Without going into much details, I think it would be quite interesting if we could see the most downloaded/popular packages directly in Pamac (within the last week/month/year). The list may be generated from the logs on the server (I’m not suggesting to spy on our users).
1 Like
There are many different mirror servers. There are no central logs for package downloads.
Still total downloads can be calculated using the sum right?
just like youtube views count work
Why? 
I am sorry, but at the end download counters and a rating systems can not be trusted in any way, because in can be manipulated very easy from the technical POV and are only a marker for felt popularity, but not real popularity. So the best way to see the popularity of packages would be that a package list is gathered on any computer and added to a central database, like that: https://pkgstats.archlinux.de/
For example this statistic would be useful: https://pkgstats.archlinux.de/fun
Clarification: This is only for archlinux and not manjaro.
1 Like
This subject was discussed before, not long ago:
Hello,
While executing systemctl start pkgstats.timer, I get the following wrror message: Failed to start pkgstats.timer: Unit pkgstats.timer not found.
@jojopara Because it is for archlinux only as I can see and not for any archlinux based distro like Manjaro.
1 Like
There is no sense in making such functionality and why would it?
Just because a number of users finds Discord is matching their requirements - it still is the user’s requirement that matters.
Are you not able to have your own opinion?
I mean - are you the kind of person that is easily manipulated into having any given opinion - just because the majority has the opinion or because some social media influencer tells you what opinion to form?
That is not - by any measure - a way of living your life.
I can highly recommend forming your own - informed opinion - whether this is the car you are driving or the computer you are using.
7 Likes
There is no sense in avoiding such functionality nowadays, where everything else has it and UX already relies on it.
We don’t decide blindly according to such list, but we could narrow down our selection more easily. It may become an additional tool for our decision making. E.g. quite recently I was looking for a note taking app. I spent lots of my time to find something useful and easy to use and none of those apps worked for me. Lastly, I ended up with Joplin, but even Joplin has some issues that I still don’t like. I don’t have time to spend on “my own research” and try all apps available out there.
Such information would surely help us to check say - the top 5, assuming the community may be larger around those projects, someone else has tested it before us, then spread the word, so we can spend less time searching for an app that we may or may not like. But the chances that we may like it will be a bit higher.
How do we decide to rent/buy/use anything today? Without those star ratings and popularity trends we would easily say “Welcome to 1996”. It simply has an added value that we rely on today.
Manipulation issues as a show stopper? I don’t believe I just read it on a technical linux forum that suppose to resolve such issues.
2 Likes
Pamac offers a reduced choice of applications in Pamac recommended apps, or in Manjaro Hello, but at the end, the user is free to choose (general maxim for GNU/Linux anyway).
There are websites more specialized at this kind of comparisons, which-is-the-best application.
https://www.slant.co/topics/6303/~note-taking-apps-for-linux
https://itsfoss.com/note-taking-apps-linux/
https://www.fosslinux.com/26068/the-5-best-onenote-alternatives-for-linux.htm
1 Like
There’s no problem enabling the service and timer (I did that long time ago and it’s still running 
), as you also say it collects stats for Arch, but they are “dirty”, as they contain stats from other Arch-based distros as well, you can find there mhwd and other Manjaro-specific references for example.
Just my 50cents:
You know, all these popularity ratings boil down to giving more and more work to the end user. Instead of asking a person who knows the subject matter, you rely on the mass of people, because the majority can’t be wrong, right? If that were the case, people would still believe today that the earth is a flat disk.
The popularity indicator is no more than betting at a horse race. If everyone says that horse X will win, it doesn’t actually mean that it will. But spreading the information that horse X will win triggers a chain reaction so that everyone bets on that horse, but only those who rely on expertise can actually assess how likely it is that that horse might win.
So in the end, the reliability of the mass psychological popularity effect is rather a hindrance in assessing whether the program is any good or not. Better ask a person who uses programs you are looking for. You will usually get a much better answer.
I hope you understand the thought process.
In any case, you can emphasize on selection of recommended programs, which have proven themselves in the past with experienced users, in the package manager, but to rely on rating and popularity as a selection criterion, I think is nonsense, even if it is modern and trendy.
In some areas popularity does make sense:
- Prioritizing work
- Enhancing the image of a company and becoming better known.
- You can direct masses the way you want.
- and so on…
But I think for the consumer the clear disadvantages simply outweigh the benefits. Why else would large providers offer such a system? Because it offers them the greatest possible advantage and not to the user, but they sell it as user friendliness.
I think I was quite clear that the popularity is not a final decision maker, but it surely may be one of the parameters to narrow down one’s search. I understand your analogy with “following the majority may be also dead wrong”, but if you want to go down that road, then we should open another, a very broad and important topic called: DECISION MAKING.
We could cite Socrates on how popular votes is useless and not efficient even in modern societies and that would be right, but I’m not talking about the algorithm to establish the decision making! The popularity of something may be just an indicator which sometimes may be helpful and sometimes not as much - it will be up to us to decide (according to “our own algorithm”).
Besides, they do display the betting odds before most of the sport events, don’t they? That doesn’t necessarily mean that the favorite will win. But it would be very interesting to see the statistics on how many times the betting favorites actually won at the end. If the number is higher than 50%, then it may be a valid indicator. Just as many other indicators in the line.
That is just one person. That person may hate something that I absolutely prefer. Why not asking 10 different people who use it? Or maybe even thousands of them? That’s exactly how the star-rating has been invented.
I don’t entirely disagree with you on this. I deliberately suggested popularity indicator, as it is something trivial. Even though I am aware it may add a smaller percentage of value to the decision making process, I would not dismiss it that easily. The real value comes with something you just mentioned between the lines, that requires way more work and is quite accurate: we are not all equal! Our own knowledge and experience makes us all different. How to evaluate each contribution/opinion/decision? Definitely not by popular voting and I completely agree with you on that, but it would require a hell of a lot more of additional work and is surely even less suitable for Pamac. 
This topic is quite interesting. Even the European Commission addresses this broad topic as “competence evaluation” and still working on it. But maybe the Open Source communities could learn how Quora, or StackOverflow do it and improve their existing ways of dealing with the same topic. But that may be quite a long term goal.