Ubuntu 18.04: Unity is gone, Gnome is back, and Ubuntu has never been better
arstechnica.com
arstechnica.com
I was curious about this and went spelunking for more information.
https://community.ubuntu.com/t/files-nautilus-v3-28-will-lose-the-desktop-icons-capability/3115/1
https://gitlab.gnome.org/GNOME/nautilus/merge_requests/46
https://irclogs.ubuntu.com/2018/01/09/%23ubuntu-desktop.html#t14:46
https://csorianognome.wordpress.com/2017/12/21/nautilus-desktop-plans/
I understand the need to remove cruft that blocks progress, but the Desktop filesystem metaphor is used by almost every single computer user I have ever seen. It seems wilfully destructive to remove something that is depended on by so many without making sure there's a replacement in place first.Really? I've never heard anything but non-intuitive ramblings from UI design people. None of them seem to ever actually use computers. Rather, they dream up use cases based on contrived scenarios. I want one thing: The icon is where it was yesterday. Gnome, unity all the others have one job: Let me launch the program I need and get the F out of my way. UI stability is why I years ago abandoned the chaos of Ubuntu for the stability of Mint.
Windows XP got it right. Start button->office->word. Start button -> Email. Most everything after that is fluff.
(Oh wait, but actually companies make money by making software into experience, so I guess that's where the overall trend comes from.)
I'd say Windows 95 in this context. XP was the first stable version of Windows (yes, the first one deserved to be called Windows 1.0), but the Start menu was working well in Windows 95/98/ME, too.
Isn’t this, quite literally, what Gnome Shell Extensions are?
-using the website to install extensions requires both a browser extension and installing chrome-gnome-shell, not installed by default by most distros
-UI to configure extensions is in Tweak Tool, again not installed by default in most distros and full of the things GNOME people don't really want you to touch but will begrudgingly allow
-the second most popular extension is an improved UI for managing extensions
-the API is poorly documented and new releases always break tons of extensions
That is a philosophy DEs should have, but (if you'd like suggestions) which XFCE most cleanly follows.
Other examples of missing stuff: - Complex shortcuts: you can't set keyA and also keyA+keyB as shortcuts work on keyPress rather than keyRelease. So no SuperKey and also SuperKey+L/D/R and such. - Easy window resize: the dragging thing is exactly 1 pixel wide. Gotta use Alt+Right-click to resize.
But you get used to XFCE's tiny quirks, it really gets like heaven after some work.
Incidentally, I'm part of the silent majority that enjoys Gnome and loves the emphasis on clean simplicity. No, it's not perfect, but it's moving in the right direction.
I'd forgotten this! Very frustrating when you're used to using this to switch to a file in that directory. I also missed split-screen functionality, and the inability to right click and fire up a terminal from that location. Oh, and also looks like they've dropped being able to double-click on an executable .sh file to run it.
Luckily none of this bothers me as there is Caja.
> I'd forgotten this! Very frustrating when you're used to using this to switch to a file in that directory.
I agree, but FWIW you can press Ctrl-L to focus the address bar, which allows you to enter a relative path, e.g. name of a file/directory in the current directory.
(or just use this overlay if you run Gentoo)
I'll try to keep it up to date.
(or the Nemo file browser, which also retained split-window functionality)
It's such a weird decision, as searching recursively something is a rare task, so having a dedicated button would be fine. While going to a file/folder in the current dir is pretty much the basic interaction with a file manager.
I don't get their reasoning there.
If somebody knows of any trick to get the old behavior back (even patching the damn nautilus), I'll take it. I'll probably upgrade in a few month and I'll need it badly.
They are removing the ability to have a Nautilus backed set of Desktop icons, with the intention of replacing this with something built in a more standard way for the rest of Gnome (which I think means web technologies?). As I understand it this not about removing desktop icons, it is about replacing unmaintained legacy code with more a modern alternative. I don't think the alternative is ready yet, but that's not to say that the direction they want to go is not having the feature.
In both OS X and Windows your desktop is just a folder with a background set to icon grid view. From right clicking to dragging, the UX is and should be identical to the file manager. So it makes sense to have that part of the UI handled by the program that's already a file browser.
Now a developer is going to recreate all of the features provided by Nautilus in a JavaScript and CSS extension. Any differences will make a janky experience for the end user.
Same thing with notification centers: another thing that both Mac and Windows added assuming it would be just as vital on the desktop as it was on phones. AFAICT they see very little use; I can't remember the last time I opened the notification center on either the Macs or PCs I use.
Not that new developments on the desktop aren't welcome, but widgets aren't it. Not every innovation that works on mobile can or should be brought to the desktop.
Certainly I use the PC one enough that I have the keyboard shortcut memorized and use it a bunch (Win+A). Interestingly, I find I often need it most in games, as the Windows version of the Xbox Guide, the Game Guide Bar, (even with the improvements in the recent April release), doesn't make it as easy to interact with things like Party Chat and Game Invite notifications as the Xbox Guide does on the Xbox One (where the Guide doubles as the Notification Center).
When I worked at a Linux distro (endlessos.com) they had modified Gnome to have a universal search bar on the desktop. It was targeted at a different group of users than me, but it seemed to make sense based on that group.
Hopefully Gnome does sufficient user testing to deliver a useful experience to some significant group.
Note that Active Desktop and ActiveX were not really related beyond being MS-powered internet technologies.
Honestly, I never use the desktop for anything so getting rid of it isn't a big deal to me. Could you explain what kind of things you use it for? What's so bad about viewing the desktop through a window?
So the desktop is a transition area, easily accessible, where I can put all the file I need for the task at hand right now. It's also the place I see everyday when logging it, so I can review what needs to be filed, and what need to be deleted. Clean desktop every morning, clean mind.
And it would fit their design philosophy to permanently remove it, which has been moving away from the concept of files as a whole. Instead, you're supposed to launch the app that does what you're trying to do and then open the file that you want to edit through that.
Another feature removal that has shown this direction, is the removal of the option to create an empty file from the context menu in the file manager.
Again, instead open the app that creates the type of file that you want and then create the new file in there.
Smartphones, tablets, and Chromebooks are incredibly popular and lack a desktop. ChromeOS is the most similar to GNOME here, in that it has a desktop wallpaper but you cannot create anything there.
Are you suggesting that Ubuntu should follow smartphones, tablets and Chromebooks in not providing the user a way to access the local file system?
Either the setup is too limitating : trying to do anything multi task on a phone OS is excuciating, you can't have your 3D rendering software/IDE/veracrypt/... installed in chrome OS, you don't have enough hard drive space,
Or on the other end of the spectrum, I even got friends who are minimalistic, they have no windows decoration, all is keyboard based, with an arch and dvorak layout, doing everything in VIM. Well in the end, they are slower, because they have to keep so many things in their head to do the most basic thing. Like 70 keyboard shorcut (on 3 different keyboard in case they are not on their perfect seup) layout and the position/index of their windows and names of files to just edit + switch. You can literally see them think before every action that requires context switching, which is brain power that they don't use for fixing the problem. And by the way, they spend a lot of time taking care of that setup, despite the fact they advertise it as rock solid.
Could cut straight to the GNOME logo after that sentence. This is sort of what they do, yeah. And they do have their niche with that, which is big enterprise-y, centrally administered systems.
If the user doesn't need to administer things themselves, then not exposing ways of administration in the GUI means less ways for the user to hurt themselves. Or rather to do something not so smart and then need tech support to fix it. Can't accidently hide your panel, if you actually cannot hide your panel at all.
I have no problem with one of the many desktop environments specializing in that, but the Linux desktop is mainly controlled by companies who then develop and set GNOME as default on their distros, which are also used by people who do need to administer their PC.
Ubuntu and Fedora have GNOME as default. Debian, too, which doesn't really have an opinion, but it's what they've always used and it's stable. Arch and Gentoo don't have an opinion at all, for obvious reasons.
openSUSE is still holding against, but SUSE uses GNOME in SLED, too, and has been spreading a sort of "We like all desktop environments in *SUSE-land"-mentality, which makes me feel like they're wanting to swap out KDE here, too, at some point.
But yeah, if you haven't yet, do yourself and the Linux desktop a favor and give KDE a fair try. Or pretty much anything else than GNOME.
I'm not saying it's impossible for you to like GNOME better, but there's so many desktop environments out there that are developed from users for users, which are actually objectively better than GNOME in so many categories. Even the most lightweight of desktop environments like LXDE/LXQt often have more features than GNOME.
And KDE is what you get when you take this feature per resource usage ratio of these lightweight environments and then multiply it by the resources that a modern system can easily provide. If you ever wanted to have all of the features and options that a desktop OS could have, KDE comes far too close to that. It's the flagship experience for power users (not the type that only ever wants to see a terminal).
As far as I know, Gnome 2 is the last example of a Linux desktop environment that underwent professional usability testing (on regular people and power users) before shipping.
Personally, these days, I prefer minimalist WM’s and CLI administration tools for stuff like WiFi, but all the newer heavy DE’s I’ve tried have obvious (to me) usability regressions vs gnome 2.
It is funny how little usability testing is done on desktop environments these days (I also wonder if Microsoft bothered with Win 8, or if apple is with iOS these days cough podcasts cough).
Also I think the general idea that options == bad started in the Gnome 2 era.
I've been very, very impressed with the MATE desktop. It's actually GNOME 2, carefully upgraded to GTK 3, then with a sensible set of improvements. It's available on most distributions, but the lead developer is also the lead for Ubuntu MATE.
not sure what you mean by 'professional', but KDE is still there chugging along (and IIRC default by suse, which is basically the 'red hat' of europe..)
Unfortunately not, no. They use GNOME as well, with some customizations: https://www.suse.com/de-de/products/desktop/features/usabili...
As much as KDE is better than GNOME in almost every way (opinion: mine), this is the one use-case that GNOME is built for. For use in centrally administered environments where the user is not really supposed to touch anything. Which is the primary use-case for SUSE Linux Enterprise Desktop (or other enterprise desktop distros, too).
openSUSE is still on KDE and I'm really hoping that SUSE doesn't violently take that over to make it their GNOME vessel. openSUSE's primary use-case is for home-users, who do need to administer their PC, and it's sort of a power user distro anyways (YaST, btrfs+Snapper, not Ubuntu), so KDE is perfect for that.
Does Gnome really have a niche there? I know of three places with large Linux deployments (as in hundreds of stations). None of them use Gnome -- its interaction model is so weird and foreign that getting people who aren't Linux/FOSS enthusiasts to use it difficult, painful and frustrating for everyone, users included.
All three are SuSE shops, actually. All three run KDE; one of them is more relaxed when it comes to software engineers so a couple of folks run Cinnamon, but that's about it.
The unsettling phenomenon is that most non-techie users don't need something super smart and intuitive, they want something like Windows, because that's what they've used so far to get their work done and they see no reason to learn something new.
If it's different, but approachable, they'll adjust (perhaps that's the value in customizable DEs like KDE; you can start with something that everyone's comfortable with, then they can tweak it to make it "better"). But if it's radically different and you need someone to explain them exactly why it's better and how they've been using computers wrong this whole time, that doesn't work in the real world.
So for most people it would be a "Windows 7"-esque GUI.
• There are the kind that want more "power" exposed through the GUI, with administration happening through the same interface as regular usage. This culminates in things like Windows Server administration via snap-ins. This is traditionally what a "power user" refers to.
• There are the kind that want to treat their PC or mobile device as, effectively, single-node enterprise systems, with administration happening through alternative interfaces that were designed specifically for administration, usually of large clusters of systems. These are the folks who are at home with Windows Group Policy Objects; Linux config files; MDM profiles on mobile; etc. Basically, sysadmins who happen to be administering their home network. These also get called "power users", though maybe "hobbyists" (that hobby being https://reddit.com/r/homelab/ building) might be more apropos.
Most modern OSes seem to be moving away from supporting the first kind of "power user", seemingly because features that support that use-case are also "sharp edges" where less-knowledgeable users can hurt themselves.
But modern OSes are exposing ever-more configuration functionality to the second kind of "power user." Windows now lets you do more things through PowerShell commands than you ever could through the Control Panel. macOS has always left the more arcane functionality to the terminal, creating POSIX-ish tools like diskutil(8), defaults(8), scutil(8), lsregister(8), etc. And Linux, underneath, is still just glued together 99% via editable plaintext files (the newest iteration being systemd unit files.)
As "the second kind of power user", I like using GNOME just fine. (It's actually a little bit "more powerful" for me than macOS is, even, because I'm also the third kind of power user: a developer willing to tweak the GNOME source to get what I want. ;)
Seems the answer is in that last link of yours:
>Nautilus had have a feature called “the desktop” which adds icons on the background of the user workspace, similar to Windows.[..] if you look you will see that the desktop is composed by more than 10.000 lines of code
That's a big mess and I finally understand why opening nautilus on i3wm always completely messes the desktop up. Using the file manager as a hack to build a desktop is a bad abstraction and I can understand why it they'd want to rebuild it.
Also, Gnome on this laptop (Dell Inspiron 15 7000) gave me (and several others users, as per forum post https://ubuntuforums.org/showthread.php?t=2358975 and some bug reports I found on the internet) a freeze or stutter every 10 or so seconds. So instead of having to start debugging shit from a fresh Linux install like I used to do in early 199x linux days, I chose to install xubuntu and just get to work and be productive.
Try XFCE today, really (and install all gnome apps should you need to run something from that desktop environment: they integrate great with xfce).
o/
However having access to things like a network manager to join wireless networks, a battery indicator, a volume control, etc is essential in today's world, which is why I now run a heavyweight xfce environment.
If they did a simple non tiling WM that had those little things in it and wasn't a piece of crap, I'd be there like a shot.
It's Openbox (which is a really capable, lightweight non-tiling window manager) plus a panel plus Xfce Power Manager and similar applications to handle the desktop environment stuff.
I use xmonad with stalonetray [1] (which as you may guess from the name is a standalone notification tray) and dzen2 [2] as a notification area for basic information ( battery / cpu / temperature / wifi status ). With them I cover enough of my needs and I don't feel the need to anything more complex.
I think that it would be nice to have some middle ground between a full desktop environment and bare-bone and manually configured window manager though.
[1] http://stalonetray.sourceforge.net/ [2] https://github.com/robm/dzen
You don't need to run a full DE to have tray icons. I'm running i3 and the network manager applet + other tray items work just fine.
Nautilus has got increasingly shitty. So there's less Gnome now in my Xfce.
I tried Win 10 recently, it was horrible.
In conclusion they all have warts. But Xfce is okay in a Win95 kind of way. It could be better.
Simple UI seems to always elude distros/desktops.
Note: I'm currently not using Xfce at home, but I use some older version at work.
Ubuntu Mate uses gtk 3 and probably in general more modern technology than XFCE, with a similar style. Definitely worth considering.
I even brought an old Macbook back to life by throwing Mate on it in place of MacOs.
The applications delivery the value on the desktop, not the desktop environment. The DE vendors (Gnome/Unity in particular) need to work that out and just get out of the way.
I used Xubuntu for a while but I was always fascinated by how bad a couple aspects of XFCE are. They maintain documentation [1] on oddball ways of resizing windows instead of just giving people what they want. I suspect MATE just does what most people want out of the box.
[1] https://xubuntu.org/news/window-resizing-in-xubuntu-and-xfce...
I've used this when I need to use Windows. Works well enough.
Win+R -> Resize
Win+M -> Move
Win+A -> All desktops
Win+Q -> Stay on top
Win+H -> Hide
Win+Arrow -> change workspace
Win+Shift+Arrow -> Carry focused window to workspace
etc.This works so well for me that I completely removed all the buttons from the window manager since I no longer need to access them anymore.
'm' followed by arrow keys to move the window around, or
'r' followed by arrow keys to resize the window
For people like me on laptops with trackpads, these keyboard shortcuts work really well.XFCE avoids the plague which affects rest of the software industry (me included) - dedicate design, marketing, and development teams who constantly have to add something to justify their jobs and bonuses. Look at the number of app updates on mobile phones, every freaking week there is a update available. If you don't update they will eventually block older versions.
Argh. I've seen this problem over and over on a variety of Ubuntu versions over the years, as if they don't want me resizing windows.
Is there any way to tweak this? I'm not at all liking it. My 16.04 installations are much better in this regard.
The width of the grab area is controlled by the theme." 1
Most people appear to circumvent the problem by using another means of resizing windows.
"As soon as your keyboard-shortcut (in Xfce upstream its Alt+F8) is hit, you can then resize the window by either moving your mouse (no need to click or drag) or use your keyboard arrow keys." 2
1 https://unix.stackexchange.com/questions/156435/how-can-i-ma...
2 https://xubuntu.org/news/window-resizing-in-xubuntu-and-xfce...
I understand that there are times when that isn't work so it's nice to have both options. ( I wrote a browser based windowing system and have experimented with having the bezels on the windows be quite large and/or auto-expanding when the mouse nears them to make resizes as easy as possible since hard to hit pixel thing annoys me so much)
I know it's not a fix, and it is still cumbersome if all you have at the moment is a laptop trackpad. But it beats having to update your destkop environment's chrome and all that jazz. I find XFCE's window chrome pleasing and minimal, the only issue is exactly that, you only have 1px of hot area at the borders.
The only thing I don't really like in the Ubuntu's model is the bi-yearly updates and the PPA system. Every time you upgrade the distro the PPA's get disabled and there's a lot of work enabling them and fixing the packages.
That being said, I installed NixOS to my other laptop and after two days got a configuration that just installs and configures everything I need. I might overwrite this Ubuntu installation with it and go rolling with the master.
I've been using i3 for years, and I tried out gaps when it first started popping up in the ricing forums (like https://www.reddit.com/r/unixporn/ [please note, it's not actual porn]).
Do you use gaps so that you can see little bits of your background image? Is it purely aesthetic? I can understand beefing up the borders so that there's some space between tiles, but the gaps seem like unnecessary waste of space.
I need Win10 for low latency music work and for the odd game, all development work happens in Xubuntu VMs.
Windows is essentially a shell where I juggle vms when I need a dev environment for a specific client.
I use VirtualBox as my virtualization solution and what I found was that using Xubuntu made the whole "linux in a vm" experience much more snappier, and since I basically work with 'text' and don't fiddle with music+video inside the vm, I often forget I'm inside a VM when it's fullscreened. With Ubuntu/Unity the whole VM experience gets slowed to a crawl because Unity wants to XDamage/Composite everything and virtualbox's "accelerated drivers" aren't that good to make it perform good.
Back on Windows 10 I re-evaluated VirtualBox, Hyper-V with the new linux enhanced session support(poor experience TBH), and VMware workstation player. VMware ended up being the best experience in terms of responsiveness IMHO. You may want to give it a go. It's free for personal use, but unfortunately workstation pro is quite expensive.
I find it smoother, much more intuitive, and about the same with regards to memory footprint
https://windowmaker.org/ -- Really, though, it's in the package repos of just about any distro...
Strongly agree. KDE/plasma continue to make great progress and it's my personal favorite. Looks gorgeous out of the box, is light on memory and the cpu, and offers great customization options. Since it uses a layout similar to Windows 7/10 people coming from the mentioned operating systems feel right at home and it eases the transition to linux (I've installed Kubuntu for a few non-technical people and so far no complaints!). If you're interested in following KDE's progress, check out Nate Graham's blog [0] where he publishes a weekly post going over new features and bug fixes in KDE applications and plasma. Really looking forward to plasma 5.13.
They still do some kookie stuff with Plasma -- from big things like Activities to little annoying things like removing the drag-resize from the application menu (I fix the size in ~/.config after every update), and removing the KDE version info from the Dolphin file manager about menu. Weird little changes that make you wonder if someone's gone insane.
Mind you I disable just about every automation and haven't used KDE apps like kmail for a long time.
Also, Alt+PPM works well for resizing the Plasma menus now. It was missing for some time due to architectural changes.
Definitely wasn't a usable way for putting icons on the desktop when plasma rolled out.
There was, in a form of the icon plasmoid. It worked and behaved completely differently than Folder View does, but if you really wanted, you could have some icons on the desktop. It still works, by the way - it gets created when you drag'n'drop something into the non-Folder View based desktop.
I think drag'n'dropping might have been added shortly after 4.0, but the icon widget was there from the beginning for sure.
I've tried Gnome and was very disappointed. No global menu integration which means losing useful screen space for no good reason. Tray icons somehow look squished and ragged. No fractional scaling is a massive issue for me since I'm working with dual 27 inch 4K displays. Tray bar only shows only on one of the monitors and you need a plugin to make it appear on both which is just ridiculous. No more convenient built in features/options like auto resizing a window to 1/4 of the screen by dragging it with a mouse to one of the corners, etc.
On the other hand you can just install KDE Plasma that went a long way and works fine if you just stick to the UI - all your pain points should work out of the box or can be configured using systemsettings. KDE applications are a mixed bag - I love dolphin but kmail and other stuff feels strange. Also printing is way better in GNOME apps, not sure why exactly.
I kinda tolerate gnome 3 (despite its marvellously obvious flaws) for now and if I find it way too annoying, I'll rip it out and use fluxbox or smg similarly low level wm.
Meaning that it is a fig leaf from the Gnome devs to power users.
I'm on a laptop most of the time and, like you, have absolutely no desire to go back to wasteful non-global menus. Have you eyed a replacement DE for Unity/Ubuntu going forward?
Edit: OTOH, Ubuntu have clearly spent a good amount of effort to customize Gnome so here's hoping they can morph it to work like Unity even more (OHD, global menu)
Global Menus is unfortunately something that I still miss, although I was able to recover some screen space with an Extension called Pixel Saver, which merges the window control bar with the panel by removing the window decorations and adding the buttons to the panel. Not ideal, but useful.
So, when you have your hand on the keyboard, this can be much more convenient.
The Mutiny layout in UbuntuMATE 18.04 is really nice, and they've implemented the HUD, global menu, titlebar integration, and it's got HiDPI support now.
I think multimonitors is still an issue though - you can only have one dock, and it's best left on the screen on the left by the looks of it.
I've got the exact same 1440p screen as you on my Thinkpad (which wants to be run at 1.5x or 1.75x) and the best solution I've found is provided by KDE. Try installing Plasma (either by installing Kubuntu or just installing the desktop w/ apt) and setting Display->Scaling to 2X and then go into fonts separately and set "force font DPI" to 168. That's working remarkably well for both GTK and QT apps for me. 2X scaling for windows/icons keeps them from being tiny/blurry and setting the font size to 168 (aka 175%) prevents the fonts from being huge.
Moving windows worked within workspaces (click, drag, drop) same as across workspaces (click, drag, drop).
Basically the principal of least surprise. Moving the mouse and clicking shouldn't have special rules across workspaces unrelated to the rules of working within a workspace.
So basically you could basically intuitively work on a space much larger than your screen without having to think about it.
Ubuntu 18.04 forces vertical workspaces (so 1x4). In the dual monitor config it only moves the left monitor across desktops. And moving windows between workspaces is a nightmare. Step one is move the mouse to the top left corner and click activities to expose. Step 2 displays all your windows as tiny icons, 2/3rd of your screen is just the background, so they aren't much bigger then icons, and theor position once exposed is unrelated to the original position. So now you have to refind your window, then drop it into one of the workspaces. But you can't watch or tell where it lands. So then you have to switch workspaces and refind that window.
Quite the pain compared to 16.04 (or any decent window manager since olvwm 20+ years ago) it's just drag and drop.
Additionally there's no spatial relationship. So a window that's off your top edge of your workspace doesn't show up in the workspace above yours. It's basically just 4 unrelated workspaces. They also prevent wrap around, so you can't go from 1 to 4, but have to visit 2 and 3 first.
Reminds me of the first hacks to OSX and Windows to get more than one desktop. Linux folks figured this out 20+ years ago with olvwm, ctwm, and many others.
[1] https://extensions.gnome.org/extension/484/workspace-grid/
Am I the only one who finds it a bit frustrating that any time someone complains about missing/bad functionality in gnome, someone recommends to install an extension, but when a user complains about stability or performance, their number of extensions are blamed?
I mostly use i3 which is perfect, but one of those days when I feel like using a full desktop environment, I just can't. On positive side however, scaling for hidpi screen seems to work out of box in Ubuntu 18.04.
There is a fix for this in display settings.
I have personally gotten used to the vertical alignment and new key patterns. Still using 17.10 but performance seems improved from unity and looking to maybe get back into a bit of gtk development ;)
When you move a window, it "lands" exactly where it started, moving a window between virtual desktops doesn't change it's X/Y position (why would it? what do you want to happen?).
Basically, you're having trouble learning your new tools. But different and bad are not the same thing. Spend a little longer and try harder.
Everyone hates new desktops when they try them the first time. "Everyone" hated Unity too. And Gnome 2. And Gnome 1. And KDE4...
I wonder if it was an effect of people who didn't like it complained and threw hissy fits. Those that liked or didn't mind, didn't say anything.
So it _seemed_ as if everyone hated and they just had to replace it. I've observed that pattern happen with product features before. A customer chimes vocally that they hate something, another does etc. An issue is created, work starts, then a whole new wave of complain come later that "hey I did like that feature, why did you remove it".
Wonder how the decision was made, did they do a more formal survey or just went by the vocal critics on HN and other sites?
(Now having said that, I don't dislike the new interface, so not complaining about GNOME, just mainly interested in how the decision process worked).
I believe it was discontinued due to Canonical's strategy of dropping mobile support. Unity was an effort to have a unified desktop across devices, I believe.
Canonical is pivoting towards enterprise software [1]
1: https://www.omgubuntu.co.uk/2017/10/why-did-ubuntu-drop-unit...
The Ubuntu team have done a great job, the "new" settings screens are especially nice IMO.
I think what really gets me is I use Windows at work and KDE is too close to windows in some respects but isn't...it's also too close to other Linux DE's but isn't. Feels really awkward, and isn't very pretty. I don't necessarily agree with the design decisions of GNOME Shell and GTK but I think they still look a lot better than KDE and Qt.
It's not.
Also, I can't stand the ugliness of GNOME, while Plasma looks mostly neat and clean.
Luckily, the search has gotten a lot better, so it will actually find KCalc when you type in "calculator".
You can also type "1+1" into the search bar and it will tell you the result right there.
And there's also an option to have the menu show applications with their generic names. So, for example "File Manager" instead of "Dolphin". Depending on the distro, this option may be default-on, too.
And now would be a good time to try it again, by the way. The last three years were spent with a major rewrite. Now it has most of the features that you'd expect from a KDE again, is stable and has been stable for long enough that most distros actually have a stable version in their repos (even Kubuntu is decent for once.)
Since I mostly work in the terminal, it improves my productivity quite a bit when I can navigate around and open apps without touching the mouse.
I'm thinking of trying out EXWM, but I'm not sure if it's stable enough yet.
When I start X11, dwm and a xterm my ram consumption is only 60MB.
Sure it was a pain to set it up (and having to configure stuff like i3-lock and suspend) but I did it once years ago and have not changed anything since then.
I feel very productive with it and I use the mouse way less than I used to. I liked it so much that at home I also do development on Arch Linux (also in a VM) and also use i3. My i3 config is in private git repo. Nothing fancy like I see on reddit.com/r/unixporn with i3 gaps. I basically copied almost all the config from https://github.com/jessfraz/dotfiles
I don't really have to make any modifications to the config either. I change the fonts to my liking and swap dmenu for rofi and I'm more or less good to go.
I never understood the fixation about ease of installing, as that's something I either did once every few years on custom hardware, or had automated already using a CFM. For stuff I installed manually (laptops, occasional desktop rebuilds) partitioning was always the thing that gave me the most pause - and a graphical installer vs text installer didn't really impact that process. Picking a new hostname wasn't any easier with a GUI than a text interface either.
The common claim was that Ubuntu had have better X, init (upstart), installer, and maybe some more recent exotic drivers -- but I was hard-pressed to identify the differentiating factors from Debian testing (what most of my friends & family were on, at my behest).
The only time I use linux is when I need a linux VM for something. This happens maybe 2/3 times a year, and I can never remember all the random linux specific installation stuff. Ubuntu goes into enough detail that I don't need to google it, Debian doesn't.
My fixation is because I don't care about what version of linux I'm using, I'm not going to be using it for much, I just need a linux kernel and a reasonable package manager.
If you're re-installing from scratch each time you're possibly doing that wrong.
An installation from scratch, in any case, should only ask you hostname, domain name, maybe wifi details (probably not for a VM), partitioning preference, and package / function (set of packages) --- none of this stuff benefits from a flashy / GUI installer.
> ... I just need a linux kernel and a reasonable package manager.
That last bit - that's kind of my point.
That's when I spent some time doing an ubuntu build, which even as of 16.04 is ncurses based installer. Boot from pxe, give an IP address/gateway/dns/hostname etc, approve the automatic partitioning selection, select which support area looks after the machine (for user generation, monitoring, etc), and that's it.
Everything gets the same base install, including things like active directory based logons (with ssh keys for those who want them), automatically adds itself to nagios. Build a dev installation (which you select form pxelinux) and auto patching gets enabled, and syslogs get fired off the the right location.
On a VM this can of course be passed as part of the creation process, so it just take 5 minutes to build from scratch. When it comes to building on real hardware it's just a matter of typing the stuff into the boxes then going and making a cup of tea.
After the base install is on, you can then either apply an ansible script to do various package installs, or do it old-skool with apt-get install $package.
It's entirely possible that debian testing may work, however from a corporate signoff (and when you've got nearly 2,000 of these machines installed you need that) saying you're basing the company on 'Ubuntu LTS' is far better than basing it on "Debian Testing".
We've run this way for over a decade over 6 continents, it just works.
Yes, of course I am, as an amateur in linux it would be reckless for me reuse an existing linux installation in a different context. The only way I can safely re-use is docker.
>An installation from scratch, in any case, should only ask you hostname, domain name
Domain name has already lost me, I'm not a website, why do I need a domain? :)
In practice, you're usually asked about root, usernames, home folders, etc, too.
>none of this stuff benefits from a flashy / GUI installer
No it doesn't need a flashy installer, but I'm equally talking about the Ubuntu Server edition's text installer. I'm not sure if you're talking about GUI vs CLI or "graphical" vs "text-based oldschool GUI". I would say totally vanilla CLI installers absolutely don't explain things well.
If you're using Docker - or even know about it - you're not an amateur.
Since you're talking about your need to 'build a Linux VM for something' this also suggests you're significantly better than average on the GNU/Linux awareness scale.
But to your question -- if you're using a VM instance, make a base build, make it just right, and then snapshot it. When you need to spin up a new Linux VM for something, replicate the snapshot, do an apt-get update/upgrade (or equiv) and carry on. This is fairly common practice and saves having to remember your domain name, root password, home folders etc.
To your last point -- I rarely install Ubuntu, and can really only compare Debian vs various Microsoft installations processes for the past twenty years. GUI GNU/Linux installers were popularised by Ubuntu, and that's why that came up. You're right that totally vanilla CLI installers don't explain things well, but that's what we used to have on Windows 3.1 / 95 / 98 ... and people somehow managed to struggle on. Either average Microsoft users have become dumber (quite possible) or IT types continue to underestimate people's ability to read or work things out. Given almost everyone that installs an OS now has another device to hand to look up guides and answers, I'd suggest it's (installer complexity) even less of a distinction now.
Aside - Microsoft installers always had it much easier than most GNU/Linux installers simply because Microsoft installers (for a very, very long time) didn't give a flying if there was an existing operating system on the box. They'd just re-partition, re-format, install, and overwrite the MBR.
Pretty much every GNU/Linux installation process took great care to accommodate other OS's already in situ, and also documented the process and correct ordering required to be sociable with other OS's.
As someone who frequently hates on Desktop Linux, I agree with the sentiment that the fixation over ease of installing is pretty silly. Only a few distros in recent memory have anything remotely difficult about their installation process, and it's usually that they don't bother with any wizards and just make you do everything by hand from the shell, but those were meant for experts anyway.
One thing that gets me is ~12 years ago I used to install distros all the time. There was a semi standard way to partition your drive, / home and swap. Now I look at the partitions that are produced and a lot of the time I have no idea what is going on. And please stop moving the bloody usb drive mount location about.
Asking the Linux community to have consistency and standards is like asking rain not to be wet I'm afraid.
With Linux, I have to check what the three identical guided partition options mean, check what a network mirror is, work out which server packages I need, work out what GRUB is.
That's four pretty technical things. Honestly, I'm probably familiar enough to make a best guess these days, but I have no idea what the consequences of the choices are and google doesn't generally help as sometimes there is no right answer!
[0]https://www.howtogeek.com/wp-content/uploads/2013/09/install...
While it is nice that you want to learn, you do not have to know any of that. Boot, select disk, username+password, click next a few times, done.
And the (or create a local) part for windows is very non-obvious: https://superuser.com/questions/904673/can-i-install-windows...
Installing Linux used to be a huge problem, and that's why Ubuntu became popular. Other distros caught up, but it was really Ubuntu that pushed it.
For a long time, anyone interested in dipping their toe into GNU/Linux would either use a live CD (Knoppix was huge) or they'd be encouraged by someone that would assist them.
At the risk of sounding dismissive, I'd suggest if someone couldn't summon the patience to follow a fairly straightforward keyboard/location/pick-a-host-name/yeah-use-my-whole-disk sequence of steps, then actually using a GNU/Linux operating system for more than a few days was probably not on the cards. I'm not suggesting it was ever an intentional gating process, but it probably did help to save some people from being even more frustrated.
Given that each release of Ubuntu starts as a fresh fork from Debian's testing branch, this is not a compelling case for Ubuntu over Debian.
> If there were drivers for linux, they were likely tested against ubuntu.
As above, and it doesn't really matter if you meant drivers for Linux (the kernel), or GNU/Linux (a distro).
I've never heard of any drivers that work on Debian but not on Ubuntu, or vice versa.
> Ease of installation is rarely talking about the install screen; it's talking about everything just working without having to dig around the internet, and installation is a good proxy for that.
When people talk about the ease of installation, they are almost definitely talking about the ease of installation.
"Just works" is a useful phrase to describe a range of (other) user experiences, but again it's not clear how Ubuntu has (and arguably had) any significant advantages over Debian.
What is that? Pretty hard to search for.
Configuration Management. The f was often shown as lowercase, as Cf is short for Configuration (don't look for a middle word :) No one knows why.
I feel like the biggest improvement by far is the Snap package manager and its integration with the Ubuntu software store. After 10+ years of doing personal computing on Linux this is the first time that installing my main apps has ever felt sane!
I was blown away by being able to install the latest Android studio, with a launcher icon, in one click. No more tarballs and command line launchers and broken dependencies.
Has this been around for a while and I just didn't know about it?
I always felt the the offical Ubuntu package manager was a dumbed down version that(though its been a while sine I have used Ubuntu). Synaptic definitely felt better for installing libraries instead of applications.
>> broken dependencies.
Not that it's Synaptic's fault, but some packages will have conflicting dependencies, especially if you're using PPA's to get recent software versions. It's not uncommon to get "foo (latest) requires bar >=3.1 but you have bar 2.8 installed. Baz (installed) requires bar 2.8". Snaps avoid conflicts by having each app contain it's dependencies, which could otherwise conflict.
It also negates the entire point of having a package management system, as you can’t upgrade a security issue in libfoo and get it fixed once, but have to wait for all the users of libfoo, and the users of users of libfoo, and so on, to provide a new version of the Snap package with a new fixed libfoo contained inside it.
Yes, you don’t get that library update everywhere all at once, but this gives each vendor a chance to make sure that update actually works with their app.
Once an Ubuntu release has been cut the debs within it can (for the most part) only receive bugfixes, not new features. Which means that as an LTS goes through its lifespan, the software within it slowly gets further and further out of date. Snaps, on the other hand, are as up-to-date as their publishers make them.
I have been using it since then, but then again I am one of the developers of snapcraft.
Regardless of my affiliation I really appreciate having a base system that is relatively static and have my apps just roll, and if you don't want to roll too far, you can stick to a major release if the publisher of the snap provides a channel with a track you can subscribe to (instead of following the default track, `latest`). If things go ad, you can always revert, which is a big plus from my perspective as it is clean and per snap.
It may be time to try/support Debian GNU/Linux again.
$ systemd-resolve --status | head -4
Global
DNS Servers: 192.168.88.1
<ISP IP 1>
<ISP IP 2>
It's not using it though: $ nslookup laser
Server: 127.0.0.53
Address: 127.0.0.53#53
** server can't find laser: SERVFAIL
$ nslookup laser 192.168.88.1
Server: 192.168.88.1
Address: 192.168.88.1#53
Non-authoritative answer:
Name: laser
Address: 192.168.88.5
Not clear to me whether this is a problem with systemd or with ubuntu integration of systemd. Glancing at /var/log/syslog, it seems systemd-resolved is restarting every four minutes, which seems unnecessary. $ nslookup dk.
Server: 127.0.0.53
SERVFAIL
$ nslookup dk. 8.8.8.8
Server: 8.8.8.8
Non-authoritative answer:
Name: dk
Address: 193.163.102.580: https://github.com/systemd/systemd/issues/6224#issuecomment-...
Of course, setting the following in /etc/systemd/resolved.conf did no good:
LLMNR=no
...because apparently systemd-resolved just ignores its configuration. (I did restart the service; it made no difference.) I re-commented that setting since it does nothing. In order to get this working, I instead had to change the symlink /etc/resolv.conf which previously pointed to /run/resolvconf/resolv.conf to instead point to /run/systemd/resolve/resolv.conf, another dynamic file which is also maintained by systemd-resolved.So now I have systemd-resolved handling the info from the DHCP ack, but it isn't handling DNS. This works, but good grief.
With a bit of effort installed Ubuntu 18.04 on it. Works great! Wifi, touch pad, even the touch screen and orientation all work out of the box. No messy driver configuration.
I did replace gnome with Unity (an `apt install ubuntu-unity-desktop` away), because it performs tons and tons better than gnome and with a 1.25 scale makes the 11" 1080P screen much more usable. Amazing how much computing you can get for this price...!
Fedora 28 and Gnome were made for each other. They look and work brilliant. Try it out (and don't let dnf vs apt hold you back)
I've been struggling with VNC on 17.10, eventually ended up installing RDP (XRDP), also buggy as hell.
Can anyone point out a good remote desktop solution for Ubuntu, that will preferably work in 18.04 ?
From https://tutorials.ubuntu.com/tutorial/tutorial-upgrading-ubu...:
If no upgrade appears
As mentioned in the Ubuntu wiki, upgrades from 17.10 will not be enabled until a few days after the release of 18.04 and upgrades from 16.04 LTS will not be enabled until a few days after the release of 18.04.1, expected in late July 2018.
You can force the upgrade, however, with the following steps:
close Update Manager if it's still running
open a terminal in the same way you opened Update Manager
type update-manager -d and press enterIt's more transparent, and if you don't care, you can install pacaur that hides all of this behind a pacman interface
If anything else irks you, you're probably looking for a different desktop environment or just to replace one application.
Other than those sorta of things... you could say there's "not much difference" between almost any two distros. Especially since the Trojan horse systemd.
I personally don't trust canonical at all. I bet more would think like this if they read up on them a bit more.
I personally use this method, which is easy to use, upgrades automatically via the normal apt mechanism, and has never once failed me.
https://askubuntu.com/a/236458/498460
I'm on the `hwe-16.04-edge` kernels, and they've been great for me, although the `hwe-16.04` kernels are still kept up to date quite well.
luckily my colleague knew what was going on. i had special characters in my password and after the update, the keyboard layout changed from german to english.
I encrypted my external hard drive with LUKS and plugging it into old Linux installs just work :)
So I wouldn't panic before trying it with a usb stick..
It's like running peak 200x Ubuntu frontend with all the up to date backend.
Responses from ubuntu folks on that bug ticket have just been "Those packages are not official ubuntu packages so we can't do anything"
1: https://bugs.launchpad.net/ubuntu/+source/curl/+bug/1754294
Does anyone know more about this by any chance?
BT applet, at least on Unity (yes, I'm one of those people) seems to be borked in that it can't enable BT after resume. I'm guessing it's because it remembers the wrong BT device ID (which gets incremented after every sleep/resume, at least on my system).
It's really light on resources and works great! (been using it for around a year now on my personal devices).
THIS! This, so many times. Xfce lets you actually use your computer to accomplish tasks without getting in the way and/or forcing you any "intended usage".
It is so good and so light, it's light-years ahead of Gnome.
Gnome was a major reason for my switch. I didn't have time for unity (just like I gave up on KDE for the same reason).
Gnome is solid and looks beautiful these days.
Unity has brought a lot of innovations, and it would have been a lot better to keep it around.
Gnome devs started making extremely unpopular changes that are hard or impossible to revert. Unity at least put some pressure on them to consider their users a little bit. If only for this reason, Unity's existence was a good thing even if you didn't like or didnt' use Unity.
The situation to me is similar to Chrome's dominance, with Firefox being around to be used by few people, but having the very important role of not leaving everything to one company/group/project.
I've tried using Gnome 3 but there are just a few things that drive me (and lots of other people) crazy and can't be changed. Along with the project's "our way or the highway" attitude that made me switch to Xfce. If Ubuntu keeps being based on Gnome, Gnome might keep leaking some of their great ideas to Ubuntu, which is bad IMHO.
Xfce is a little too barebone but very usable, and with an extremely positive community. Devs are normal human beings that never look down on you (from my experience)--which can't be said of all open source projects.
Unity was just so clunky, and no matter the hardware always seemed slow.