(Steve Summit's take copied here from Wikipedia_talk:Reference_desk/guidelines, now archived here, diff)
The purposes of the Reference Desks are (in order):
- To improve the encyclopedia. They do this in two ways:
- by discovering, via people's questions, which of our articles need work (or are missing entirely), and beginning the process of improving/adding them; and
- by helping Wikipedia editors with their questions, as they work on articles.
- To help other people on the Internet by answering their questions.
- To have fun showing off our knowledge, expertise, and erudition.
Now, even if I were singlehandedly writing the guidelines, I would not necessarily state these purposes this way, because (as written) they're exceedingly easy to misinterpret. But I'll explain what I mean by them, and why I've listed them that way:
First of all, my list is a mixture of the way things are and the way they ought to be. (That is, I'm being both prescriptive and descriptive, thus maximizing my chances of pissing everyone off. :-) ).
I do believe that "helping the encyclopedia" has to be our first priority. Anything else, and it becomes impossible to answer the question, "then what are these desks doing as part of the Wikipedia project?" Saying that this is the top priority does not mean that it is the only priority, but it's important to list it that way so we're clear on what we're supposed to be doing here.
I also believe (in answer to a poll from a couple months ago, further up the page) that yes, ideally, we should answer every question by citing an existing Wikipedia article. In fact, I'll go further than that: for any question, if an existing Wikipedia article doesn't answer it adequately, such that an aspiring RD answerer is tempted to launch into a finely-wrought explanation, the right thing to do, theoretically, is to instead add that finely-wrought explanation to the appropriate article, and then link to it.
With that said, though, I do not actually believe or expect that any of the Reference Desks would ever actually meet this "ideal" in practice. In fact, I wouldn't want them to; a Reference Desk with nothing but clipped citations of articles, with no nice friendly explanations fine-tuned to a particular questioner's situation, would be deathly boring for all concerned, and I for one would have no interest in participating. (True confession: I love launching into finely-wrought explanations right here on the desks, existing Wikipedia articles be damned, and I know I'm not alone.) But I do think this ideal is worth keeping in mind, to (again) clarify our thinking about what our real purpose is here, even though I have negative expectations of achieving it. (But stating a primary goal or purpose that you do not ever expect to actually meet is a terribly risky proposition, which is why I say I wouldn't necessarily word these priorities this way in the actual guidelines.)
Number two is obvious. It's what the apparent purpose of these desks has always seemed to be and probably always will seem to be. But (selfish though it sounds), I think that helping the encyclopedia ought to be a higher priority than helping any one individual questioner. This does not, of course, mean that we're not interested in helping individual questioners, that we won't always try to do our best in helping them. But if there's ever a conflict between helping the project versus helping an individual questioner, the project has to come first.
Finally, number three. I'm sure I'll get scolded for putting that one in there at all. It's solely in the "way things are" camp, not necessarily "the way things ought to be". But I think it's just as important to acknowledge, because it explains why those of us who participate are actually here. It's all well and good to state highfalutin virtuous altruistic principles for these desks, but people aren't going to come here and volunteer their time and expertise to answer questions unless they enjoy doing it. So their enjoyment is important. In fact, to the extent that a certain amount of humor and friendly banter are enjoyable, those aspects are important, too. They certainly can't be denied and shouldn't be discouraged. But, again, they come at a lower priority: if (when) they come into conflict with the helping-the-project and helping-the-questioners goals, they've got to give way first.
—Steve Summit (talk) 19:14, 7 July 2007 (UTC) [tweaked 21:16, 7 July 2007]