Actually neither. One provides the reference to the work, where the equation appeared and explains how it’s being used in the new work.

Direct quotation is never used. Meaning matters, not the words. Scientific texts aren't scripture, where the words matter.

It depends on whether you're asking the way things ought to be or the way things are.

The way things are, people just vaguely mention the formulas of other works, assume that the reader has the resources and wherewithal to go and look at that resource, and they also anssume that all readers have an understanding of that work, and that that understanding is equivalent to the work of the author of the current paper. NONE of these premises are likely to be true for most readers.

The way things are, people just vaguely mention the formulas of other works, assume that the reader has the resources and wherewithal to go and look at that resource.

However, the equations are given a shroud of mysticism, as though they have an objective meaning, even if no one explains what the variables mean, and if no reason is given for the equation to be true.

I take the view, personally, that both words and equations are important, and the equations are a language, where the numbers are adjectives, and the equal-sign is the verb. The nouns, though, absolutely must be supplied by the surrounding text, or you have adjectives and verbs without nouns, and the equations are meaningless on account that you have no idea what the adjectives refer to.

I have a method that generally seems to work for me, that in a long derivation, I put it all into one aligned equation. Over each equal sign, I put a letter, and then under the derivation, I put reasons to say WHY the equality (or inequality) is true.

In my view, the "why" of each equation really needs to have an answer for every equal-sign of the derivation, though sometimes it is so obvious that it need not be explicitly stated. The "what" of each equation is absolutely necessary in that you MUST define what each variable represents and its units. These are things that go in the text surrounding an equation.

That being said, opinions may vary. But my own opinion is, without the words, you don't have the meaning. What's quite often done in practice though is the equation is given without explanation, or a reference is made to the equation without even being given. This is especially done in papers where the argument is that "the equation does not make sense". Of course, in these conditions, no effort is made to convey the meaning of the equation, because the author does not know that meaning.

If you think of paraphrasing, something is wrong. Either the statement is not essential in the reasoning, and a citation is enough (mentioning the section and the equation number is better.) Either it is, in that case just spell out the reasoning as it should appear in your text, possibly with a notice like "I follow closely…" If that doesn't work because it is too close, keep working, you don't master the material enough, it is not useful to have a second copy.