[aur-general] How the "Popularity" value is being calculated for a package in the AUR4? - Aur-general - lists.archlinux.org

newer
[aur-general] Signoff report for...

[aur-general] How the "Popularity" value is being calculated for a package in the AUR4?

older
[aur-general] Signoff report for...

Andrejs Mivreņiks

12 Jun 2015 12 Jun '15

9:26 a.m.

Hello! I noticed that the "Popularity" colon has been added to the list of packages. What does this value actually mean? How is it being calculated? -- Andrejs Mivreņiks PGP Key Fingerprint: 3872 5DEB BCA5 9460 09B2 E867 F34B C7DA D782 DAB8

Show replies by date

Marcel Korpel

12 Jun 12 Jun

9:34 a.m.

* Andrejs Mivreņiks (Fri, 12 Jun 2015 10:26:08 +0300):

I noticed that the "Popularity" colon has been added to the list of packages. What does this value actually mean? How is it being calculated?

Lukas on aur-dev: "Popularity is the sum of all votes with each vote being weighted with a factor of 0.98 per day since its creation." Best, Marcel

G. Schlisio

9:43 a.m.

Lukas on aur-dev: "Popularity is the sum of all votes with each vote being weighted with a factor of 0.98 per day since its creation."

thats quite a fast wear-off, i think. after a year, there is virtually nothing left (6e-4 only). does this really meet the intention?

Lukas Fleischer

10:01 a.m.

On Fri, 12 Jun 2015 at 09:43:00, G. Schlisio wrote:

...
Lukas on aur-dev: "Popularity is the sum of all votes with each vote being weighted with a factor of 0.98 per day since its creation."

thats quite a fast wear-off, i think. after a year, there is virtually nothing left (6e-4 only). does this really meet the intention?

Yes. The main purpose of the popularity field is to give very popular newcomers a chance to appear on the front page. Also, we don't care about packages that were very popular a year ago and are no longer used today. If a package is still interesting, it will continuously receive new votes such as is the case with yaourt. Regards, Lukas

LoneVVolf

1:11 p.m.

On 12-06-15 10:01, Lukas Fleischer wrote:

Also, we don't care about packages that were very popular a year ago and are no longer used today. Low number of new votes != not used anymore

Many high quality, useful packages target a specific group of users. Does that make them less valuable ? LVV

Magnus Therning

1:38 p.m.

On 12 June 2015 at 13:11, LoneVVolf wrote:

On 12-06-15 10:01, Lukas Fleischer wrote:

...
Also, we don't care about packages that were very popular a year ago and are no longer used today.

Low number of new votes != not used anymore

Many high quality, useful packages target a specific group of users. Does that make them less valuable ?

No, and valuable != popular :) Any popularity contest is going to be biased in some way. This is as good a bias as any other. /M -- Magnus Therning OpenPGP: 0xAB4DFBA4 email: magnus@therning.org jabber: magnus@therning.org twitter: magthe http://therning.org/magnus

Martti Kühne

1:45 p.m.

On Fri, Jun 12, 2015 at 1:38 PM, Magnus Therning wrote:

No, and valuable != popular :)

Any popularity contest is going to be biased in some way. This is as good a bias as any other.

And let's leave completely out of focus that the bias that is needed for a package to enter the official repositories is the personal opinion of a TU or Dev. The rate recording is optimized to emphasize wide adoption trends, which to me appear completely spot-on. What aspect is missing there? cheers! mar77i

Connor Behan

18 Jun 18 Jun

6:33 p.m.

On Fri, Jun 12, 2015 at 1:38 PM, Magnus Therning wrote:

...
No, and valuable != popular :)

Any popularity contest is going to be biased in some way. This is as good a bias as any other.

And let's leave completely out of focus that the bias that is needed for a package to enter the official repositories is the personal opinion of a TU or Dev. Does anyone know an easy way to hide "Popularity" with a userscript or userstyle? I currently call $("input[type=text]").unbind().attr("autocomplete", "on"); to disable search suggestions since I can't stand websites suggesting that I do

On 12/06/15 07:45 AM, Martti Kühne wrote: popular things. I'm sure this is a useful metric for some, but I'm not fond of seeing a column of almost all zeros in searches. I think the only packages that maintain a high popularity will be things like yaourt where people keep voting it up because they mistakenly think it will one day go to [community].

carstene1ns

19 Jun 19 Jun

3:35 p.m.

Am 18.06.2015 um 18:33 schrieb Connor Behan:

[...] Does anyone know an easy way to hide "Popularity" with a userscript or userstyle? [...]

Given that the index of the column not changes, this works: $('table.results > thead > tr > th:nth-child(5)').hide(); $('table.results > tbody > tr > td:nth-child(5)').hide(); First line removes the header, second line the values. It may not be the best solution, but I did not find a better way than counting to get the right column. It uses jQuery from the AUR page, if you want to use it from a user script you may need a way to inject it. best regards, carstene1ns

Hugo Osvaldo Barrera

20 Jun 20 Jun

2:26 a.m.

On 2015-06-12 10:01, Lukas Fleischer wrote:

On Fri, 12 Jun 2015 at 09:43:00, G. Schlisio wrote:

...
...
Lukas on aur-dev: "Popularity is the sum of all votes with each vote being weighted with a factor of 0.98 per day since its creation."

thats quite a fast wear-off, i think. after a year, there is virtually nothing left (6e-4 only). does this really meet the intention?

Yes. The main purpose of the popularity field is to give very popular newcomers a chance to appear on the front page. Also, we don't care about packages that were very popular a year ago and are no longer used today. If a package is still interesting, it will continuously receive new votes such as is the case with yaourt.

Regards, Lukas

Given how it's calculated, and the purpose it intends to serve, I believe the proper word is "trending", not "popularity". -- Hugo Osvaldo Barrera A: Because we read from top to bottom, left to right. Q: Why should I start my reply below the quoted text?

3243

Age (days ago)

3251

Last active (days ago)

Download

9 comments

10 participants

tags

participants (10)

Andrejs Mivreņiks
carstene1ns
Connor Behan
G. Schlisio
Hugo Osvaldo Barrera
LoneVVolf
Lukas Fleischer
Magnus Therning
Marcel Korpel
Martti Kühne