When to use inline versus displayed equations in a publication?

Question

On a publication (thesis, paper, other), when should I write an equation inline and when should it be separated and numbered from the text?

I've seen publications containing both styles, but I don't know if there is a rule to choose among them. My field is applied computer science.

Answer 1

A numbered equation always takes a separate line, so there is no question here — if you need to refer to some equation later in the text, it deserves a separate line and it needs a number.

If your equation is as long as a line, or even longer, then again — it should always be placed out-of-text.

To answer the rest of your question, we should think a little how we (e.g. your readers) read inline equations, and how we read equations on separate lines. First of all, inline equations merge into the text, and provide some smoothness (it is just like you speak and draw on a whiteboard at the same time). In contrast, equations of a separate line break the text (just like you write a long equation on a whiteboard and take a big pause to let your audience reflect on it).

If there are no pauses in the text, it is hard to comprehend. Too many pauses raise similar problems, because it becomes difficult to focus on the most important pieces then. My advice is to think about the role of each equation and put only the most important equations in a spotlight of a separate line.

Answer 2

It's matter of style. But always think about the reader - what is the most convenient for her/him.

I agree with most of points raised by Dmitry. However, when reading publications I often see too many inline equations, which make it harder to comprehend. It often happens due to space restrictions.

My personal stance is that only very short and simple things should be put as inline equations (like $x \in {1,2,3}$ or $y=Ax$). That is, ones that at the same time are too technical to deserve a separate line and short and easy, so they can be read with the sentence without any pause.