I think it would be helpful to focus on the desiderata (and why they are desirable), because those then are the values that we would be working towards, and obviously we would want to organize those values hierarchically.
I propose that the purpose of such a site is to provide a venue where good questions from people with a certain familiarity with Objectivism can receive high-quality answers. (This serves a higher purpose, the promulgation of Objectivism, which serves an even higher purpose, all of which we can assume). Obviously, this purpose would be a matter for discussion, but assuming that this is the purpose, then reaching that goal requires attention to two specifics: “good questions” and “high-quality answers”. What are “good question”, and “high-quality answers”?
I think that a “good question” (not just about Objectivism, but in general) is one that fully exploits the assumed knowledge context and gives a reason why the answer is not self-evidently known to the level of being certain. What is most crucial here is saying what that assumed knowledge context is, i.e. “who is the audience?”. If you assume a relatively low level of exposure to Objectivism (e.g. “I am currently reading a Rand novel”), then certain kinds of questions would count as “good” – because their answers are not obvious, given that level of knowledge – as compared to the assumption that people will have read the novels, VOS, CUI, ITOE, and OPAR. I’ll refer to these contexts as the “elementary exposure” and “advanced familiarity” contexts
“Good questions” must be differentiated from “bad questions” – presumably, via moderation, “bad questions” would either be converted to “good questions” by re-writing, or would be excluded. A bad question could be “not actually a question”, that is, using a question mark or question form as a means of making an assertion – a favorite tool of trolls, who will say “Many people say X about Rand (X is some insulting characterization); how can you disagree?”. Assuming that the question really is a request for information, it should request further information not only on what conclusion Rand has specifically drawn, but also why that conclusion is necessary, given facts and more basic assumptions; and there should be some reason for that conclusion to be possible, given the assumed knowledge context. For instance, in the advanced familiarity context, asking “Since one’s own interest is what defines moral choices, is a life of theft a morally proper choice if, to the best of you knowledge, you can get away with it undetected?” would be a bad question; on the other hand, it is a credible question for a noob in the “elementary exposure” stage.
So in my opinion, one of the most important issues that has to be addressed is the level of background knowledge that is assumed. I don’t think it should be too high; it should also not be zero. The problem of “noise” is caused not by people who don’t know enough about Objectivism, but by those who will not take serious action to overcome their lack of knowledge.
