQ item, would advise questioners to make specific improvements in their queries. Adding information to a random task question can produce an interesting, non-random task, even if identical in mathematical content.

Example of a [random-task] question: If $(a^{n}+n ) \mid (b^{n}+n)$ for all $n$, then $a=b$

(In addition to the lack of sources, motivation, or difficulty calibration, this task is random in the sense that the same question could have been asked about any other family of functions in place of $f_n (a) = a^n + n$. There is no indication of why this family is worth considering, or if the functions themselves are not of particular interest, why the proof would be interesting, as in the case of textbook or competition problems.)

I edited the question title to reflect the possibility of an FAQ item on this phenomenon, without creating a tag. However, the random-task postings are frequent and likely to continue until specifically discouraged, and I think having both tag and FAQ would be useful.

Can we stop promoting these kinds of (meta) tags? If a question is bad, downvote it, or even close it, not abuse the tag system.

• It is totally misleading, and factually incorrect, to lump all these proposals as "meta-tags" instead of reviewing their individual merits. Both the [unsourced] and [random-task] tags are objective attributes of the question content (does it contain sources, is it a proof task without information about the problem or its difficulty, etc). They can, if you like, be presented as deterministic algorithms applied to a question. This is the antithesis of a meta-tag as defined in the "death of metatags" posting linked from SO. Please stop implying that these proposals are meta- or abusive. – T.. Sep 4 '10 at 8:33
• Concerning the merits of the proposals, do you have a specific objection to the [unsourced] or [random-task] tags that does not boil down to confusing them with the dreaded, forbidden metatags? I concede that [generic-question] may (or may not, depending on definitions) qualify as a metatag, and that [solution-request], although objectively definable, has gotten reasonable criticism. No specific objection to the other two has been posted so far. – T.. Sep 4 '10 at 8:37
• Also, I call your attention to the proposal for defining and discouraging "random task" postings as an FAQ item irrespective of whether it is added as a tag. Whatever your specific objections to the tag (though you have yet to articulate any), do you agree that discussing this in the FAQ may be in order? – T.. Sep 4 '10 at 8:44
• @T..: (a) 1-7 are not all objective. 'No obvious reason to be interested' is probably just because someone is not familiar with that field. 'No given reason to believe that the solution ... would be interesting to those who attempt it' is again very subjective. What if one person is agree with it while the other disagree? Edit war. – kennytm Sep 4 '10 at 8:55
• @T..: (b) Also, these tags describe how the question is formulated, not the content of question itself. [unsourced], [random-task] can be resolved just by providing the source or point of interest, but these won't change the topics ([number-theory], [topology], etc.) the questions belong to. Personally, I would discourage all non-topical tags, but that is debatable. – kennytm Sep 4 '10 at 9:01
• Yes, the point is to have tags that get the problems resolved. Math subject tags do change for many questions. In neither case is stability a prerequisite for the tags to be useful. You concede that [unsourced] is objective and not a metatag. As for [random-task], items 1,2,3,4,7, and the frequently-applicable parts of 5 and 6 ("no reason given [i.e., not explicitly stated]", "not [cited as] from a textbook or competition") are objective, and at least as objective as the math topic classifications of questions, which also may differ slightly or not so slightly between users. – T.. Sep 4 '10 at 9:31
• @T..: If the point is to get the problems resolved, leave a comment, downvote, or edit the question. Tagging it is just redundant since the user won't know the problem just with the [random-task] tag anyway. – kennytm Sep 4 '10 at 9:45
• Comments don't suffice. (1) Experience shows that comments to the questioners don't necessarily lead to improvements, are time-consuming, and do cause back-and-forth arguments and disputes, inter-user animosities and so on. (2) Comments don't signal other users that there's a problem with the question until they read it. (3) One way to "get the problem resolved" is to not work only through the original questioner, but to help other users decide whether they want to invest time reading posts that are [unsourced], [random-task], [computation-request], or whatever else. – T.. Sep 4 '10 at 9:59
• @T..: (1) Neither are non-topic tags better. (2) Downvotes do. (3) Could be subjective; or not in sync with whether the problem is resolved. – kennytm Sep 4 '10 at 10:18
• @KennyTM: (1) what's better is to have non-topic tags as an additional option rather than comments being the only choice. One could use either or both depending on preference. Do you agree more choice is better? (2) Downvotes can also be applied, and do not subsume the information in the tags. Again, more options is better. (3) Problem can be resolved or not, there is never a guarantee. In my approach, we don't need to assume it will be resolved; in yours, possibly futile/acrimonious attempts to get it solved are the only option. I like having more options. – T.. Sep 4 '10 at 10:36
• @T..: (1-3) That means non-topic tags are redundant. And no, more options aren't always better, in particular these tags just make the user experience more complicated. (3) We don't want to keep bad questions. Problems should be resolved, or should be expelled from the site (by closing or deleting) if OP or high rep users refuse to improve it. Giving it a [this-question-is-bad] tag won't make the site look good. – kennytm Sep 4 '10 at 11:02
• Non-topic tags are clearly NOT redundant with downvotes. The information content of nontopic tags like [big-list], [puzzle], [homework], and [reference-request] -- or the proposed tags such as [unsourced] and [random-task] -- is more specific than anything carried by up/downvotes. We have one user in math.SE who has generated a huge volume of complaints for posting unsourced random tasks, but has accumulated a large net number of upvotes (and thus is ignoring many of the complaints) on those postings because they contain other positive features. Up/down count is only one dimension. – T.. Sep 4 '10 at 19:12
• @T..: I would like to emphasize that my opposition is on the abuse of tag system. Actually I should not bring on "meta tags" as these kinds of tag usage has been banned long time before the concept was introduced (to kill [subjective]). (1) It is redundant to downvotes & comments together. BTW, the question you linked to has zero downvotes and no specific complains. – kennytm Sep 4 '10 at 20:02
• @T..: (2) It is hard to ignore because the tag sits right next to normal tags used for classify the question in topics. Plus, what's the point of having a feature that you don't want users to use? (3) If you want the mod to clean a question, "flag" it for moderator attention. (BTW, on the FAQ part, please suggest over this question). – kennytm Sep 4 '10 at 20:04
• @KennyTM: The statement about meta tags is not the gospel truth. It's a matter for each site to decide independently. – 97832123 Sep 5 '10 at 9:19