How much background research before posting a question?

Question

Q: how much background research should we expect users to conduct before posting a question?

So far, most of the questions asked on Emacs.SE have been remarkably good. Every once in a while we get one, however, that suggests that the user has not done much (any?) background research to try to solve the problem him/herself. Anecdotally, it seems like it comes primarily from someone a) new to the Stack Exchange sites, or b) new to Emacs.

I presume there is a sweet spot we want to hit with this site:

• welcoming to new people who are just learning the ropes and are willing to do their homework, but
• preventing the site from being swamped with help vampires (e.g., SE thread, SO thread).

So, for clarification: how much homework should we expect of people? I presume we want people to RTFM, but it seems plausible that someone new to Emacs would not know about the manual or how to go about finding information in it (because, you know, s/he's new to Emacs). Should we expect a baseline of "have at least tried a tutorial," or "have tried googling my question"?

Moreover, what are our meta-norms about answering questionable questions and voting on other people's answers to questionable questions?

• should we provide answers, or comments with links to the manual pages or tutorials they "should have" found in their background research?
• for other people's answers: should we upvote or not (i.e., should we provide second-order condonement of questionable questions)?

Answer 1

Baselines

I don't have a strong intuition about what the baseline should be, but there should definitely be one. For people new to Emacs, "Have tried a tutorial" or (perhaps preferably) "Have completed THE tutorial" (C-h t) seems like a good first approximation. Once we agree on a baseline, it might be a good idea to create a series of Community Wiki posts that present, e.g., the contents of the built-in tutorial in Q/A style:

• "What is the area at the bottom of the Emacs frame called?"
• "How do I move to the beginning and end of the current buffer?"
• "How do I cancel a command?"
• ...

This would enable us to close these types of questions against the CW posts, and make it possible to point novice users to on-site resources that are under our control (Take that, link rot!). Of course, we'd also be pointing users to built-in documentation, but in some cases people might be more comfortable reading some introductory material outside of Emacs, especially if they're just starting to learn about Emacs.

Meta-norms

If we want to keep help vampires at bay I think it is very important not to answer questionable questions. At least not right away. Comments pointing to manual pages or tutorials are fine, but should also contain hints that we do expect askers to make some sort of effort before posting questions (and to share their research if they haven't been able to find what they are looking for). If it is apparent that someone is completely new to StackExchange, we can always point them to relevant sections of the Help Center.

As for second-order condonement of questionable questions: No. We should not upvote answers to these types of questions. I wouldn't go as far as saying we should downvote them, but rewarding people for rewarding (and thereby encouraging) questionable questions seems like a pretty bad idea to me.