[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Napster clients, so much diversity.



On Thu, Mar 02, 2000 at 12:23:22AM +0000, bug1 wrote:
> Jacob Kuntz wrote:
> > i took that statement to be sarcastic. with the size of the distribution
> > growing so fast, shouldn't we have some restrictions on level of
> > completeness?
> 
> If we start limiting the size, who is going to judge wether a package is
> "worthy" of inclusion ?  Personally i hate EMACS cause i havent had time
> to understand it, so to me EMACS is pointless and otherwise useless, but
> i recognise its not my position to judge EMACS usefullness in other
> peoples eyes.
> 
> IMO the only restriction on what deems an acceptable package should be
> technical and legal.

Jacob was talking about a technical restriction: if most of a program's
features are unimplemented or unreliably implemented, it is technically
incomplete. We're feeling around the edges of some rational metrics for
buggy/useless packages, and it would be nice to have some guidelines on
this point.  But we can't even have a discussion about what might be
useful in determining removal candidates without someone throwing in a
hand-waving argument that the distribution's size is irrelevent and that
any trivial package that someone throws together in an evening is
sacrosanct. (I snipped the page you spent expanding on that basic
theme.) The only person who brought up removing a package because of
personal dislike was you! That kind of argument only serves to distract
people from the real issue: how do we deal with junk packages? We know
that personal dislike is an invalid justification; more useful would be
to discuss some valid criteria.

Here's a start: Let's say someone installs a package and finds that it's
mostly a menu bar and an about box. Most of the menu items do nothing,
and the few that do something lead to immediate crashes. He brings it up
on debian-devel and a couple of people confirm that the package does
nothing. No one chimes in that the package is useful. Is it too much to
ask that the developer justify its inclusion in debian?

-- 
Mike Stone

Attachment: pgpMHYqefROeU.pgp
Description: PGP signature


Reply to: