Unification of terms "ring unit" resp. "multiplicative identity"

[avekens]

According to issue #4512, we have several different names for 1r already at the beginning of section "Ring unit". With this PR, these differences should be levelled out.
As proposed (and decided in principle) in issue #4512, 1r should be called "multiplicative identity" or "(ring) unity" or "unity element" uniformly in (i)set.mm.

An overview over the terms used in literature, and a discussion about the ambiguity of the term "unit" in the context of rings, is contained in the header comment of the section "Ring unit", which is changed to "Ring unity (multiplicative identity)" (see first commit).

In the following commits, the comments in set.mm are revised. I will do the same with iset.mm after this PR is accepted and merged (to avoid the effort to modify two files simultaneously if there are review remarks).

Here the details (provided in separate commits):

• Title and header of section "Ring unit" -> "Ring Unity" (including addition of bibliographic reference [BeauregardFraleigh])
• Replacement of "unit" by "unity" in comments (contains most of the changes)
• lattice unit replaced (Two meanings: (Hilbert) lattice one and lattice unity)
• unitary ring -> unital ring
• nonunital -> non-unital
• multiplicative neutral element -> unity element

[avekens]

I think the level of section "Semirings" is wrong (currently subsubsection of "Ring unity (multiplicative identity)", formerly "Ring unit").
It should be a subsection of section 10.3 Rings. I will change this in the next commit.

[jkingdon]

To be honest, I'm not completely sure what the "best" term to adopt is (if there even is such a thing), but this is clearly an improvement over what we had.