It is perfectly acceptable to answer non-research level questions in comments if you feel like doing it. What is trivial to an expert is often hard for anyone else. I wouldn't send a colleague who knocks on my door with a trivial question to go take a graduate course. We of course should firmly discourage users who flood MO with trivial questions.
Clarification in response to comments by quid and Niemi:
The definition of a "research level question" (recalled in comments) is very inclusive, and pretty much any question asked by a professional mathematician in the process of his/her work will be research level regardless of how trivial it is. I like it that way. I see a trend to on MO to agressively close trivial questions, and I see that many users do not even want us to answer such questions. The latter is (in my view) harmful to the MO as it makes the site less welcoming, and frankly less useful to my own work.
As to whether one should respond to trivial questions in a comment or in a formal answer, I do not really care. My preference is to respond in comments because I find distasteful to earn reputation points on such answers, but then I also find the whole reputation point system somewhat silly--it ought not be a motivation for answering questions. What is essential is that those trivial questions get answered. If they don't, we are telling our colleagues that they are not welcome, and this is self-destructive for the MO.