Re: [nongnu] main 74116339a8 2/3: * elpa-packages (anzu): New package

[Philip Kaludercic]

I agree with all of this, but this seems to imply that some sort of
identification would be necessary (I would guess automatised, using some
kind of public key cryptography, instead of manually with accounts).

Jean Louis <bugs@gnu.support> writes:

>> > I also think we should have a "rating" system in place, which should
>> > also somehow include information on which version/date the user rated
>> > the package in.
>> 
>> Would also be nice. 
>
> While it is good idea, it would be good to add the care of genuity of
> such ratings.
>
> If such rating system goes over HTTP anonymously and without
> registration, that opens door to false statistics, thus false reports.
>
> It is good to have genuine rating system that would tell something
> about background of the rating.
>
> If such rating is automated then full description of the rating system
> in simple English shall be provided.
>
> Let me give the example of what I think is false report based on
> download count on melpa.org domain, that is similar to ratings
> system. It gives download count.
>
>> dash, A modern list library for Emacs, 20210826.1149, github, download
>> count being 3,951,288
>
> That line alone is vague to its meanings:
>
> - I am asking myself, was the version 20210826.1149 downloaded
>   3,951,288 times?
>
> - or was that the total download count for all versions?
>
> - then I am pretty sure that "dash" is not a package users would
>   download manually, rather automatically
>
> - it appears most popular package by download count, while it is
>   probably not well mentioned package among people in their written
>   and verbal communication;
>
> Thus ratings method in ELPA shall be described in public for people to
> understand the background.

-- 
        Philip Kaludercic