[Popcon-developers] Bug#418891: The number of etch installations is rocketing...
- Subject: [Popcon-developers] Bug#418891: The number of etch installations is rocketing...
- From: Bill.Allombert@math.u-bordeaux1.fr (Bill Allombert)
- Date: Fri, 15 Jun 2007 19:11:39 +0200
- Message-id: <[🔎] 20070615171139.GU1198@yellowpig>
- In-reply-to: <1176390364.5359.59.camel@dorfl.liw.iki.fi>
- References: <2flveg2uf7r.fsf@saruman.uio.no> <20070412092913.GA29982@kitenet.net> <20070412093150.GA6424@ulisse.dmgroup.hu> <20070412100835.GA6893@ouaza.com> <20070412101656.GB6424@ulisse.dmgroup.hu> <1176374372.5359.55.camel@dorfl.liw.iki.fi> <2flps69vmm8.fsf@saruman.uio.no> <1176390364.5359.59.camel@dorfl.liw.iki.fi>
On Thu, Apr 12, 2007 at 03:06:04PM +0000, Lars Wirzenius wrote:
> Package: popularity-contest
> Severity: wishlist
>
> On to, 2007-04-12 at 13:49 +0200, Petter Reinholdtsen wrote:
> > [Lars Wirzenius]
> > > However, changing the popcon debconf question to something like the
> > > following might be an acceptable compromise:
> > >
> > > Would you like to participate in the Debian package popularity
> > > contest meter? Blah blah blah. You may also choose to only
> > > report once that you have installed this release of Debian on a
> > > new machine. In that case no other information from you or your
> > > machine gets recorded.
> > >
> > > [ ] No, I do not want to participate at all.
> > > [ ] I want to report once that I have installed Debian.
> > > [ ] I want to participate in the package popularity contest.
> >
> > Interesting idea. Perhaps it deserve a wishlist request in BTS?
I am not sure report-once is workable. Periodic report provide us with
basic data integrity checks that are necessary for the data to be
meaningful. This favors short-lived installations (someone resinstalling
every three months) over long-lived installations (someone installing
woody and upgrading all the way to lenny).
Also I am afraid this will reduce popcon participation by leading users to
choose the 'middle-ground' option.
Generally, collecting data is easy. What is hard is to provide
meaningful interpretation of them.
Cheers,
--
Bill. <ballombe at debian.org>
Imagine a large red swirl here.
Reply to: