To me, the answer is obvious: Ask for advice on methods for solving the problem at hand, instead of a solution to the problem at hand. That way you can learn more than you might if you just went at it alone.
Consider, for example, my own experience in a recent thread on
microcontroller ADC calibration. We talk about the problem and suggest solutions on it. However, after re-reading the thread, I realized that explicitly describing the method I was using to find the solutions I suggested, would be worth way more than any of my suggestions:
this. None of my other posts in that thread actually help that much, but knowing the method on how to find answers to such problems is definitely useful.
This means that instead of asking
"Should I use a crimped or a soldered connection for this?", one should try to find out
"When should I use crimped connections, and when to solder them?". In practice, you start with the first question, but when you get an answer, you explore
the basis of that answer, and find out the logic of how they arrived at that answer. I often do this myself, and I do claim it works well.