[aur-general] Official discussion period - Rules governing packages entering [community]

[Aaron Griffin]

On Fri, Dec 5, 2008 at 1:41 PM, kludge <drkludge at rat-patrol.org> wrote:
> example 2: flagging packages as anything doesn't guarantee any course of
> action, in any of the repositories.  bitlbee, in [extra] is 0.2.1
> versions behind the upstream release.  it has been flagged-out-of-date.
> the maintainer has been contacted individually.  an updated pkgbuild has
> been sent to the maintainer and posted on the forums.  but it's only
> getting older in [extra].

CCing Jeff on here so he remembers this. :) The rest is a response to
the actual content.

> so, while the aaron/thayer amendment to the proposal (or is it a
> separate proposal?) provides a couple useful new statistical measures, i
> don't see that it would actually generate better statistics.  several
> ideas have been floated to solve this particular problem, like download
> statistics.  those all need more consideration and development, though.

Err, I didn't think of it as an amendment or anything, I was just
babbling. If you personally would like to develop a way to track
download statistics, then please provide us with something to do this.
I do not think this will give us very accurate statistics, unless we
keep very heavyweight statistics (downloads / day or something,
tracked over the life of the package, so we can see waxing and waning
popularity). That's the way it works with open source - I don't think
this is a valuable/good idea, so I have no desire to write the code to
do it. If others think it is a good idea, then they must at least help
out with the work.

> it seems to me that there won't be any real consensus on a concrete
> proposal to regulate [community] until there's a mechanism for
> generating accurate, reliable usages statistics.  i would anticipate a
> close vote; given the furor that's surrounded this proposal, i would
> also anticipate a lot of bad feeling on both sides arising from a close,
> binding vote.
>
> if i were a tu, i'd move to table this proposal and form a working group
> to study the social and technical problems of generating good usage
> statistics.  it would put off a resolution to the resource consumption
> problems, but i feel that, sometimes, "now" is not better than "better."

This sounds like way more bureaucracy (man, I can never spell that
word) than we need or want. And with regard to your now/better point:
I think it'd a good general rule that 75% now is better than 100% a
year from now

> in the meantime, it seems that moving all the games out of [community]
> and into [games] is one concrete and non-controversial way to take some
> load off the server.

Yeah, as was pointed out a while back, it's not exactly about
resources anymore. We've solved that particular problem by throwing
money at it. It may, however, be important to note that not all
mirrors have the resources we do. Some mirrors only have 5 to 10 gigs
of space to offer.