If you compare my remarks in

Do partitions of unity exist if we impose additional conditions on the derivatives?

and

Exponential sums

you'll see that the problem and my request were exactly the same in nature. In the first case, the OP cooperated and we had a happy ending, while in the second case he didn't and we are still at square one.

Questions:

1) Did you have the same experience?

2) What is preventing the people from asking exactly what they need? Are they afraid that if they reveal too much, I'll steal their work and publish it as my own, or what?. (By the way, if there are some good (able to understand other people's ideas and present them well) thieves out there, I always have some yummy stuff for you ;)).

3) Should my frequent request "Ask exactly what you need" be added to the "How to ask a good question" article?

• Re 3): Yes. Also: Vote to close as "unclear"; the vote can be reverted once/if the issues are clarified. – Andrés E. Caicedo Sep 6 '13 at 14:15
• Re 2): impossible to answer. Rhetorical question? Re 1): sometimes, I guess. But I rather suppose that (and this can happen to anyone in the course of doing research) one doesn't always know exactly what one wants to ask. I think Halmos remarks on this in his automathography: often people seeking help are really looking for help finding the right questions rather than black and white answers. (For another data point, I had to guess what the OP really wanted here: mathoverflow.net/questions/141363/reference-for-stasheff-operad/…). I think you're doing well here, @fedja. – Todd Trimble Sep 6 '13 at 14:53
• @ToddTrimble Yes, ok, that's true. So rather than "ask exactly what you need", perhaps words to the effect that the question should be as precise as possible. If the point of the question is an attempt to clarify what the right question ought to be in a specific setting, this should be made explicit. – Andrés E. Caicedo Sep 6 '13 at 15:35
• "one doesn't always know exactly what one wants to ask." True in general, but in both examples I gave it clearly wasn't the case. I mean, the situation is often like I need $a+b<a^2+b^2$ under the condition $ab^2=1000$ and ask "What is known about quadratic inequalities under cubic constraints?". I doubt Halmos would recommend such approach :). 2) Of course, you cannot tell what the reason was in each case, but can you just imagine any reason that would make you behave like this? @Todd Trimble – fedja Sep 6 '13 at 16:15
• @fedja: Provided that people know exactly what they need for some research purpose, I'd suppose in many cases that the urge not to hone it down to the exact specification comes from a hunch that something more general (and more illuminating) could be said. It could of course also be paranoia, either in the sense you mentioned (concerns about stealing or being scooped), or alternatively in the sense that by exposing one's true thoughts, one might look stupid. I'm not sure which (of any) of these possibilities would apply to your test cases. – Todd Trimble Sep 6 '13 at 17:28
• @fedja the attitude/action you mention in your most recent comment may also stem from an alarming attitude to maths that I see occasionally (or when I referee, frequently): "I want to prove X, there must be some Big Theorem that implies X, just like in my classes". So for such people, solving the question becomes a matter of being told which large book of recipes to look in. – Yemon Choi Sep 6 '13 at 20:17
• something more general (and more illuminating) could be said Yep. I'll sort of buy that though if you want A, cannot do it yourself, and post a generalization B, you forfeit the possibility that people may know other "general and illuminating" approaches C,D,E, all of which imply A, but not B. @Yemon Choi The phrase "like in my classes" may reveal more about the state of affairs than you intended to say. I should remember it when running my graduate course. :-) – fedja Sep 6 '13 at 20:43
• Part of what is going on is probably that the "How to Ask" page specifically recommends generalization as a way to make your question more interesting. – Charles Staats Sep 15 '13 at 17:52