RE: [External] : Re: [ELPA] New package: dired-duplicates

[Drew Adams]

> > Why should I not use :group for customization? I thought this makes it
> > easier to explore customization?
> 
> It's redundant - see (info "(elisp) Variable Definitions"):
> 
>      If a ‘defcustom’ does not specify any ‘:group’, the last group
>      defined with ‘defgroup’ in the same file will be used.  This way,
>      most ‘defcustom’ do not need an explicit ‘:group’.

All true.
But "not need" doesn't imply "shouldn't have".

> If :group is specified, it will most of the time mean the defcustom should
> be assigned to some other, not to the default, group.  So it's better
> style to not explicitly specify the default.

FWIW, I disagree, stylistically; I think it's worse
style.

1. I disagree that it's important either way, and
that removing "redundant" :group's should ever be
suggested as a convention.

2. It's clearer, less error-prone, and more
practical to use :group explicitly, even when
"redundant", because moving defcustoms around (e.g.
reordering) doesn't risk implicit changes to the
wrong group.

Yes, this is a personal choice - and IMO it should
be.  I don't see a reason for suggesting any
particular convention for users wrt including
:group when it's "redundant".

A convention for code that's to be included in
Emacs itself is a different matter.  I'm talking
about a suggestion that there's something bad or
wrong with users including :group redundantly.

(I realize that here it is about code submitted
to be included in Emacs itself.  I also think
removing "redundant" :group's shouldn't be a
requirement for Emacs itself.  But that's not up
to me.)