Guidelines for writing question bodies when one is self-answering
It's often the case that I think of something that could be explained very nicely on Codidact and I want to make an answer for it in Q&A format (as mentioned e.g. in this post).
It's pretty easy to formulate the answer. But I wonder how to formulate the question body. E.g. if I wanted to explain a theorem, the question would be like, "How does one prove Theorem X?". Can I merely repeat the question in the question body? Should I use the body to contextualize the title question, say by talking about the connection of Theorem X with other results or about why it is interesting? Should I invent a fictitious confusion about Theorem X, e.g. "I attempted to prove Theorem X using steps A, B, and C, but failed to prove that every D object is an E F."1? What does the community recommend?
1 Of course, I can just use my own previous confusion about a problem if I remember it. Sometimes I don't remember what I was previously confused about, though, and in any case I'm not sure whether that's what preferred.
2 answers
You are accessing this answer with a direct link, so it's being shown above all other answers regardless of its score. You can return to the normal view.
The following users marked this post as Works for me:
| User | Comment | Date |
|---|---|---|
| clemens | (no comment) | Apr 10, 2026 at 15:48 |
I would not advise inventing a ficticious confusion: that only serves to confuse the reader. Explaining why the theorem is interesting is definitely worthwhile.
I would add that the conjunction of Q&A should add value. Standard theorems with standard proofs would be clutter. When I have created self-answered questions1 it's been for results which I couldn't find proofs for.
1 Do the Faber partition polynomials have integer coefficients?, Matrices with rotational symmetry
0 comment threads
Karl Knechtel gives an answer to a related question in which he talks at some length about how self-answered questions can written in a way that's most useful to the community:
People who actually need an answer to a question are often in a uniquely bad position to actually ask that question. Out of all the information they've gathered about the problem, they don't know what's actually relevant to explaining the situation of implementing a solution - that only comes with hindsight. They also don't necessarily know the right way to frame the problem, or the right scope for the question. An expert may have seen hundreds of people fail in a similar manner, and thus understand what those cases have in common. Finally - and especially for technical issues - people experiencing a problem might not know how to describe it accurately and unambiguously - because they don't fully understand the terminology of the problem domain.
The Q&A format can be an excellent device for experts to communicate their expertise. It's not always the best choice - which is part of why the Codidact software supports other category types. But when there is a specific problem that can be described - either "why does doing X fail in Y manner?" or "how do I do Z with W tools?" - asking that framing question and answering it is a fantastic presentation method. A well-chosen title can be easily found with a search engine; the body of the question confirms the nature of the problem; and an answer can convey all the necessary information - whether it's a three-step guide, a pages-long conceptual explanation, or anything in between.
0 comment threads
0 comment threads