+1, I'd add two things: 1) Questions about non-mainstream physics should be ok, e.g. ("Have Podkletnov's antigravity experiments been independently reproduced?", "Why do we believe cold fusion can't work?"). Questions in fringe physics aren't ok ("Where does the square root come from in this mass-formula?"). 2) We should be careful about "questions that propose a new concept or paradigm". Those can be very useful, and generate research-level content, but they can also invite half-baked speculation (like "Are all particles the same string weaving the fabric of the universe"?)
The author of the answer replied:
@jdm afaict (1) is not allowed, but I'm not sure. Regarding (2) That's actually OK. While the "one string to rule them all" question wasn't worded too well, it didn't fall afoul of the current non mainstream policy.
Then @jdm replied and said:
Regarding (1), sometimes I come along something in a paper or a pop-sci article, and would like to know whether it's legitimate or not. Say, the Woodward effect, Mallet's "Time machines", and so on. I'd like to be able to ask "Is this mainstream science?" and if it's not, "Where's the author's mistake, why can't this work?". Do you think that's OK?
Both comments by @jdm have been upvoted, but the middle comment by @Manishearth has not. However, since @Manishearth is the original author of the answer, I am a bit confused about whether questions of type (1) are allowed on the site, and for that matter, questions of type (2) as well.
What is the site policy regarding the two question types noted by @jdm in their comment?