Re: doc elisp intro cross reference fixes

[Luc Teirlinck]

Per Abrahamsen wrote:

   And I don't really see it is as a case where "legitimate" is proper
   word to use.

Let me be more precise and "define" what I mean with "legitimate".

We see a defcustom without a :group. Now we can take one of the
following two attitudes:

1.  Unless, for some reason, we have the definite impression that
    there is an obvious group this option _should_ go into, we leave
    that alone and do not worry about it.  The author is the person
    best placed to decide whether the option belongs in a group or
    not.

2.  We try to do "something" about it.

By "legitimate" I mean (1).  We definitely need to decide whether we
do (1) or (2), regardless of whether "legitimate" is the right word to
denote the difference between (1) and (2).  What would you suggest as
the "proper" adjective?

Currently (2) is done by either putting it in some group, using a
rather wild guess that gives no guarantee whatsoever, or putting it in
the `nil' group if the wild guess failed to guess anything.  (For
instance, if the defcustom is defined in simple.el, the option goes
into the "paren-blinking" group.)  If we should be doing (1), that is
definitely wrong.  Actually, I personally believe that even if we
should be doing (2), it is still wrong.  I personally do not believe
that if somebody forgot a :group for, say, `eval-expression-print-level',
`kill-read-only-ok' or `yank-excluded-properties', it is better to put
them in the "paren-blinking" group and forget about them, rather than
to, say, put them in the `nil' group and periodically check that group.
That is, if we decided on (2), we could use he `nil' group as a
debugging tool and empty it before an official release.  We also could
make the compiler issue a warning for defcustoms without a :group.

But, first of all, do we want to do (1) or (2)?

Sincerely,

Luc.