RE: Incorrect terminology in Customize doc

[Drew Adams]

    > Some of the Customize documentation incorrectly refers to user
    > preferences, which can be either faces or variables, as "user
    > options". In Emacs terminology (elsewhere, and longstanding), a "user
    > option" is a `user-variable-p', that is, a variable that users are
    > intended to change (its doc string has "*" at the beginning).

    Since all custom vars are user-variable-p, I think it make
    sense to expand
    the notion of "user option" to also include other custom-izable objects
    like faces.

All custom variables are user-variable-p, but of course not all
user-variable-p variables are custom variables. In any case, I don't see the
connection between that and the idea of expanding "user option" to faces -
that is a separate matter.

Of course, we are free to change the terminology at any time, but 1) there
should be a good reason, 2) the new terminology should be able to name all
the denoted entities, 3) the change must be made throughout, and 4)
preferably, the name change would not conflict with past usage (to avoid
confusion).

What we have now is a partial name change. The result is confusing.

I personally have no problem with "user option" as a synonym for "user
preference", and denoting both user variables (`user-variable-p') and faces.

But if that is the choice, then that change needs to be implemented
correctly and thoroughly. And it does involve a _change_ to the previous
definition of "user option", which explicitly was identical with
`user-variable-p'. Because of the possible confusion, I thought it would be
better to use a new term, "preference", for user variables and faces.

I do agree, however, that the name "user option", stripped of its Emacs
history, does not imply any restriction to variables, and it would be an
appropriate term.

The point is to be clear and consistent. We should choose one way to do
that, and then stick to it. The current state is bugged.

 - Drew