GNOME Software and KDE Discover support both Flatpak and the distribution's native package manager together. For example implementations on Ubuntu-based distributions that work out of the box, see the elementaryOS AppCenter and the Pop!_OS Pop!_Shop:
Curious... do you think that the world has moved on from the "download from website and run installer" model? Obviously that has serious drawbacks from a security perspective, but up until less than a decade ago, that was the only way to get software on Windows and macOS, and they were perfectly mainstream.
I think the majority of users out there don't understand sandboxing or permissions models or why they are useful. While I think those are the users that probably benefit from them the most, it doesn't follow that we need these things before average users will embrace a particular platform.
> A lot of people are mad at Canonical for not open sourcing the backend but no one seems to be offering to build one.
You mention Flatpak and then two sentences later claim this?
I think maybe you're just missing the point. I have been using Linux as my daily driver for a good two decades now. I don't care one bit about Linux being a mainstream desktop distro. Sure, if it was, new (and even not-so-new) hardware would get better support, and that would be a win. But getting that is not a fair trade off if what we have to accept in return is a closed-source, walled-garden software delivery model. Hard pass, no thanks.
For the most part, the only people who really care about "the year of the Linux desktop" are people trying to build a business around desktop Linux. Those aren't the sorts of people I want making decisions like this, but, unfortunately, they're often the kinds of people who have the ability to force these things on the community.
Ubuntu seems to have made clear that the snap client will not be able to connect to alternative stores. What is the point of building an alternative store.
Further snap is controlled by one company. Canonical has a history of developing groundbreaking new things for the Linux community before unceremoniously killing them off.
* Ubuntu One * Unity Desktop * Ubuntu Touch * Upstart Init system
Do you see a trend here? There is a reason the community is skeptical of something completely controlled by Canonical.
Ubuntu One file syncing, also server part, has been Open Sourced with much effort from Canonical: https://ubuntu.com/blog/ubuntu-one-file-syncing-code-open-so...
Ubuntu Touch https://ubports.com is run by community and has rapid progress, it's running on PinePhone etc.
Canonical would have liked to continue Unity: https://www.omgubuntu.co.uk/2017/10/why-did-ubuntu-drop-unit...
I would presume Upstart Init system was changed to systemd because other distros started using systemd.
I don't see a trend here.
As another long time Linux guy, I have a feeling most of those obsessed with Linux becoming a mainstream desktop platform are Mac or Windows users.
They just want to get Chrome and VSCode running without any headaches.
I used to care, nowadays I just focus on Apple, Google and Microsoft platforms instead, and leave Linux for cloud VMs.
Flatpak doesn't impose closed-source, we have distros which only use open source flatpak repos.
Models like Flatpak have the advantage of "download from website & install" i.e you get the apps directly from the producer (which is better for security) while having no walled garden app store but having the advantage of auto updates & permission systems
I do not like Flatpak for technical reasons but Flatpak itself is in no way a walled garden solution, the problem is Flathub.
> I don't care one bit about Linux being a mainstream desktop distro
If you still want to be able to run Linux at all (not even on new hardware) on a consumer hardware, then you should care.
- Linux doesn't run (decently) on Microsoft latest surfaces - Linux won't run on new Apple hardware - Linux doesn't run on Android phones (with a few exceptions) - once Google switch to Fushia no more Linux kernel on mobile hardware - on the long term Huawei will ditch Android and other OS for their own thing.
Linux will die on consumer hardware (except on servers) if it does not become more popular. According to Apple's keynote the future of Linux is in its cage, in a VM running on walled garden MacOS. According to Microsoft the future of Linux is WSL, running on top Windows.
If you do not like walled garden then you should definitively care about Linux getting more mainstream before no comsumer hardwares support it anymore
Why? Desktop computers worked just fine for 35 or so years before app stores showed up. It's completely unnecessary.
I'm not sure that "desktop computers" as we know it were around in the 60s (I certainly wasn't), but in the early 90s software tended to come from floppy disks, before that it was a C90 cassette on a spectrum.
The iPhone launched without the AppStore. The AppStore was not the cause of Apple's success, Usability and Power were - in the sense that it allowed users to do more things (access the web from anywhere) with less effort (no booting times or ceremony). It was the first mainstream device that made it really easy to browse the web; that is what made it successful.
The AppStore was just a play to entrench that initial success by coopting third-party developers and ensure they could only sharecrop on Apple's land: "If you do anything serious with our platform, we get 30%". That's it. Nothing to do with usability. The AppStore is actually a terrible model from a usability perspective, because it collapses discovery so much that it does not scale. "Tap and run" could have been implemented as a web protocol without a centralized UI and it would have worked just fine.
The same is true of Snap. Linux does not need a centralized store, from Canonical or anyone else; that's just a play for them to get a pound of flesh from third-party developers. It has been tried so many times before in the Linux world, and it has never succeeded. Even Nokia threw a lot of money at it, and failed.
What Linux really needs, as you say yourself, is to make it easy for developers to guarantee integrity of their solutions once they reach the user. That's why Snap is getting some traction: because it promises to move power from packagers to developers and to make it easier to guarantee that a program will Just Work, without having to deal with the packaging vagaries of this or that distribution or risking that an overzealous maintainer will "improve" your stuff with crippling patches.
If Canonical were sincere in their attempts to make Linux succeed, they would concentrate on that side of things and drop the AppStore shenanigans. But that's not what they are really after.
Also, it's not like modern developers don't sell users down the river. In fact, that's one of the very few sustainable business models that the AppStore forces you to have.
I do not want that to be the target. Going mainstream never helped with the quality of anything. Quite the contrary: mainstream means lowest common denominator.
If you want to compromise on how much you control of your desktop os for more vendor support you already have windows and macos.
If this is what you're looking for out of a computer, then why not just use a phone, tablet, or something like ChromeOS?
Sometimes there can be tradeoffs between convenience and security, but in the case of an app store the convenience actually enhances the security, because it diminishes the chance of people running random harmful commands from the Web.
[0] https://blog.system76.com/post/616861064165031936/whats-new-...
Who says it needs to be? What if what appeals to Linux desktop users is precisely what differentiates it from the mainstream?
Snap is regressive. What we have now works better, faster, lighter, and more secure. No, you can't have a convenient means of shipping your proprietary electron 10G garbage monster to end users. We don't need it to be "mainstream".
https://en.wikipedia.org/wiki/Synaptic_(software)
Sandboxing is a laudible goal but curation is more likely to result in a system that isn't pwned in the context of current sandboxing being at present being vastly insufficient to contain real threats.
I'm not sure in what fashion flatpak is behind snap given that the former supports custom software sources served over simple existing servers, local and offline installs, runtimes, fine grained permissions via portals, and turning off automatic updates.
>A lot of people are mad at Canonical for not open sourcing the backend but no one seems to be offering to build one.
This is a very strange defense to a legitimate criticism. In what universe would anyone on earth offer to spend their own time to fork snap and write their own backend instead of using flatpak, appimage, apt or 17 other choices.
Needless to say, that was not enough for Linux to go mainstream.