While downloads may be inaccurate, and plenty of people will uninstall
it, you would most likely be able to at least see some sort of trend
with a package.

When package A has got 50 downloads and B only has 5, there's
obviously much more interest in package A, and possibly more demand
for it.

As for whether the package is any good... we dont know!

There's no perfect way of tracking a package on the AUR's success.

We can monitor the package's installation, via Archstats or a custom
daemon, but many either wont run archstats or will bitch about privacy
issues.

We can make people comment/vote themselves on the package, but we know
already that a lot of people wont do this.

We can use a system that checks how many downloads a package has had,
however here we dont know whether people have continued using the
package, unless they respond with comments.

I think in the meantime, a system relying on a combination, of votes,
and downloads, and a TU's judgement should be fine. If a package is
being downloaded, therefore there being demand, and it's also being
voted for, then we can reasonably say that there's interest, possibly
enough to include it in community or a repo.

Ideally, we would get more TU's/devs, then adopting more packages
wouldnt be as much of an issue. I'm willing to do my bit and keep a
group of packages up to date.

iphitus


On 7/31/05, Dimitrios Apostolou <[EMAIL PROTECTED]> wrote:
> > Maybe instead of voting, the AUR should track number of downloads, or page 
> > hits ... something more automatic.
> 
> I agree with number of downloads, since not everyone wants to register
> just to vote (me included).
> 
> _______________________________________________
> arch mailing list
> [email protected]
> http://www.archlinux.org/mailman/listinfo/arch
> 


-- 
iphitus - www.iphitus.tk

_______________________________________________
arch mailing list
[email protected]
http://www.archlinux.org/mailman/listinfo/arch

Reply via email to