My problem with voting is that stuff like...Say i use one of the firefox-like packages in the AUR (swiftfox, swiftweasel, firefox-beta, what have you) and I vote for it, but then I switch to Chrome/Chromium. It's unlikely that i'll ever remember to un-vote when i switch which would skew the popularity/vote rating for the firefox package. Perhaps the fix for that is to reset the votes on all packages every 6 months or something? As far as actual voting though, i think be best option might be a sliding vote scale, possibly like the vote at the Vim website for Vim scripts. Offer 3 "vote options" like: "Great package." "Meh." "Rubbish." and that'll likely give the best idea of package usage/quality. 2009/12/26 Sebastian Nowicki <sebnow@gmail.com>
Hi,
I believe this was discussed on aur-dev some years ago, but it seems that discussion was lost (no longer in archives). I'd like to bring up the subject again. What do you think the best way to indicate package popularity is? The two main ideas were votes (the current implementation) and a download counter. I can't really recall which one was preferred.
The issue has been raised because we're deciding which to use in "AUR2", as a patch has been submitted to implement votes.
I'd like to know if voting works, how effective it is, and how much significance it has on a TU's decision to put a package into community. Basically whether it's "broken" and needs to be "fixed" or if it's fine the way it is.
P.S. I didn't send this to aur-dev as it doesn't really concern the developers. It's an end-user feature, and mostly a feature for TUs, so I posted here.