[aur-general] Proposal: package's votes
Hi all, I propose to make package's votes equal zero after update. All users vote when they update package. Maintainer has true statistic about package usage. Whan do you think about? -- Sincerely yours, Vitaliy Berdinskikh (Виталий Бердинских) xmpp:vitaliy@berdinskikh.in.ua <xmpp%3Avitaliy@berdinskikh.in.ua> skype:skipper.barbossa
Vitaliy Berdinskikh wrote:
Hi all,
I propose to make package's votes equal zero after update. All users vote when they update package. Maintainer has true statistic about package usage.
Whan do you think about?
I don't believe that this is a good idea. People's votes go towards a package and not a specific version of it. I highly doubt that anyone (myself included) would bother to vote after each package upgrade.
Evangelos Foutras wrote:
Vitaliy Berdinskikh wrote:
Hi all,
I propose to make package's votes equal zero after update. All users vote when they update package. Maintainer has true statistic about package usage.
Whan do you think about?
I don't believe that this is a good idea. People's votes go towards a package and not a specific version of it. I highly doubt that anyone (myself included) would bother to vote after each package upgrade.
Agreed. While it sounds like a good idea (get to see how popular a package /currently/ is and not how popular it was last year), voting every time a package is updated would be ridiculous. -- Your Fortune... --------------- "In the fight between you and the world, back the world." --Frank Zappa
2009/1/4 Evangelos Foutras <foutrelis@gmail.com>:
Vitaliy Berdinskikh wrote:
Hi all,
I propose to make package's votes equal zero after update. All users vote when they update package. Maintainer has true statistic about package usage.
Whan do you think about?
I don't believe that this is a good idea. People's votes go towards a package and not a specific version of it. I highly doubt that anyone (myself included) would bother to vote after each package upgrade.
For statistical purposes, would it not make sense to have the system record when a vote was submitted and package version information at the time, but still display a full vote count? That way consideration could still be made based on a general vote count, but a clearly visible lack of votes in the last year or two could be used to ascertain whether it is still considered as useful, as perhaps the lack of new votes means another package has replaced and surpassed its functionality.
On Mon, Jan 5, 2009 at 4:24 PM, Kennedy Skelton <kiirani@gmail.com> wrote:
For statistical purposes, would it not make sense to have the system record when a vote was submitted and package version information at the time, but still display a full vote count? That way consideration could still be made based on a general vote count, but a clearly visible lack of votes in the last year or two could be used to ascertain whether it is still considered as useful, as perhaps the lack of new votes means another package has replaced and surpassed its functionality.
I suppose you also have to keep in mind existing users. You can't vote more than once so what if a package has a fairly large but stable userbase? No new votes but a definite need for the package. -- Callan Barrett
I don't believe that this is a good idea. People's votes go towards a package and not a specific version of it. I highly doubt that anyone (myself included) would bother to vote after each package upgrade.
Also votes are not meant to specifically reflect usage, just interest. The votes do not mean that the person uses that particular software, nor does it mean they think the PKGBUILD itself is of high quality, etc. Resetting the number for each version doesn't make sense for the purposes of tracking usage since that's not really what the votes are for in the first place. Aaron "ElasticDog" Schaefer --
2009/1/4 Vitaliy Berdinskikh <skipper13@root.ua>:
Hi all,
I propose to make package's votes equal zero after update. All users vote when they update package. Maintainer has true statistic about package usage.
Whan do you think about?
-- Sincerely yours, Vitaliy Berdinskikh (Виталий Бердинских) xmpp:vitaliy@berdinskikh.in.ua skype:skipper.barbossa
I think this is bad idea. Users mostly vote when they download it first, at least I do. Moreover votes doesn't depend much on version (Hey, I like foo 1.0.0.92, foo 1.0.0.95 sucks!) and if someone disliked new version, he/she could easily remove his/her vote. Regards Lukas "stativ" Jirkovsky
participants (7)
-
Aaron Schaefer
-
Callan Barrett
-
Evangelos Foutras
-
Ghost1227
-
Kennedy Skelton
-
Lukáš Jirkovský
-
Vitaliy Berdinskikh