Re: [Gnu-arch-users] Re: Re: Obsoleting abrowse

[David Allouche]

On Mon, 2005-08-01 at 08:38 +0200, Matthieu MOY wrote:
> Mikhael Goikhman said:
> > missing --categories and company,
> 
> --categories and --branches are missing. Personnally, I think adding them
> to rbrowse would complexify the code (and the spec, since the current
> rbrowse never displays categories and branches individually, I don't know
> what the output of baz rbrowse --categories should be) a lot for very
> little benefit.
> 
> > full match rather than prefix match.
> 
> If you mean "full match of category, branch, or version", this means the
> spec will have to change again with the future namespace change. I
> strongly prefer opting for the flat namespace now, and not having to
> change anything later.

I completely agree.

I see very little value in treating the c-b-v components of a branch
name individually. All searching and listing needs can be satisfied
using prefix match like "category--" or "category--branch--".

That would also perpetuate a meme that has plagued many Arch user
interfaces: that c-b-v is a generally practical and meaningful way of
presenting branch lists. In my experience, this rigidly structured
hierarchy obscures the contents of categories containing few branches
(aka versions), and is inadequate for categories containing many feature
branches.

Then there is the added implementation burden, CLI complexity, and the
fact that encodes a rigid naming policy that is explicitly going to be
phased out.

Did I do my weekly rant against the a/c-b-v namespace yet?
-- 
                                                            -- ddaa

signature.asc
Description: This is a digitally signed message part