I've been thinking many things related to Pamac for some time, and finally pushed to post my thoughts due to recent events.
Super fast context: at home we are Arch+derivatives users: me running Parabola/Arch/Manjaro and others running Manjaro only. Those who run Manjaro only happen to not know anything about computers (nothing wrong with this). We all run the vanilla GNOME desktop (no extensions or theming).
Recently had my parents hooked on a "Checking conflicts" for too long during a Pamac update.
This worries me, as it made GNU/Linux look yet again seem like a "hacky unstable OS", uncomfortable for them to use. The truth is, the background process is not the problem (it would happen in the terminal as well), but it is how it was displayed to users.
The problem here I think is not Pamac design alone, but the existence of Pamac, as I will describe in this longer post:
Pamac is too verbose and convoluted, even possibly scary to complete beginner. On the other hand, Pamac is too "clicky" for power users (plus insert all reasons power users tend to prefer the terminal). Pamac users can very much use GNOME software if they wish to use a GUI, or jump into CLI/pacman if they want to learn more. The middle option sounds good, but doesn't seem to fit any practical case, apart from installing many packages at the same time or accessing the AUR (more on that later).
Another problem which I don't know if is still present is that Pamac relied on an extension to show the user whether updates were available. While there were no updates, the icon would remain visible, while offering no information. This is a step down from GNOME software notification system in which just a system notification is used: clean, informative, and without the use of extra tools.
GNOME Software problems
It boils down to two problems:
One is that it consumes system resources, varying from 150 MB of RAM to almost 500 MB. The other one is that it doesn't have AUR support.
The thing with this is that GNOME desktop already implies more modern hardware for the normal user, and their developers don't hide it. There is the exception if you know what you're doing, but I'm leaving this case aside.
Also this might be just coincidental, but I usually have to remove the updates notifier extension from the top bar as I get the comment "Why is my desktop full of -things- ?" (just n=3).
This means that to get notifications from updates in way they can easily click, use and understand, I need to have running GNOME software. Plus if I want to give AUR support to the novice user, I need Pamac running as well.
I have requested people to install software in their Manjaro machine, leaving them the option to use "the bag app" (GNOME software) or "the boxes app" (Pamac).
In all cases (n=2), GNOME Software was preferred because it follows a familiar and good paradigm: you see you "app storefront" and choose from there or search for something, and install it.
Also you can see what's installed and if there are updates from well defined sections.
Compare to Pamac presentation mix of installed and available packages.
If you click them to install, they don't, the interface displays there is an "operation" and you have to "apply a transaction".
Again, I'm discussing on the grounds that Pamac is shipped as the default installation tool in a user-friendly distribution.
Arch based distributions may attract users because of the release schedule, simplicity, modernity, etc. But none of those are exclusive to Arch-based distros. One of the ultimate only reasons one user might be on Arch instead of Void or Gentoo is the AUR.
In a Arch-based distribution that trades user-centricity for user-friendliness, a GUI should be provided.
This seems the partial role of Pamac, and the reason why it still has a place, but not as it is today.
My proposal would be to repurpose Pamac as an AUR graphical frontend that follows the GNOME Human-Interface guidelines, while making clear that the software it offers is not officially supported.
The AUR offers killer app reasons for some people to switch to GNU/Linux, as they find the software they were used to and more, easy to install.
Summary of the porposal
- Drop Pamac support for normal packages in the Manjaro GNOME edition in favor of GNOME Software.
- Maintain AUR GUI support either by integrating it in GNOME software, repurposing Pamac as an AUR tool, or use/adapt another AUR GUI developed elsewhere.
- Drop GNOME extensions that check for updates.
- Reinvest all these saved resources.
The possible question would be if Manjaro is right for these people (extreme beginners). I've seen less annoyances in Manjaro compared to its Ubuntu-family counterparts.
For many users, the fact that they will adopt GNU/Linux is only a matter of three specific things: looks good, works out of the box, and they have plenty of software available. In order to have plenty of programs available at a click, I need them to access the AUR in an easy way. Plus I can use all Parabola/Arch tricks with friends and family, too.
I hope it's useful, or at least some food for thought.