On Fri, Jun 21, 2002 at 02:36:12PM +0200, Aaron Isotton wrote:
> > It's not about censorship (why the hell do people always insist on making
> > any objection to a package an argument about censorship anyway?)
>
> It *is* about censorship. What you are saying is "remove the packages
> which *I* don't think have a 'real use', or which *I* don't think have
> enough users". In my eyes, that's censorship and nothing else.
You are acting like a moron.
The dockapp in question is self-described as "stupid". It is admittedly a
JOKE written by its author and is not intended to be a serious applet. It
does not provide any features which have any actual value, even in the
eyes of the things author. It's a JOKE, and a bad one.
The fact that someone thinks it's worthwhile to put this in Debian is
nice, and clearly I don't have the authority to say that he can't do it.
But I do ask for application of common sense when packaging things, which
it seems you are incapible of doing. This is not my problem. If you want
to bloat the archive with this, a similar applet which posts messages
about Natalie Portman's hot grits to Slashdot, another containing a few
gigabytes of porn, or anything else for that matter, there's absolutely no
way I can prevent you from doing so, other than appealing to your common
sense that these things do not really belong in Debian.
Since this has clearly failed, so far as I can tell due to a complete lack
of any such common sense to appeal to, do as you like. Just don't come
crying to me the next time someone hands out 100 free Mandrake CDs because
100 sets of Debian CDs would have been too expensive at an installfest.
(Hey, I've watched it happen a few times before, and the Debian people
always bellyache that nobody was giving away Debian sets, which take too
many CDs for the vendors to donate a whole pile of them for giving away..)
> BTW, how are you going to remove the "useless" packages? Do you
> really think that anybody saying "I am the boss here, and I decide
> what goes into Debian and what doesn't. Your package is useless.
> Your work sucks. Fuck off." will improve the general quality of
> Debian and the collaboration between the developers?
What the fuck are you smoking? Useless packages should be removed by
their maintainers because they are no longer useful. I realize that's too
much to ask for from this project, but that is what should happen.
> > bzip2 was packaged because someone thought it would be a useful thing to
> > have in Debian. And it is, lots of people use it. We're not trying to
> > keep useful things out of the distribution. However, I'm not convinced
> > this thing is actually useful for any real purpose.
>
> As Anthony pointed out, I was saying something completely
> different. Using Packages.bz2 instead of Packages.gz would save 23%,
> which is probably more than you'd ever obtain by eliminating so-called
> useless packages.
I don't know, there are a lot of packages in the archive which are neither
maintained nor used (widely or otherwise), some of which are not even
installed on more than a few machines according to popcon. Removing these
would be a good start toward warding off archive bloat.
> > > Please define "real use".
> >
> > Use your brain! That's what you've got it for. I'm asking for people to
> > apply a little common sense. I realize this is a lot to ask from people,
> > especially in this project, but it's the only way to prevent more and more
> > useless crap from filling the archives, making Debian CDs cost more, and
> > slowing down the release process while people try to fix silly bugs in
> > packages that nobody uses, rather than worrying about the not so silly
> > bugs in packages people do use.
>
> Debian is made by volunteers. They fix what they want, when they want
> and only if they want. You cannot force anybody to accept your
> priorities as his own.
And when a thing that was packaged is no longer useful to the person who
packaged it, or anybody else who might take it over? How many packages do
we have up for adoption or outright orphaned? Many of these are useful
things, many are not really used by anyone now. Several of these haven't
even seen a recompile since potato, and some of them even since before
that.
> > Again, you're dodging the issue.
>
> Not at all. What you are saying is to package only what is popular
> and what has "real use". (Not defining what "popular" and "real use"
> means, of course.) Doing only what's popular and used/liked by most
> has lead to Windows XP. I, and many others, prefer a bloated
> Packages.gz because if we want some applications only few others are
> using, we must accept that other people will want some other
> applications only few are using. Or, in other words: if I want to
> have that program with "no real use" to you in Debian, I have to
> accept another package with "no real use" to me.
Popularity has little to do with it. There are certainly applications
which have a small target audience. When the target audience is totally
non-existant, however, the package needs to go. And if the target
audience for an application is microscopic to begin with, the maintainer
_should_ be thinking about the value his package will add to Debian. If
he can't think of anything, maybe he should not package it. This concept
is too complex for you, I think.
> > If the software is only useful to a small handful of people, then those
> > people should have the software. If it's useful outside that small group,
> > it probably fits well in Debian. However, it's not our job to serve as
> > the central repository for every single coder's joke program, CS homework
> > assignment, porn collection, or anything else that someone, somewhere
> > wants - even if they are the only person in the world who does.
>
> What is a "small handful"? 5? 10? 50? How do you know how many people
> will actually use a program? If it takes an hour to package a
> program, and 10 people install it, are these 10 saved hours worth less
> than a tenth of a second of download time of Packages.gz?
If it takes you an hour to install this silly dockapp, then something's
wrong, but I'll concede that there are certainly packages which take me at
least as long as your example to _compile_ which are not yet in Debian.
The amount of time and headache saved can certainly be a good reason to
package something, but probably should not be the only reason a package is
uploaded.
> > If you cannot apply common sense to the question "should we have this
> > package?", then I can't help you.
>
> Define "common sense". What *you* think, maybe?
No, what the maintainer thinks. I think this package is as stupid as it
has been described to be and think it doesn't belong in the distribution.
It's not my call, however. I've registered my opinion; this is the extent
of my ability or interest in this particular case.
Hey, we've even got intercal packaged - that's gotta be the most complete
waste of packaging effort ever. I would assume (but actually highly
doubt) that the person who packaged it had a better reason for doing so
than "hahaha, someone should package that just so we can say we have it."
Such logic would be the epitome of archive bloat, and I'm quite convinced
things have been packaged for precisely that reason in the past. This is
a practice which should stop.
--
Joseph Carter <knghtbrd@bluecherry.net> You expected a coherent reply?
Basically, I want people to know that when they use binary-only modules,
it's THEIR problem. I want people to know that in their bones, and I
want it shouted out from the rooftops. I want people to wake up in a
cold sweat every once in a while if they use binary-only modules.
-- Linus Torvalds
Attachment:
pgp6lilZfL_33.pgp
Description: PGP signature