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

Re: Work-needing packages report for Sep 6, 2002

On Mon, 9 Sep 2002, Michael Cardenas wrote:

> My main reasons for suggesting this were:
> 1. to save archive space, and thus all the associated costs in money
> and time. 

A legitimate concern.

> 2. to have more useful wnpp reports that aren't filled with lots of
> useless stuff, so that more people will read them

Again, quite reasonable.

> 3. if a program has been orphaned for 2 or 3 years, I'd be willing to
> bet that there are few or no users.

I wouldn't be making that bet, if I were you.  Lack of development does not
mean lack of users.  The problem that Debian has is that it is basically
impossible to make any sort of accurate survey of what people are using, so
we can't make any sort of assumptions about usage.

> If the package gets removed, we
> will at least hear complaints from users and maybe someone will then
> be willing to come forward and maintain the packages. 

That's one way of working out if it's used.  <g>

One thing I did just think of - in the case of existing systems, rather than
new installations, the last version which was in the archive will be kept on
people's systems on upgrade, which should help reduce the issue - as long as
the person doesn't blow away their system and reinstall, and the newer
versions of libraries the package depends on don't change their semantics,
there shouldn't be any big hassles.

I'm still not keen on having packages removed from the archive for no better
reason than "there's no DD willing to maintain it".  "It's broken" is a good
reason, and (as I mentioned in an earlier e-mail) "package <foo> provides
similar functionality" is another (IMHO, the best) reason to remove a
package entirely from the archive.

A bigger problem, in terms of archive bloat, is the supposedly maintained
packages which were uploaded by new 'one-shot wonder' DDs, who have now
disappeared and left useless cruft in the archive.  I don't know of *any*
reasonable way to fix this, however - every option I can think of would
either seriously screw up the NM process, annoy real developers, or other
unpleasant circumstances.

> When looking at the wnpp for a package to maintain, and seeing that
> something has been orphaned for 3 years, my only assumption is that
> noone is using this program, so why would I want to maintain it? If

We're working from a different set of base assumptions, so I don't think
we're going to agree on this.  I can't assume that just because there isn't
a DD who wants it, nobody else must either.  We're here for the users, not
just the developers.

> the wnpp list said it was orphaned for a year and was about to be
> removed from the archive, and I wanted it to stay in, then I would
> volunteer to maintain it. 

That's great.  If there are any packages you care about on the orphaned
list, you should take them.  If you don't use it, why do you care if it gets
removed or not?  And, similarly, if you do use it, you should probably adopt
it anyway - after all, it has at least one user - you.

> I wasn't suggesting that we should force users to be maintainers,
> rather that we shouldn't keep packages around that noone is using. 

Agreed.  As soon as we can make some sort of decent approximation of usage,
without (a) inconveniencing users while doing so (hence, removing and
waiting for the howls is out) or (b) invading users' privacy (hence, no
mandatory popularity-contest).

The problem is assessing usage.  You seem to make the assumption that if
there is no DD who wants it, nobody else must either.  I make an orthogonal
assumption - that just because no DD wants it, doesn't mean there isn't a
userbase out there which doesn't get involved in debian development.

Matthew Palmer, Debian Developer
mpalmer@debian.org     http://www.debian.org

Reply to: