I kinda found it by accident… VMware announced the “free” VMWare player (decades before they enshittified themselves) and a “VM MarketPlace”… You could browse a download list of various prebuilt VMs - one was called “Browser Appliance”…
That’s all it did - ran a Linux distro with Firefox… I dug deeper - it was a distro I’d never heard of before (I’d been using Slackware, then Red Hat, then SLES [Suse Linux - not OpenSuse], and had tried Debian once or twice) - so I took a look at Ubuntu…
It was great! Installed it on a Toshiba laptop - and - EVERYTHING WORKED! Everything including WiFi (I suspect it probably used that old method of loading a Windows driver : NDISWrapper?)!
Installing Windows XP on that laptop took hours locating and downloading the drivers (and a lot of hit and miss with the wrong drivers!).
Historically, it was always changing the UI for no apparent reason.
It insists on many snaps which are (it says):-
“Snaps are app packages for desktop, cloud and IoT that are easy to install, secure, cross‐platform and dependency‐free. Snaps are discoverable and installable from the Snap Store, the app store for Linux with an audience of millions.”
So massively duplicated software which can only be found in just one place - Canonical’s website.
To add insult to injury, installs of a .deb package can, in certain cases, actually cause a snap to be silently installed in its place e.g. Firefox, Chromium.
So trust of this organisation is, quite reasonably, zero.
I thought the word ‘bias’ sort of implied an unreasonable ( ie illogical) attitude.
Anyway, your arguments are reasonable.
My main issue with Ubuntu is it uses systemd exclusively… there is no choice of init system… but I have the same issue with Debian and Arch.
It is reasonable to have your opinion and your reasons.
I don’t think they change the UI willy nilly. It seems like Gnome, KDE, Windows, and others change the UI and tick off users. They went from Gnome to Unity and ticked off a group of users. Then they dropped Unity and went back to Gnome and ticked off a group of users. That was part of my point. They can’t win for some reason. There are also official spins for KDE, XFCE, Mate, and others.
Snaps are pretty good stuff. Similar to Flatpaks. Somehow, I don’t hear the same complaints about Flatpaks. Snaps run really well. As far as where they come from, I install software from my distro’s repos anyway. It is possible to just download a deb and run an install, but I’d sure rather get it from the official repo, in this case, Ubuntu.
Yeah, the .deb installing a Snap doesn’t seem quite right. Not sure why they do that really.
I trust them myself, but you do have reasons. Just be open to re-evaluating your reasons please.
@nevj , I agree that there is nothing wrong with Ext4. Before I switched to ZFS for data volumes, I was using the following:
mdraid
dmcrypt/LUKS
LVM
Ext4
I build ZFS in RAID 10 (a pool of mirrors), and ZFS adds the following:
redundancy with bitrot protection
encryption
compression - Typically, QCOW2 files (virtual disks) are reduced in size by approximately 40% using ZFS’s LZ4 compression with no performance penalty.
I am using filesystem type datasets–not ZVOLs. I create a new ZFS dataset for each QEMU-KVM guest, and within that dataset create the QCOW2 file(s) using “qemu-img” to my specifications.
Everyone gets their Flatpaks from Flathub. It’s just as centralized. Plus, there is nothing stopping another source from creating their own store or hub for Snaps. Snaps are an open format.
Canonical is not trying to replace the system package manager. You can still use the graphical updater, apt in a terminal, snap, flatpak, and appimages all side by side.
Flatpak uses OSTree-based deduplication, allowing identical files and runtimes to be stored only once across multiple applications, reducing disk usage when many Flatpaks are installed. Snaps achieve some storage efficiency through shared base and content snaps, but do not generally deduplicate files across different snap packages.
I dont like qcow2 snapshots … they are not separate … they add stuff to the qcow2 file. … I use virt-clone instead of snapshots.
You have a big system. I guess you need ZFS. For me it would be just another complication … I do have multiple disks, including hotplug disks, and at last count I think I had 38 partitions, but I seem to manage OK.
and this. That flathub is the de-facto standard, does not make it the only flatpak repository out there. Someone here mentioned in another thread the Fedora project hosting their own flatpak repository. There are more.
Yes, there are several hurdles to be taken:
The source code of the snap repository software is not published.
The snap system is able of atomic updates. This translates into it being non-trivial to implement alternative repositories, and the developer needs to build their own snap repository software from scratch, because canonical does not publish this software. Plus they need to alter the entire snap package to accommodate for the atomic updates — which they probably made sure is also non-trivial work.
Here’s some criticism from the Linux Mint maintainer.
The whole fact that the snap system is a centralized system is a huge problem for several distributions. Flatpak hasn’t got that problem.
The Snap format itself is completely open source. The only proprietary part is the store. If you’d like, I can spend a few AI tokens and create a repository for demonstration purposes.
I’m not claiming there aren’t alternative Flatpak repositories. There certainly are. My point is that I don’t personally know anyone who uses anything other than Flathub. You’ve correctly pointed out that Flathub has become the de facto standard. In practice, the Snap Store occupies a very similar position in the Snap ecosystem.
It’s perfectly fine to dislike Canonical or Snaps. I don’t. It’s a bit like Coke vs. Pepsi or Ford vs. Chevy. People are free to choose what works best for them.
Maybe I should even mark this as the solution. At the end of the day, I believe Linux and the open-source community are stronger because Canonical exists and continues to invest in the ecosystem.
@nevj , yes, the snapshots are added to the QCOW2 file(s).
I read about “virt-clone”; however, it appears to be more for cloning an existing install for duplication. Yours is a clever use for it. In my pre- QCOW2 snapshot days, I would simply copy the relevant QCOW2 files to another location for testing something.
I have found nothing better, faster, or more reliable than QCOW2 snapshots for rapid testing, revert, and repeat until you get it right.
Yes. If I am going to try adding something experimental, I make a clone, then I have something to go back to if I mess it up. It is sort of like forking.