Rethinking Ubuntu Desktop: a more thoughtful default installation

I’m not commenting on your mom’s tech aptitude (though I’m glad to hear she’d helped contribute to Ubuntu). I was referring to your suggestion that we keep in mind that things “we” do “should pass the mom test” (you didn’t say “my mom test”). Words matter. Plenty of tech-savvy moms out there might feel a bit meh being pigeon holed as a litmus/shorthand for “basic”.

That’s all I meant (and you don’t have to agree).

Something like “new user test” or some such would make the same point.

1 Like

then you didnt read the paragraph til the end :wink:
but i’ll watch my wording in the future …

…and yes, she has been testing installs and desktop UI changes for about 10-15 ubuntu releases and gave very good feedback i wouldn’t have gotten without her (resulting in launchpad bugs)… she’s one of the reasons i do care about making my own UIs userfriendly as well and i would love to see more people caring about making their software usable (and tested) by their moms… there is absolutely nothing disparaging in this and i dont get how anyone could think this …

PS: this is about driving development by love, not about badmouting tech savvy moms …

1 Like

I don’t see how this approach offers a more “thoughtful” installation. It sounds a bit like a solution in search of a problem. Presumably there’s some user-research to show people are struggling with the choice of “minimal installation” and “full installation” and the impacts thereof?

And if shipping LibreOffice and Thunderbird by default is affecting the usability of Ubuntu, does the latter solicitation of whether to bundle in software like GNOME Weather and Solitaire, help or undermine the goal?

I appreciate I may be missing something (wouldn’t be for the first time) — but rather than making installations faster, forcing users to sift through a list of software to check boxes or hit buttons to install the software they want — yay choice — sounds like it’s going to slow things down…

And that assumes, as has been pointed out, that users even know what apps they want!

I don’t think “app bundles” are a solution. They’re cute on paper — “creative super pack”, “social media power up”, etc - but what’s the criteria for what’s included? Who decides? And aren’t bundles of apps just repeating the problem the default install (supposedly) has given it effectively already comes with a bundle of apps (not every one of which people want)?

I mean it’s not like we collectively lack the ability to install the apps we want as things stand - Ubuntu comes with a package manager and software center.

A smaller ISO would be welcome but since the proposition here is that we all have super-fast connections allowing us download the software we want on demand… renders ISO size moot.

You say “clearly neither [current] option suits the varied needs of our users” — but I have to confess: it’s not clear to me!

Ubuntu should ship sane defaults that “just work” for the broadest set of people and, crucially, offer the best software open-source has to offer. I think the current curated install ticks all those boxes :slight_smile:


Here’s the thing. 1 week after you build that bloated cd, everyone who uses it to install a new system will first need to download 2 gigs of outdated data on the iso, and then re-download the fresh/updated packages. It’s infuriating for those of us on a metered connection or spoty internet.

I much rather prefer a slim 650 meg (or as close to possible) “starter” cd. It gives you an installer which installs a basic plasma/xfce/gnome desktop depending on the variant you chose, and then from that desktop you get a basic set of tools installed like file manager, terminal, browser, app store, and anything needed to bring you up to date by downloading the freshest packages.

Software suites and media players, I can live without all that because I’m just going to redownload it all again so why bother?


By the way, another old trope is “desktop productivity is useless as everyone uses the cloud.” Or that “archive managers are pointless; let’s remove them from the default install!”


Hi all many thanks for the feedback :pray:

Data points to share (sourced from opt-in data via Ubuntu Report):

  • 41% of users click through the installer without changing any options
  • 17% choose minimal install
  • LibreOffice adds 1GB to the ISO

A few notes from my perspective:

  • My feeling is that the default install should be slim. I’m avoiding “minimal” because there are packages in minimal today that we may want to remove (ie gamemode) and similarly there are packages that one could argue should be included as it completes the oob experience (ie gnome-clocks).

  • There are users who don’t interact with the installer at all (for example those who purchase pre-installed Ubuntu laptops).

  • There’s no plan to have users select apps during the installation process. In fact I’m recommending the opposite – lets reduce complexity with reasonable defaults and rely on better app discovery once the user is logged into their desktop.

  • If someone tries to open and no installed app supports that type, then we should recommend ones that do. I don’t think this is something the installer should try to solve.

  • From this discussion and others I’ve had, I see a range of opinions on the matter. Ultimately, we want to enable personalised desktops and I don’t think a great solution can be found in the installer alone. I’d like to see a solid base and then make personalisation straightforward. In time, this also means tooling to allow for backup/recovery of your personalised setup … but now I’m speculating way ahead.

Since I have a captive audience :joy: I’d like to focus this discussion and propose a strawman user-facing applications list for this slim install:

  • gnome-terminal
  • gnome-control-center
  • gnome-text-editor
  • gnome-font-viewer
  • gnome-disk-utility
  • gnome-calculator
  • gnome-characters
  • gnome-system-monitor
  • gnome-power-manager
  • gnome-logs
  • gnome-clocks
  • gnome-weather
  • eog
  • file-roller
  • seahorse
  • nautilus
  • firefox
  • app store



Yes to all plus the gnome-calendar, gnome-contacts, Thunderbird and Backups.
From my experience, even most of the time a web version or a PWA app does the job, a native app sometimes needed for advanced, or complicated tasks, e.g., an email client.
Additionally, a backup app is a must or at least should be recommended.
As for the music and video player, I agree with your choice to get rid of them since cloud services are used almost exclusively for basic needs, e.g., spotify, youtube, etc.

1 Like

I think backup is a good one to call out :pray:

@kakosf do you consider gnome-calendar/contacts as offering similar features to Thurderbird? And do you think those who opt for for web-based/PWA apps for those functions would also want these installed in the default configuration?

as a long time evolution user (which simply still has the only proper calendar integration with gnome) i admire the omission of thunderbird and gnome-calendar :slight_smile: and all the hassle it brings along having to uninstall these when wanting a lean install …


No audio/video file player? That seems like a feature even a basic OS should cover. I guess Firefox will play many of these files. (I say this even as someone who would uninstall this if it’s totem or vlc or something and replace it with my preferred celluloid and quodlibet.)

I’ve been proposing that we drop the email app from the default Ubuntu Desktop install for years.

I use webmail and I think most individuals do (even in office and school environments, G Suite and Microsoft 365 are very widely used).

Those that do use email apps don’t all use the same one. Thunderbird has a big fanbase which has enabled it to remain Ubuntu’s default for 12 years. Evolution has better integration into GNOME. Other people prefer Geary or Mailspring.

In summary, I feel like it’s better for most people to not have an email app pre-installed. Those that want one should be able to easily install the one they want. Both Thunderbird and Evolution are fine recommendations for people who know they want an app but don’t know enough to already have one they prefer.


I’d certainly hope that one wouldn’t have to load the web browser to play local audio and video. It’s NICE that it can, especially for niche scenarios, but Totem and Rhythmbox are applications that I use heavily and I agree that this is basic functionality in a modern everyday desktop OS.

This may not be popular sentiment (then again maybe more than I think) but I appreciate that LibreOffice and Thunderbird are present in the ISO. Thunderbird 115 will soon be overhauling the GUI and looks like a great PIM.


At the moment, gnome-calendar/contacts offer similar, but simpler features than Thunderbird and often buggy.
Thunderbird for me is a must because I’ve to handle several email accounts, domains, etc. If it were for example for only one email account, I’d safely say that the email client wouldn’t have been necessary.

As for the web-based apps/PWAs, most of them are concerning communicating and productivity applications, so definitely not needed to be preinstalled since it’s a matter of preference. Some examples are the MS Office and the MS Teams (lately MS stopped its official native client for Linux). Add Skype, which still remains as a snap.

Strictly speaking, a browser would be enough for most of the basic to medium tasks, assuming that the machine is always connected. Necessity for extra tools comes when we talk about specific and advanced tasks.


I think there should be a minimal (we no longer have a mini.iso, do we?) and a normal ISO. This would give choice to the internet limited and to the resource limited (and I’m not only talking about disk space or HDD vs SSD vs …, I still have a 1 GB flash drive).

If the user picks minimal, he knows what he is doing and is willing to install his stuff. Strip it t the minimum desktop and let him roll with it. I wouldn’t even include a browser.

If he picks normal, then it’s better to have as much as usual, including things like Libreoffice, VLC etc… I don’t think we need to be concerned about bloat because, after all, it’s Ubuntu, not Debian, not Gentoo — to be clear, distributions I also love —, and things should work out of the box.

Now, choice is always good, in my ideal world installers would have a collapsed tree view and the user could tick/untick stuff, and expand it to get better fine tuning, e.g.

- [x] Text processing
  - [x] Libreoffice
  - [ ] Latex
- [X] Text editor
  - [X] Gedit
  - [ ] Vim
  - [ ] Emacs
  - [ ] Vscode
- [X] Media player
  - [ ] VLC
  - [ ] MPV
  - [X] Gnome

but I this is probably out of the scope of the question.


@tim-hm if we really go down the route of relying on the app store this heavily, can we pretty please make sure we have a well maintained and regularly updated “recommend apps” and “featured apps” section again…

this has been completely stale since our beloved @popey left and i think having that well maintained (and in my view that includes that people managing the category actually install and test the apps they recommend in such a section) is a must if we will put that much focus to the store app…


What I personally didn’t fully understand is why there can’t be two options anymore. One with a preset (whatever that might be) and the other one with only the browser and terminal. Is it really such a big maintenance work? (yes, I am clueless)



I programmatically scanned all the recommended categories a few weeks back. They were either untouched (same apps on show) for a year, which isn’t great for freshness, or they contained applications which haven’t been touched for between 1 and 3 years. Some were even abandoned, marked private or had “DO NOT USE” on the store pages :frowning:

I provided feedback to the team, and I understand there may be a cleanup at some point. But the fact is, the momentum on snaps has dwindled a bit. If a fully snap desktop is to be successful, it needs fast, fresh, well maintained, secure applications published in a timely manner.

Also, thanks @ogra


@elcste @xa-hydra good point ref media files will amend (probably totem as its currently in full)

I’m personally excited by the work being done on Thunderbird, but I feel like there is no consensus around an email client/calendar (or just using the browser). So my feel is to keep it out for now.

@nteodosio I agree ref distinct ISOs for different use cases and I think that aligns with my proposal here. One set of packages for the default ISO, but there’s nothing stopping us/anyone from providing other ISOs with a more opinionated list of packages. We do need to improve the tooling to better support this though. But we can’t fix all of the things in one go :slight_smile:

@popey and @ogra +1 to the app store comments. More to share here in due course, but in order for this to all work well these are critical (we need recommendations, ratings, whats new recently updated, etc).


I think most of us on the technical side (who have opinions about which applications to use) would be in favor. Many of us customize.

However, I see two use cases who may be impacted negatively, and some design work outside of the installer seems to be in order:

  • Non-technical users, the historic target of the Desktop installer. They are quiet, but they greatly outnumber us when something goes wrong. “Oh, I tried Ubuntu, but it couldn’t open Jeff’s wedding invitation. It just didn’t work.” We can patiently explain that wasn’t an error message, it was an offer to install LibreOffice. But they won’t care: They expected X, they got Y, Y was too confusing, they threw it away.

  • Offline users. We’re seeing a lot of these over in AskUbuntu. Ubuntu seems popular in, for example, air-gapped secure networks. Desktop is popular because it’s complete for their workflow.

Of course, there are ways to mitigate and overcome both of these, and I’m sure you’ve already thought of all that. The point is merely that the path involves non-installer design goals (like UX: “Hey, just a moment while I install Gnome Foo to handle that”) which should be included as milestones so we don’t leave use cases behind.


I agree.

For specific mime types, do you think it would be a good experience if the OS said: “Hey you tried to wedding-invitation.doc but you don’t have an app installed to handle docs. Would you like to see applications in the store that support *.doc?”

For systems with problematic connectivity, see my response to @nteodosio in my post above. I think we can do better than the Full install for those folks.

1 Like