Solus repositories small or big?
- Edited
murbert Debian & it's variants have something like 50,000 native apps (.deb). Solus has far less.
Most of which are not designed to current professional standards or rigorously maintained. The fact that applications in the Solus repository are curated, and as a result, fewer in number, is an advantage, not a disadvantage.
tomscharbach yeah, and it's strange that reviewers don't talk about it. About the quality of repositories. Always it's a short summary that there are less packages compare to e.g. Fedora. Sometimes they install some packages/apps and they discover that Solus have the basic ones. I'm just wondering what is missed here. As I said from my testing Fedora is lacking for a desktop user like me more native apps than Solus. I would understand if the review consider to be about distribution as a server comparison.
Solus tends to focus on applications for general desktop computing, so a lot of the more specialized tools are not available. Meanwhile, the Arch / AUR repos and as already mentioned, the Debian repo are quite expansive. I generally maintain at least one machine with Manjaro / Budgie, in case I need some off the wall piece of software. If it's not in the AUR, it's nowhere.
- Edited
WhiteWolf It's strange that reviewers don't talk about ... the quality of repositories.
I doubt that reviewers, who are journalists writing under deadline, have either the time or the expertise to evaluate the quality of repositories, and are, for that matter, probably indifferent to the issue. Mostly what reviewers do is rewrite blog announcements and echo conventional wisdom.
I posted a recent Solus review article link, that was positive, on this forum, not long ago. The reviewer said he was "surprised" everything he "liked and used" was in the repo and remarked on how "large" the repo was.
'Small repo' was a knock on Solus from way back when it started and many reviewers just parrot each other when "researching" their "review' of the live iso..(Plus the .deb thing).
If that Linux guy I quoted thought it was large enough then more reviewers will in time.
tomscharbach The fact that applications in the Solus repository are curated, and as a result, fewer in number, is an advantage, not a disadvantage.
+1
The good question to ask yourself is which software do you use is missing in the Solus repository.
In most cases, when you can't find a software, the team or the community can suggest an alternative. You still have to possibility to make a package request.
Quantity isn't an indicator of quality.
Another consideration: The base repos often do not tell the whole story. Adding repos in Fedora and Debian derivatives is fairly easy, as well as many sites offer a .deb and/or .rpm file for side loading.
The simple fact is there are a LOT more available programs in these distros than in Solus. Some are good, some old, out of date, uncurated & crappy.
I am on Solus for a reason
- Edited
kyrios The good question to ask yourself is which software do you use is missing in the Solus repository.
Assuming that you include Flatpak (e.g. Zoom) and Snap applications as part of the "repository", absolutely nothing, except for Microsoft Edge. Microsoft will probably release an Edge Snap eventually, because Canonical is pushing hard for Snap development.
Solus was designed for ordinary desktop users, and the available apps are more than sufficient for that purpose. The fact that I can trust that the apps curated by the Solus team are properly maintained and reasonably secure is a plus. I don't intend to start reading source code or spend my life checking Githubs.
kyrios Quantity isn't an indicator of quality.
Usually the opposite, in my opinion.
- Edited
I honestly really like how things are today. Maintain strictly necessary packages and a little more and have Flatpak and Snap at your disposal. I had not had my PC so stable before, XD .
A heartfelt thank you to the maintainers.
Big repositories in my opinion will be less and less relevant. I use more and more flatpaks and I think the usage of flatpak in general is probably growing strongly.
kaktuspalme I'd agree. I feel like flatpak adoption is outpacing snap or appimage. This is not scientific and purely based off observations from podcasts, other Linux media, and the inclusion of flatpak into many distributions own app centers.
Brucehankins that's a great observation. if you ignored podcasts etc and looked at the solus forum as a fishbowl, I'd say flatpak is the front runner here, too, when given appimage and snap options. I confess to two appimages I can't find in flatpak, but I have no snaps at all and Flatpak for three necessary things. It all works out. I think solus core team was ahead of conventional wisdom when they eyeballed Flatpak to someday be the repository of all things third party (someday) as well.
- Edited
Solus repo is not small, we have over 4k source packages, you have to remember that we do not split packages as much as other distros.
As an example run eopkg info qt5-base-devel
and see all the pkgconfig stuff, many Distro do split them into an own package we do not
- Edited
Girtablulu Solus repo is not small, we have over 4k source packages ...
I suspect that the only folks who characterize the Solus repository as "small" are reviewers who get background from Solus reviews written 4-5 years ago and can't be bothered to update the facts.
Nonetheless, the Solus repository is significantly smaller than .deb/.rpm repos, and that is a plus, in my opinion. The Solus repository is focused on apps that are needed for ordinary desktop use, which cuts the number of packages by eliminating server and niche packages, and is curated to eliminate the chaff, which dramatically cuts the number of packages by eliminating outdated, poorly maintained, poorly conceived and poorly designed apps.
Less, in this case, is more.
Except with respect to the kernel, which is tightly controlled, the Linux community prides itself on allowing anyone to code and distribute just about anything, whether or not it is useful, well designed or properly maintained. I don't think that is necessarily a strength. We do not need 300+ distributions, most of which are ego-driven niche projects, or 50,000+ apps, most of which are used by a handful, at best. It diverts time and talent from creating a professional-quality Linux desktop ecosystem. Torvalds observed years ago that Linux would not gain a significant market share unless and until the Linux community developed enough self-discipline to focus on quality rather than quantity. Time has proven him right, in my opinion.
tomscharbach I suspect that the only folks who characterize the Solus repository as "small" are reviewers who get background from Solus reviews written 4-5 years ago and can't be bothered to update the facts.
Not to mention many of these reviewers you mention 'review' Solus 'comprehensively' in live .iso mode, believing that is representative of a full install.
tomscharbach That can also be a disadvantage. Sometimes having a somewhat crappy app is better than having no app.
With that said, there's potentially a better solution.
That is to add some sort of a "verified" badge to apps that are verified to work well with the latest version of Solus.
This badge would then have to be earned by the developers of said app by making sure it remains compatible with the OS, and if they don't then they can lose their badge.
This way we could grow our repos, but also keep an eye out for quality.
- Edited
FanboyStudios With that said, there's potentially a better solution. That is to add some sort of a "verified" badge to apps that are verified to work well with the latest version of Solus. This badge would then have to be earned by the developers of said app by making sure it remains compatible with the OS, and if they don't then they can lose their badge.
Some distributions do this. Ubuntu Budgie, for example, provides basic information about each app in its repository:
I don't know who does the analysis/review, but someone responsible for the distro -- presumably the responsible maintainer on the team -- is taking a developer-independent look at the apps.
I'm not sure that self-reporting by the app developer would be sufficient, because Solus is updated on a rolling basis, and team-based maintainers are more likely to spot potential problems than a disconnected app developer not involved in week-by-week development/maintenance.
I'm not sure how much work by the Solus team would be saved by self-reporting in any event. If nothing else, even if Solus elected to rely on self-reporting, someone on the team would have to audit app-developer compliance.
I've never been involved in distro development or maintenance, so I might be talking through my hat about the issues presented by your proposal. But I trust the Solus team to provide a high-quality, rock-solid distro on an ongoing basis, and that is the reason I use Solus and support Solus financially.
FanboyStudios This way we could grow our repos, but also keep an eye out for quality.
A "verified" badge, as you note, means only that an app is "verified to work well with the latest version of Solus". It says nothing at all about the quality of the app.
FanboyStudios tomscharbach I think quality is a relative term also, in some areas at least. Things like ux (appearance, layout, design) may be of high quality to one set of users while also being of poor quality to another set. Strictly looking at functionality could be a better option. Does it do the thing it says it does without breaking? Do all features or functions work properly as stated? Functionality of an app is tested and verified by maintainers and team members before an update is pushed or an app is included (my assumption) in the Solus repos. This somewhat alleviates the need for a verified badge.
Also there's nothing stopping someone from installing via flatpak, snap, appimage, or building from source. So expanding the repos just for the sake of having more apps available even if they are "somewhat crappy" doesn't make the most sense.
- Edited
Brucehankins I think quality is a relative term also, in some areas at least. Things like ux (appearance, layout, design) may be of high quality to one set of users while also being of poor quality to another set.
That is certainly true, but I would submit that generally accepted modern design standards exist for GUI apps, and that goes beyond personal preference.
All major Windows, MacOS, Android and iOS apps, although often quite different in appearance, layout and design, are remarkably similar in terms of meeting modern design standards. Major Linux apps (e.g. LibreOffice) do so, as well.
But many Linux apps fall far short of the mark, designed, it would seem for 2001 rather than 2021. The apps are functional, but not up to the design standards that Windows, MacOS, Android and iOS users expect.
I use LinSSID, for example, a completely functional app that seems stuck in the Windows XP days in terms of design, when compared to competing Windows/Android/iOS apps with the same functionality.
The outdated UI** doesn't bother me, but I think that a Windows/Android/iOS user, comparing the apps that they use on those platforms, would look askance.
My comments about quality were made in the context of Torvald's remarks a decade ago to the effect that Linux would not gain a significant market share unless and until the Linux community developed enough self-discipline to focus on quality rather than quantity. Torvalds was speaking in context -- to gain market share, Linux has to directly compete with Windows and MacOS, and I think that Torvalds, when thinking about quality in that context, was speaking about then-current generally accepted GUI design standards. At least that is how I interpret his comments.
A significant number of Linux apps were behind the design curve then, and it seems to me that, in light of developments in both Windows and MacOS GUI design over the last decade, not to mention the advent of Android and iOS apps, that a significant number of Linux apps are well behind the design curve now. Users expect apps to meet modern design standards, and are right to do so.
Brucehankins Strictly looking at functionality could be a better option. Does it do the thing it says it does without breaking? Do all features or functions work properly as stated?
I think that functionality (in the sense that you are using the term) defines the bare minimum threshold of what is acceptable. Any app that doesn't meet that minimum threshold should not be in any repository, period.
Brucehankins Functionality of an app is tested and verified by maintainers and team members before an update is pushed or an app is included (my assumption) in the Solus repos.
That's my understanding, too, and one of the reasons that Solus remains rock-solid.
**Edit/Update:
By way of comparison to LinSSID, this is the app used at the railroad, available for Windows, Android and iOS: