Gregory Maxwell said:
When it comes down to it, if we only use the criteria of "verifiable and NPOV" we end up with the prohibition on original research as being the only real control on what can go into wikipedia after a little tidying up.
That's basically it. With the proviso that an editor must want to write about it. Unwanted articles should never get written. There is a slight flaw here if we permit people to produce articles using bots, such as rambot, but that seems to be a result of a general decision that a certain class of geographical feature should be in Wikipedia. Not an unreasonable assumption, but not without its problems in the early days. This is amusing because the direction to avoid original
research is often taken merely as advice and only used as a rule when there is conflict
That is not my understanding. "No original research" is official policy.