A difficult question. Since it was me who pulled the trigger, let me give a “confession”.
I have seen mainly three different kinds of "bad" questions in MSE so far:
- please-do-my-homework-I-don't-know-anything-about-it type questions;
- the question is mathematically ill-defined but might incidentally trigger some interesting/meaningful discussions or answers, sometimes even deep further questions;
For the first kind of questions, there is no doubt that it would/should be deleted immediately by the moderators. (Yes, I do have a very narrow definition for "spams".)
The second kind are (usually) mathematically well-defined homework problems or standard textbook exercises. Such questions are discouraged by ([Edited:] some members of
?) the MSE community for not only that one would learn nothing from merely copying a problem here (OK, you might learn some LaTeX) but also that if such behavior was encouraged, MSE would soon become a plain homework problem site. On the other hand, closing such question as soon as possible to put the question
[on hold] could force the asker to try to at least put some efforts into the problem. If such question has never been improved and have been
[closed], it would be useful for the readers who can see a complete/sketch answer under the post. For this category of questions, I think closing them would be enough.
The third kind questions are a little bit complicated. The asker for such question usually does not know what s/he is talking about. The question is mathematically ill-defined or certain terms in the question need further explanation. Ironically, the people who notices such issue would usually not ask such a question in the first place while the reason for the asker to ask such a question is that s/he doesn't know the necessity of a well-defined mathematical question at all. In such situation, one of the most beneficial things to the asker (and the interested readers) might be pointing out that the question is ill-defined and how it is ill-defined. Learning why a dumb question is dumb is meaningful. For instance, in this question:
The dilemma of $\pi$
besides that the last sentence is nonsense and the asker obviously didn't notice that, he didn't know that in order to answer his question one should know the following things:
- what is the definition of $\pi$;
- what are irrational numbers and rational numbers;
- what "represented" means.
Anyone who knows elementary real analysis would know that these are all non-trivial questions, but the asker didn't know that. In this case, MSE is useful for letting people (like the asker) know that such knowledge is indispensable in order to understand/answer the question.
I should have noticed that "bad" questions are different from "bad" answers. With such reflection, I would have been hesitant to vote to delete all the high-negative-score questions.