Your Comments and Feedback on the Desktop Experience AND the Community/Contributor Experience

  • Let user choose between snaps or flatpaks
  • Add ligatures and slashed zero support to Ubuntu Mono font
  • Replace pulseaudio with pipewire
  • Improve battery life
  • Improve nvidia optimus switching experience (not exactly problem of Ubuntu)

Honestly Unity type of computing could be easily recreated with KDE Plasma. I think it is better to stop wasting resources on GNOME, and making it look similar to Unity and switch to KDE Plasma desktop and modify that with Yaru theme/icons, and have Unity Dock with Latte dock, and have a bar at top with global menu.


I’ve selected bits of your paste to reply to; but Ian has already replied expertly in my opinion.

Lubuntu announced they were switching from LXDE to LXQt years before the switch occurred; the only unknown was when. Lubuntu Next ISOs were released for a number of cycles so people could see what was planned, and where the future was (again this should have provided a clue as to what was happening).

Anyone watching upstream would have known it was occurring; after all it a very large portion of the LXDE developers who’d joined with Razor Qt to create LXQt as the replacement desktop; given GTK2 that LXDE needed was deprecated, and the early ports of GTK3 were found heavy and didn’t meet the goal of LXDE being light as was blogged about (by PCMan).

The Lubuntu team even created documentation on converting a Lubuntu 18.04 LTS system into a Lubuntu 18.10 LXQt system; but the result wasn’t always clean, and problems were often encountered; thus in the end it was not supported as we lacked resources (but for anyone who asked about it; I provided the link so they could read what was suggested to complete the upgrade by themselves).

Lubuntu QA-test re-installs over an existing partition; I recently documented for QA-testing impish or 21.10; though the testcases were written for 19.04, so aren’t new (just difficult to find sorry). I regularly use that method of switch systems from one flavor to another; it was one of the things that really impressed me about Ubuntu when I started using it about Ubuntu 10.10. Switching desktops isn’t difficult either; I always used Ubuntu ISOs (quota free download) then added the desktop I wanted to use on top of it (also bandwidth free) avoiding my ISPs quota restrictions for years (they no longer exist), where as downloading a Ubuntu flavor was not quota free for me.

If an ISO contains all flavors it would be huge and no-one would want it. I recall downloading and writing Debian ISOs (fullset) and they used 5-9 DVDRs each release… and whilst it was nice to have the installer ask what packages/desktop/etc you wanted to install; it was a pain to download, then keep swapping the ISOs in the DVD tray during install. Even if it’ll all fit on a 32GB thumb-drive today; it’d still be a HUGE file to download. I think picking the ISO you want to install, then grabbing just that makes much more sense.

1 Like

As an user I’d like if Snap apps asked for permissions when needed, just like Android and iOS. Want to open a file in another partition but the Snap is sandboxed? Just type your password when prompted and the correct permissions will be given. Want to give an app access to your camera and microphone? Type your password when the prompt appears after the app tries to access them.

I realize this kind of functionality might be complicated to implement and might even need support from apps themselves, but it would really help with the user experience and make using Ubuntu be more safe while at the same time giving more power and freedom to the user. That’s what Linux is all about anyway, no?



Something that hasn’t been mentioned, I think it would be great if swap on zram is enabled by default in 22.04 LTS.

Fedora 33+ enables swap on zram using systemd-zram-generator ( and, and if not enabled by default, at the very least systemd-zram-generator should be packaged in Ubuntu’s repositories.

I’ve found swap on zram to be extremely helpful when memory is running low:

  • While running KDE Neon 20.04.2 on an i3 laptop with an HDD and 4 GB RAM, creating 4 GB of swap on zram with zstd compression made a night and day difference for me having Firefox, Discord and Visual Studio Code open at the same time. The system felt far more responsive overall, with no freezes when opening more apps or tabs.
  • A friend of mine installed Ubuntu 20.04 to use Blender, but since he only had 12 GB of RAM, when rendering a project on Ubuntu, Ubuntu would freeze and take 4-5 minutes to complete a render, while rendering the same project on Windows would take a bit over 2 minutes. Creating an 8 GB zstd zram device as swap brought render times for the same project down to 1m57s.
  • I also use 4 GB zstd zram on a Debian server with 4 GB RAM to run a few extra apps without suffering from major slowdowns, as well as 8 GB of it on an Ubuntu server with 64 GB RAM, and so far I have experienced no performance degradation, mainly since the kernel only uses zram when it is about to swap, and swapping to compressed memory space is significantly faster than swapping to the hard drive/SSD.

This is largely anecdotal, but I’m sure these experiences are not only limited to me. Since Windows does something similar by default, and Fedora enables this by default as well now, I’m hoping that Ubuntu can consider doing the same by 22.04 LTS, at least on Ubuntu Desktop, if not Ubuntu Server, since it provides a much better user experience on systems with relatively low RAM, or when a system is under memory pressure, by providing some extra breathing room before resorting to swapping to hard drive.

Hoping you can look into this suggestion, thanks for reading and thanks for all the work you guys put in Ubuntu!


I read the code of conduct, and I find this sentence:

We want a productive, happy and agile community that can welcome new ideas […]

Your way of answering to other’s proposals is cold and does not give desire to join the community

My question is simple though, and it’s not really related to the community: is the desktop a priority for Canonical like before?

I’m following with interest what’s going on with Ubuntu Frame and Flutter, and I want to get involved with the community

1 Like

This is a good idea. It is essentially the default on windows and mac os. I also enable it on all my linux devices.

1 Like

Thank you for your detailed answer, most of which I vaguely understand. As is clear by now I use an OS because of what it enables me to do - communicate, research and write, listen/watch media etc. in the same way that I drive a car because it gives me personal mobility. In both cases I try to understand sufficient of what goes on “under the hood” to keep myself out of trouble and to do very basic preventative maintenance - but I would not attempt to swap out my diesel engine for a petrol/hybrid.

The link you provide on Install using existing partition is I think of potential great interest and I will study it in detail. That however is a “support” solution (for which I am grateful, and for which I may turn out to be very grateful) to what was only meant as an illustration of why in response to a request for “Comments and Feedback on the Desktop Experience” - which I received because I do try to follow some of what is going on on planet Ubuntu - I was trying to advocate a particular route on a road map.

Your reply detailing the needs for massive download files almost makes my case for the need for different thinking!

If I am changing my car, I want the controls to at least remain familiar - approximately the same function and in approximately the same position. Possibly I am wrong in imagining that Linux Desktop is still in the evangelising stage - a minority wanting to attract new followers - in particular disaffected users of two “majority systems”.

I did an OU course on Linux (using Ubuntu 8.04/10.04 in VirtualBox if I remember correctly) and did not like the interface - too much change. I think this was also at a time when the vanilla desktop environment was going through spasms of radical change. However I found flavours and Lubuntu and saw that it was an easy learning curve to get to using the OS to do things that I wanted to do. If I was starting now and was not directed to Ubuntu by the OU course I might have ended up elsewhere.

I was not suggesting an ISO with all flavours (or even multiple flavours) but just wondering if the installation process (which I imagine for most is now download and install rather than buy DVD and install) could install a “base Ubuntu” and then in a late stage of the install prompt users to selected Desktop/Flavour, which would then be downloaded. Will many ordinary users be too bothered about the distinction between a Desktop Environment and A Flavour - is the distinction useful or necessary? The prompt can be a user-friendly guided prompt “Select X for most Mac-like, Select Y or most Windows 7 like, or just try any flavour - Run “Swap Desktop Experience” if you ever want to change your mind - your data and application configurations will be safe!”

Can Ubuntu be developed so that to most users (as in people who send emails, browse the internet and Word-process) the flavour or Desktop Environment is no more than a Skin or Theme that can be easily swapped at will (and with a bit of investment in download and processing time)? I suspect that will be an architectural issue with significant development effort required - can we distinguish between a common base Ubuntu layer and DE layers? But as the original invitation to contribute comments and feedback said “If releases were a roadtrip, this is when we pull over, have some snacks, and find the best way to get to our next destination.”

The advantage is that people exploring Linux opportunities will then see a “one Ubuntu” offering not a further array of choices, and the single installer will guide them to an initial “experience”.

1 Like

Regarding the desktop as first-class-citizen: I see 21.04 and 21.10 (I suspect 22.04 too) already add gnome-shell-extension-desktop-icons-ng, which is basically what I want… :+1:

Regarding Snap: Yes. I’ve reported problems, but they don’t seem to be easy to solve (~1.5 years).

This is a feature that is being being worked on by multiple teams at Canonical, but it is complex and requires changes in both snapd and the kernel and so it will definitely take a good chunk of time but it’s something we’re very excited about and when we have something to share we will make sure to have some noise about this :slight_smile: Due to it’s complexity I also can’t give any sort of timeline unfortunately, but hopefully we can have some beta level things available for testers to try out sometime in 2022.


Push companies to adopt interim releases.
Explanation: For a developer waiting two years for the next LTS is just too long.

Companies in general don’t get it, that newer hardware requires newer kernels. At the same time they often rely on proprietary software that have poor Ubuntu LTS support since their business models are more related to windows/mac desktops or clouds. Pushing for Interim adoption would force 3rd parties to be more supportive.

Provide installer with option to use BTRFS as rootfs

If OpenZFS does get pushed do it also for BTRFS which in my opinion fits better for a developer or user workstation. Don’t get political about that and just remove BTRFS (like RHEL did; but Fedora now defaults to BTRFS;.

Sponsor bcachefs development - that would be awesome!

Provide a tool to install hardware drivers and vendor repositories
Explanation: if you get a windows laptop in a company most people go with default Ubuntu installation and miss installing hardware specific power profiles or drivers and software. Example: Dell XPS ( – get those vendors on board!

Full Pipewire Chain - drop the Pulseaudio server!

Adopt more Clear Linux Patches

Switch over to systemd-boot

Like PopOS! or Fedora did long ago.

Full Systemd Chain – from boot to poweroff

If that’s still an issue - maybe meanwhile obsolete.

Provide a vanilla GNOME experience without the Ubuntu Desktop

Honestly, ubuntu dependencies on gnome packages is crap.

Provide option to replace Chrome and Firefox snaps for advanced users

Honestly, they have still downsides and are not perfect. My Firefox couldn’t connect to GNOME Extensions Host after upgrade to impish.

Drop Yaru and that ugly default themes and get designers on board!

Honestly, ever installed Arc Theme? Know how New Adwaita looks? Yaru is better now, but compared to Windows or macOs or even default GNOME Ubuntu just looks like the operating system someone would run a 50$ tinkerbox – not a high end laptop or desktop.

Be more like Fedora or PopOS!

Honestly, they made cool changes and progress and do focus on technology – something Ubuntu community has lost sight for in the struggle to provide an OS for “everyone”. Now skilled Linux developers are no more part of “everyone”.

1 Like

Rafael, thank you so much for joining us! This is one of the reasons we extended the transition, because Kubuntu and Ubuntu Studio would be affected by this, and this should give all of us time to address the issue before making the full transition for the flavours.

As someone who is still a bit of a new user, this is a great point, and I wonder if there’s a way to make this much more intuitive for someone who might be a first time Linux user or not too tech saavy.

First of all, thank you for registering and giving us such thoughtful comments! I am going to tag @local-optimum in to here, since he really wants to help make the gaming experience better in future releases, and he can discuss your concerns better than I can.

@hellsworth do you have a possible answer/suggestion for the LibreOffice comment? And do you know who might know the answer to the Thunderbird question?

Operius, thanks for joining us and for your comments on theming and our apps, which I know are important to a lot of people. Maybe @kenvandine can offer some insights?

Madwatch, thanks for joining and commenting! Have you taken a look at our snap/snapcraft forum at ? I think those comments could spark some good discussion there - and take a look at the discussion on lol, which is an open version of the snap store still in very early development!

Oh my goodness, we definitely want the flavours like Kubuntu to get the love and attention and support that they need, and while the spinner might not be the most seemingly important thing, it’s the little things that count. Thank you for your comments!

Thanks for joining and posting! Not only are those bugs annoying, but they’re often great bugs for new contributors to work on. Let me find out who to talk to about finding a way to bring that back, and to bring it back even better.

Thanks for posting, christian, and thanks for the encouragement! :+1:

I’m an Ubuntu MATE user for my daily driver, and the Software Boutique is just fantastic at curating apps and making them super easy to find and install. I don’t know what the answer is, but I wonder if it’s possible to have something like that in stock Ubuntu as well.

I know @kenvandine definitely agrees and we want to help community contributors find that documentation, maybe even help expand it, and make some truly beautiful and useful apps for Ubuntu!

Thanks so much for signing up and for your really useful comments, including the one above! Have you joined the Discourse at to let them know about those concerns, because they seem extremely valid and something we should be working to improve.

Hello paulox! Welcome to the community hub! This sounds like an interesting and potentially very useful feature, and thanks for suggesting it!

**Well, it is very late, btu I’ll try to answer the rest of these or at least address the tomorrow. And keep an eye out because in next Thiusday’s office hours we’ll be organizing this feedback so we can make it a resource for future releases!

  1. Better hardware enablement for bleeding-edge hardware.
    I feel that we’ve seen frequently a lot of issues with Ubuntu not properly working with a brand new hardware out of the box it makes me I worry about this when I build my next PC.
  2. Pipewire
  3. Snaps asking for permission, instead of manually plugging the permission.
    It’d be convenient if there’s a way to make it temporary (day-long, one-time, etc.)
  4. Giving desktop better attention
    Nowadays, it feels like Ubuntu’s focus on the cloud market and abandoning its loyal users. By giving us desktop users more attention, I feel this may solve my first feedback.
  5. Better collaboration with flavors & Ubuntu-based distros.
    I don’t know if it’s just me, but it feels like Ubuntu’s in its own little world, even among flavors.

As a KDE person, i have the feeling that snaps/flatpaks&appimages are lacking integration into MUON the great packet manager. I therefore agree with “collaborate more”.

The question for hardware enablement is: Is it enough to switch to a more recent version of the kernel & mesa for new releases and maybe delay the distro release a few days to iron out the last kinks or is it more about forming relationships with hardware manufacturers and explain to them that they need to be more open source focussed and get the driver side going?
I often see a good chunk of a device working years after the production but the community lacks man power and documentation and knowledge of the hardware so components don’t reach their full potential, if being worked on at all.

1 Like

The TB thing is partially solved by Addons
owl Addon from

Kewischs gdata-provider

But as long as they stay paid or a free extension, its difficult for an “out of the box experience” to be good enough. Maintenance is a burden, for external developers as well as admins :wink:

LibreOffice has
but it is moving slow.

3 posts were split to a new topic: MSR and SecureBoot

We know that we need to improve things like our narratives and the website so people know that the desktop very much matters. The recent Flutter developments are a part of that desire to make the desktop better, and I’d invite you to watch the most recent Indabas - the September Indaba was with two of our Flutter engineers, and in the most recent Indaba, we talked about our commitment to the desktop at the end. You can find all our videos on our YouTube channel.

I do think, however, that Ian was being both direct and respectful in his replies. This thread obviously got a lot of attention from newcomers, and one of the most useful things he did (out of many) was to point out the most effective ways to communicate on the Forum.


BTRFS may be OK for desktop users but I wouldn’t trust it on any of our servers. I want to see the Ubuntu (server) installer get better ZFS support, mainly the ability to install to a RAIDZ1, RAIDZ2 pool etc and the partitioning tool should be able to create and delete ZFS partitions. It would be nice to see better BTRFS support too, such as grub-btrfs getting packaged and configured as standard similar to the current ZFS grub history integration.

Even if ZFS support doesn’t make it into the 22.04 server installer, zfsutils-linux should still be installed by default on the ubuntu server ISO because I suspect ZFS is pretty popular amongst Ubuntu server users (I use it on all our Ubuntu servers) but the ISO is useless for recovery purposes without having zfsutils-linux installed, which brings me onto my next suggestion.

The Ubuntu iso has always been pretty useless for recovery purposes. Thankfully, those who install using ZFS have snapshots to fall back on but not everyone opts for ZFS and it can’t help you with some GRUB problems and indeed GRUB is the major issue I have in mind. In an (my) ideal world supergrubdisk ( ) or something that does the same thing would be integrated into the Ubuntu ISO boot menu, grub-customizer would get packaged for the main repos and included as standard (because grub configuration is painful, especially for new users) and Boot-Repair should also be included as standard on all Ubuntu ISOs. Those three can usually fix most boot issues.

I’m also not a fan of snaps. Both snaps and flatpaks are flawed because you need to install an app to use either hence they are not truly universal or portable like AppImages are. I don’t care about sandboxing at all personally, I can use a VM or a container to achieve that. I don’t think any of these three ‘universal’ packaging formats belong in Ubuntu as standard.

Another +1 for Pipewire as standard.

Thanks for asking! Lots can be done to improve Ubuntu and many important items have been mentioned already. Here’s a few that would really make a big difference IMO:

Fix Firefox so that I never see this error message in Firefox again (it seems to be an Ubuntu issue as I haven’t seen it happen on any other OS or distribution):
“Oops we updated your Firefox, please restart Firefox to continue”

There’s multiple software applications which make it really confusing. Some of them share the same logo as well, a grey version and a purple version but one goes to software settings and one goes to software updates. Very confusing. There’s also Ubuntu software. Gnome’s Software application is much nicer. Why isn’t that being used?