I think that in all writing, and especially in mathematical writing, signposting is extremely important. So a calculation-heavy proof should include a short summary, and then the calculations should be very clearly marked so that the reader knows that "here are the nitty gritties of the calculations, and here's where they stop, so you can skip them and pick up at the end".

For short papers, I like the following format. Begin with a very short introduction, setting the paper in context and giving a couple-sentence outline of the paper. In the second section, state all definitions and theorems precisely, but do not provide any proofs that are more than one or two sentences. Then the rest of the document, in as many sections as necessary, provides the detailed proofs, with all non-immediate calculations. The point is that most readers can read sections 1 and 2 and get everything out of the paper, and the only people who will read sections 3+ are: the reviewers (one hopes), and anyone trying to generalize the actual proof to a different setting.