I also sighed with relief. Generally this is what happens with Linux desktop through several technologies compete and one comes through after a while as the winner. Now that Wayland is actually shipping and with the whole community shortly behind it I hope to see some nice improvements in the Linux desktop in the next few years.
If someone could fix Network Manager that would be nice though...
Seriously. When they announced Unity, the only thing I could think of was "too many cooks in the kitchen." I have no trouble with choice, but too many options is devastatingly fragmenting. The investment should have gone with Gnome to begin with, as Ubuntu had already made the commitment to that project.
I've been running everything using supervisord lately -- then I only need to maintain a single system startup script. And I can use supervisord under OSX easily as well to run the same stack of stuff.
I use supervisord a lot as well, though I actually had issues getting it up and running on the latest Ubuntu. It seems that it isn't enabled immediately after installation via apt-get, but has to be manually enabled. Minor hiccup, but a small pain.
"Six years after making Unity the default user interface on Ubuntu desktops, Canonical is giving up on the project and will switch the default Ubuntu desktop back to GNOME next year."
Yes, yes yes! That's fantastic! I hate Unity and I love GNOME. Good choice. Too bad it's going to be April 2018 until then, but this should teach anyone: It's never too late to revert a stupid design decision!
I think part of it, is it was too fragmented of a codebase that migrating to Wayland would have been more effort than re-polishing Gnome 3 would be. There were few advantages since the non-desktop (phone, tablet) targets are also going away.
I think the stupid decision was to stick with in interface you hate instead of going for an alternative distro that uses GNOME for a desktop environment, but I'm glad it panned out for you in the end.
I think the stupid decision was to assume that I do something that I do not do and then respond to me based on that assumption.
First of all, I am a Windows user 95% of the time. But when I wanna spin up a VM or Linux distro real quick, Ubuntu is my choice. Of course, if I use it longer, I will install KDE or something like that. But with lots of window managers, I noticed instabilities or redundancies (e.g. there's a Unity and a GNOME service running for the same thing). It's best to use a distro with the window manager it comes with. Which was Unity on Ubuntu, but will finally be GNOME again.
Unfortunately there are plenty of applications designed for gtk, qt, wx, ... that you'll likely still have various supporting libraries and services installed.
> "Six years after making Unity the default user interface on Ubuntu desktops, Canonical is giving up on the project and will switch the default Ubuntu desktop back to GNOME next year."
Now imagine if 6 years ago Canonical had spent all that time and effort into improving existing DEs and implementing Wayland-support instead of their (already then) NIH initiatives.
Where would we all be now? Certainly not still stuck in this same, unimproved X11 sinkhole (yes yes, apart from Fedora-users, I know).
Such a shame they had to spend more than half a freakin decade to realize their mistake.
I agree, but at this point, I had given up hope. It's better to realise one's mistake late than to never realise it at all. So I'm excited they correct it after all.
I prefer it there.. if Windows start menu weren't weird on the left, I'd have it there too... my mac dock is on the left as well. I have more horizontal space than I typically use, while vertical is at a premium.
"Shuttleworth made it clear that the cloud isn't the financial future of the company. He wrote:
'The <latest-big-idea> story for Ubuntu is excellent and continues to improve.'"
I don't have anything against Ubuntu, but I do have something against chasing the latest fad. I know, the cloud isn't a fad, but mobile wasn't a fad either.
> Android is killing it and there is no viable open alternative.
Because of all the services Google provide on top of it. That's where the real value is, just like iOS, CyanogenMOD and co don't seem that popular. If Microsoft wants to make a come-back in the mobile space, that's where it needs to focus.
I never liked Unity and I'm comfortable with Gnome, so I think this change is good.
But, fer cryin out loud, Ubuntu, how many goddamn times do you need to change things around? UI? init system? Firewall management? It goes on and on.
This is the stuff of nightmares if you manage more than one distro version, and I feel I'm starting to slowly hate this distribution, even though I use it a lot and it works great (provided you remember which version does what all the time).
Their leadership projects a weak, hesitant image. There's no one at the rudder and the ship is being blown by the winds every which way.
Init system? are you brain washed by propaganda dude? Red HAt used and stil uses upstart in older versions, how can you blame Canonical creating it instead of using systemd that was not yet created???
The difference is that Ubuntu is already the de-facto dominant player here. As far as AWS goes, the Canonical Ubuntu AMI is the most popular one, except maybe the default Amazon one.
but dominant != revenue generating. Sure they make decent money with juju if you're needing help setting up openstack. Juju is actually underrated for what it does, and the engineers have done impressive work. But of "90%" of the startups using the Ubuntu AMIs on AWS, how many of them are actively paying Canonical anything at all?
Meanwhile, old boring Redhat just made 2 billion dollars in revenue and is still just slowly improving.
I might be wrong about this, but I'm pretty sure I read that AWS pays Canonical for a support contract on the official Ubuntu AMIs, to make sure they stay stable and up-to-date with the new Xen versions AWS uses, etc.
Anyway, even if they aren't making (significant) money on it now, having market dominance certainly puts them in a better position to profit eventually, than in mobile where they first have to unseat the three largest software companies in the world, on those companies' most lucrative markets.
Canonical has a commercial relationship with dozens of public clouds, ensuring that the Ubuntu in each of those clouds is securely produced, actively maintained, and at the standards and quality Ubuntu users expect. See the list of excellent commercial cloud partners at:
"How many of them are actively paying Canonical anything at all?"
All of them. Ubuntu gets a cut for every instance of Ubuntu on AWS. The Ubuntu AMI's are not stock Ubuntu, they've been modified. So Amazon has to pay Canonical to use the Ubuntu name or it has to call them something else. That's why Canonical is suing OVH; OVH is using a heavily (and poorly) modified Ubuntu without renaming or paying.
Hi Dustin, sorry for offtopic but why not use MATE, you have some MATE developers working for Canonical and I think Ubuntu MATE is more popular then the Gnome version.
I don't really think this can be chalked up to chasing fads. Any sane business owner is going to want to spend time and resources going where the money is. He makes it pretty clear in the blog post that cloud and iot is where the money is, and since Canonical is a for-profit entity that's where they've gotta focus their time and effort on.
That's fine; but the silliness is the "convergence" mantra. Different devices have different requirements, one size does not fit all; if you don't have the will to spend on multiple teams to build different gui shells, don't get in the game.
Microsoft learnt this the hard way, almost losing its desktop dominance to the chimera of converged UI; and it looks like all desktop vendors are slowly reaching similar conclusions.
Well, I hope they do not decide to start moving away from desktops.
While I am technically capable of installing most distros or compile my own kernel to enable the modules I need, usually Ubuntu has a good enough default kernel configuration that is compatible with laptops to a level I am satisfied.
I like the windows taskbar a little more... but prefer the top-bar (similar to mac), and don't care for the mac dock nearly as much as either. Unity is definitely better out of the box than most linux DEs imho.
Was anyone forcing you to use Unity against your wishes? Why are you glad I have one less option? In fact it was my daily driver for quite a while, I'm pretty sad about the news.
That's a shame. I prefer the current Unity (16.10) interface to the other available options. It's simple, clean and intuitive IMO.
Haven't looked at Gnome since the disaster that was 3, having previously used it exclusively. Here's hoping it has improved (dramatically) since. Unity is lightyears ahead of where Gnome was last time I checked.
I'm hoping some outsiders (outside redhat/fedora/gnome) will have enough influence to stop some of the dumb things going on with gnome. This could be good for everyone ;-)
If that would be possible we would not have MATE or people/projects droping GTK for Qt, Gnome has a vision and it will follow it, and a big chunk of people like it for some reason.
If you've been using unity, you've already been using most of the gnome stack, the main difference being a different shell and window manager on top.
I do slightly prefer unity to gnome (mainly because, in my experience, unity/compiz is has far smoother performance/animations than gnome shell), but gnome is pretty good these days.
Also, it's worth mentioning that the first release of unity was every bit as much of a disaster as gnome 3 was (I remember unity in 11.04, it was buggy and the performance was horrific)
I am somewhat exercised by this. I like Unity and have grown accustomed to it over the years, and in any case I think having choice in software is a positive thing. I hope that the community will continue development of Unity.
Well, this spells bad news for Qt and by extension any non-Gtk Linux environment. I was really hoping there would be more traction behind it to get a better non-C++ dev story; but it seems we're stuck with developing for two separate Linux environments (and other OSes) instead of writing once and having a binary that looks native in every OS with just a few platform-specific stylesheets here and there instead of entirely separate UI codebases.
That's one of the two environments I was talking about (coupled with LXQt and potentially if it ever gets anywhere, Liri), but KDE gets far less dev mindshare for a variety of reasons, and it shows in the quality of releases and size of ecosystem.
New Unity was to be built using Qt/QML. Qt plays ball with any environment and looks native everywhere (although perfect adherence to HIGs requires tweaks). Gtk looks native only in GNOME, thus, it is a horrible choice for portable GUIs (even throughout different Linux desktops).
Canonical getting behind Qt meant more effort was going towards making it a better development platform outside of C++, in fact early on they sponsored the development of Go QML bindings.
But now that's gone, and we're back to having no hope for that.
Last week there was a "Rust + Gnome hackfest" in Mexico City, attended by a bunch of devs for both projects. There haven't been any major announcements about it yet, but the general goal is to make it nice for writing applications for Gnome in Rust, and to expand Gnome's efforts to use Rust themselves.
Vala isn't in a good state today, it virtually has no maintainer. Gtk isn't in a good shape either. API keep changing, bugs are never fixed and maintainers seem to care very little about stability. Actually the whole GObject ecosystem is in a difficult situation.
IMHO, using Python is right now the only viable way to write Linux desktop apps without using C or C++ .
> The built-in official Gnome programming IDE ("Builder") seems to support "C, C++, Python, Vala, Rust and Ctags" too. [2]
Gnome Builder isn't financed by the Gnome Foundation AFAIK.
Yes, Qt looks good in Gnome because Qt tries to use your system dialogs and icons , GTK does not respect the user enough and it will use it's own dialogs on all systems. This is not only about the look but also the fact that my native dialogs are customized( shortcuts ) and those are missing from GTK apps dialogs
Gnome 3 is the best desktop environment I've ever used - my comment history on HN over the past 5 years definitely indicates as much. I'm not going to claim I've used all of them, but compared to what I've seen in Unity, macOS, and various versions of Windows (except Windows 10, which I have not tried), it's modern, fluid, and intuitive. To see Gnome return to the fold of Ubuntu is a huge deal to me and I couldn't be happier or more excited on this decision. Shuttleworth deserves a lot of praise and admiration for taking this bold (and likely painful) path.
(On a related note, I was reading up on Apple's switch from PowerPC to Intel earlier today, and it feels very familiar to this.)
I originally quit Ubuntu because I didn't like Unity ... Gnome is great, I can live with KDE but now I'm moving towards even lighter solutions. I won't miss Unity at all!
I'm personally using a tiling window manager (StumpWM, in my case, but dwm & i3 are both nice too). It feels, in every way, like the future of desktop interaction. And it's a lot lighter-weight than the GNOME desktop.
Doesn't do as much, granted, but isn't that kinda the point?
I've got XFCE installed but I'm working on learning RatPoison (a tiling window manager that doesn't use the mouse). It's a bit of a learning curve, so depending on what I'm working on, I switch between them (for now).
It's also a bit more of a pain to run multiple monitors with RatPoison and I've become well acquainted with reorganizing my screen configurations using the console commands to xrandr.
I also should have noted that I switched back to Xubuntu from Linux Mint.
Canonical is also ending development of Ubuntu software for phones and tablets, spelling doom for the goal of creating a converged experience with phones acting as desktops when docked with the right equipment.
I'm surprised there isn't more discussion here about that part. I'm really saddened to read that. When the project was first unveiled, it looked like exactly the sort of experience I want out of a smartphone.
+1
I had to double check the date to make sure this hadn't been posted on April 1st. I knew it was a ways away but I was looking forward to being able to have an Ubuntu phone.
This! I'm all glad about going back to GNOME and Wayland, but I really hoped Canonical was strong enough to put up a new mobile OS, even if just as a small player. It takes so much time (and money) to do so, and if even Canonical can't do it, I guess we can totally forget about a truly open source mobile OS in the near future, and maybe ever. That's a pretty sad future we're to, IMHO.
Agree. I was excited for this- even to the point of using and growing to like Unity. (That may not be fair though as I spend a lot of time in xterm so gui isn't so important).
The barrier to entry in the smartphone OS market is just too high for a small company to have a chance. Mozilla learned this, and now Canonical has too.
Neglecting the core product to pursue a phone OS that was very unlikely to succeed was bad for the company.
The barrier to entry in the smartphone OS market is just too high for a small company to have a chance. Mozilla learned this, and now Canonical has too.
Not only for small companies. Microsoft threw billions had at it, bought a phone manufacturer, had a pretty nice phone OS, had a name in business, and was years earlier then Mozilla and Canonical. And they failed miserably.
(Of course, the whole WP7 -> WP8 transition was handled badly, but it shows that even a company with extremely deep pockets will fail to capture significant marketshare.)
The thing is, Microsoft started pretty early: in 2010, when Windows Phone 7 was released, the Apple and Google app stores were not nearly as large as they are now, and a lot of people didn't own smartphones yet. I just looked up some statistics that show that less than 50% of American adults owned smartphones that year. The barrier to entry was a lot lower then, and it should have been surmountable.
I think Windows Phone failed primarily because Microsoft did not take early steps to cultivate a competitive app store, which they probably could have done back then -- these days, it would probably be impossible. They also didn't have very many desirable flagship phones, especially compared to the iPhone and the Samsung Galaxy series.
Microsoft's failure was partially self-inflicted, but Mozilla and Canonical would have had no chance even if they did everything right. They don't have the scale, and they are now up against app stores with millions of apps. It doesn't matter how nice the OS is if it can't run Facebook, Snapchat, and a bunch of popular games on day one.
I've always been sceptical of the idea of a converged desktop across different devices. One size just can't fit all in my view. Even a 'hidden' desktop on a mobile phone that reveals itself when docked to a large monitor will still be a constrained experience.
A dedicated desktop OS is capable of rich and complex interactions. Could a mobile 'desktop mode' ever match such capabilities?
I'd rather have a dedicated desktop OS that maximises features suitable for the desktop hardware it runs on. Not an OS that has to lower it's capabilities and interactions to match the smallest device it runs on.
Ubuntu being debian based is a huge reason for me. Apt is both more familiar and easier to use than yum/dnf, at least for me, but I'm sure for many others as well since most of the time, the first distro they try out is some form of Ubuntu.
Dnf? Oh, so they replaced yum? Which in turn replaced whatever it was RH used before...
In the meantime, apt-get has been there (and "good enough") from before yum even existed. I really don't like re-learning for the sake of it; the less I need to know to run a box, the better.
It's really not about which package manager is better. The fact of the matter is apt is "good enough" (as toyg mentions below) and more importantly, it's the ubiquity of the package manager that matters. Even Windows has Ubuntu now as the "bash subsystem".
People are used to apt. Switching to Fedora means learning an entire set of commands. While dnf's feature set is better in a lot of ways, that's just too much to ask when apt is there and it works for the majority of users, newcomers to Linux or otherwise.
Are you running KDE Neon? For the longest time I was using Kubuntu until I learned that it uses an ages-old version of KDE. KDE Neon uses a stable Ubuntu core with an up-to-date KDE desktop package.
Well I'm disappointed about them dropping their phone project. There's too few players in the phone OS market and Android just has a woeful security ecosystem; most shipped devices only receive updates for a year or two at most. I'm still on Android 4.3 and really don't have any trust in my phone.
I'll still hold out hope that someday an OS that is similar enough to desktop Linux - one that can ship updates independent of the carrier or manufacturer - will rise and fill that void.
And it's really unfortunate that more devices aren't more open... considering there are plenty of third party ROMs with newer builds, but only as long as the binary drivers can be ripped out of other devices for new versions.
I've been using mostly devices supported by third party roms from the start, but that doesn't always mean the latest and greatest hardware. Currently on a Nexus 6P and will probably move to Lineage when it's no longer supported.
Thank god. At no point did I like Unity. The way it integrated with the rest of the system seemed broken and all I ever got from it were countless problems.
I have grown used to Unity. Now they are stopping development on it? I checked immediately if this was dated April 1st...so this true?
I decided to stick with Unity because hit the sweet-spot for me when it comes to the functionality I actually want in a desktop environment. I was not happy with GNOME 3, nor KDE, nor any other desktop environment I tried. Now Canonical is pulling this again?
I just hope that there will be a UUbuntu flavour for the 18.04 release.
Next up ... will Mir be killed off too? ;) (Whoops ... didn't see that they are killing Mir too... :))
> By switching to GNOME, Canonical is also giving up on Mir and moving to the Wayland display server, another contender for replacing the X window system. Given the separate development paths of Mir and Wayland, "we have no real choice but to use Wayland when Ubuntu switches to GNOME by default," Hall told Ars. "Using Mir simply isn't an option we have."
This is the one part of the announcement that is a clear benefit. MIR has always been unique to Ubuntu and a waste of effort. If they put even half of the resources behind Wayland and Gnome, the remaining issues with them should get resolved quickly.
Killing off Mir would probably be a good idea as well, and just tracking back to Wayland... I also really liked Unity a lot, not as much as the windows start/taskbar, but more than mac's dock.
Hopefully they can do the canonical magic with Gnome 3, without too much reinvention this time.
Gotta admit- it's really cool. But I still more often than not fire up a linux under virtualbox on my windows box. I'm actually pretty delighted at how well vb works.
That is something entirely different. It's not like you are going to see difference in display servers and GUI shells, when you are using only console apps.
I don't care about window management, launchers, file browsers and all that other fluff. Usually you can customize that thing by e.g. using a custom launcher.
What I do care about is good hi dpi support, good font rendering, nice consistent controls etc. Somehow it seems that all these Linux window managers are just variations of superficial stuff (multi desktops, launchers, search...).
> What I do care about is good hi dpi support, good font rendering, nice consistent controls etc. Somehow it seems that all these Linux window managers are just variations of superficial stuff (multi desktops, launchers, search...).
Other than hi dpi support, none of that is under the control of desktop environments. And, at least for me, it's mostly fine these days.
152 comments
[ 3.0 ms ] story [ 223 ms ] threadIf someone could fix Network Manager that would be nice though...
edit: specifically, I think this was my issue: https://github.com/Supervisor/supervisor/issues/735
which seems to be patched now.
Yes, yes yes! That's fantastic! I hate Unity and I love GNOME. Good choice. Too bad it's going to be April 2018 until then, but this should teach anyone: It's never too late to revert a stupid design decision!
Does the desktop on Linux have a viable future?
First of all, I am a Windows user 95% of the time. But when I wanna spin up a VM or Linux distro real quick, Ubuntu is my choice. Of course, if I use it longer, I will install KDE or something like that. But with lots of window managers, I noticed instabilities or redundancies (e.g. there's a Unity and a GNOME service running for the same thing). It's best to use a distro with the window manager it comes with. Which was Unity on Ubuntu, but will finally be GNOME again.
Now imagine if 6 years ago Canonical had spent all that time and effort into improving existing DEs and implementing Wayland-support instead of their (already then) NIH initiatives.
Where would we all be now? Certainly not still stuck in this same, unimproved X11 sinkhole (yes yes, apart from Fedora-users, I know).
Such a shame they had to spend more than half a freakin decade to realize their mistake.
I don't think Canonical was unusual in having a failed investment.
I really don't care about it being X11 under the hood, I care about the sidebar on the left being gone.
https://micheleg.github.io/dash-to-dock/
I think that had they made its position configurable as half the Internet asked for since the start, Unity would have been way less hated.
Anyway, there's still a top bar, the one that has the clock and so on. Two bars doesn't save space.
"Shuttleworth made it clear that the cloud isn't the financial future of the company. He wrote: 'The <latest-big-idea> story for Ubuntu is excellent and continues to improve.'"
I don't have anything against Ubuntu, but I do have something against chasing the latest fad. I know, the cloud isn't a fad, but mobile wasn't a fad either.
Because of all the services Google provide on top of it. That's where the real value is, just like iOS, CyanogenMOD and co don't seem that popular. If Microsoft wants to make a come-back in the mobile space, that's where it needs to focus.
But, fer cryin out loud, Ubuntu, how many goddamn times do you need to change things around? UI? init system? Firewall management? It goes on and on.
This is the stuff of nightmares if you manage more than one distro version, and I feel I'm starting to slowly hate this distribution, even though I use it a lot and it works great (provided you remember which version does what all the time).
Their leadership projects a weak, hesitant image. There's no one at the rudder and the ship is being blown by the winds every which way.
Meanwhile, old boring Redhat just made 2 billion dollars in revenue and is still just slowly improving.
Anyway, even if they aren't making (significant) money on it now, having market dominance certainly puts them in a better position to profit eventually, than in mobile where they first have to unseat the three largest software companies in the world, on those companies' most lucrative markets.
https://partners.ubuntu.com/programmes/public-cloud
All of them. Ubuntu gets a cut for every instance of Ubuntu on AWS. The Ubuntu AMI's are not stock Ubuntu, they've been modified. So Amazon has to pay Canonical to use the Ubuntu name or it has to call them something else. That's why Canonical is suing OVH; OVH is using a heavily (and poorly) modified Ubuntu without renaming or paying.
Quite the opposite, in fact: https://www.ovh.com/us/news/cp2445.ovh_joins_ubuntu_certifie...
Does that mean that if I choose an Ubuntu image when installing an OVH server I won't get the OVH crap any more?
Microsoft learnt this the hard way, almost losing its desktop dominance to the chimera of converged UI; and it looks like all desktop vendors are slowly reaching similar conclusions.
EDIT - I am curious why I am being downvoted. Is the article not a reasonable summary?
Source link.
While I am technically capable of installing most distros or compile my own kernel to enable the modules I need, usually Ubuntu has a good enough default kernel configuration that is compatible with laptops to a level I am satisfied.
it is wayland vs mir, which had the potential to much more radically fragment linux GUI applications over the long run..
Haven't looked at Gnome since the disaster that was 3, having previously used it exclusively. Here's hoping it has improved (dramatically) since. Unity is lightyears ahead of where Gnome was last time I checked.
I imagine if there's enough interest, the same could be done with Unity.
https://github.com/mate-desktop Look at the commit graphs.. Their effort is super!
I do slightly prefer unity to gnome (mainly because, in my experience, unity/compiz is has far smoother performance/animations than gnome shell), but gnome is pretty good these days.
Also, it's worth mentioning that the first release of unity was every bit as much of a disaster as gnome 3 was (I remember unity in 11.04, it was buggy and the performance was horrific)
Canonical getting behind Qt meant more effort was going towards making it a better development platform outside of C++, in fact early on they sponsored the development of Go QML bindings.
But now that's gone, and we're back to having no hope for that.
The built-in official Gnome programming IDE ("Builder") seems to support "C, C++, Python, Vala, Rust and Ctags" too. [2]
Edit- A complete list of fully- and partially-supported languages for Gtk+ 3 is here: https://www.gtk.org/language-bindings.php
Edit 2- I was wrong about Vala! See comments below. :)
1. https://wiki.gnome.org/Projects/Vala 2. https://wiki.gnome.org/Apps/Builder
No, some Gnome developers blame Vala for all their problems and want to rewrite everything in Rust :
http://www.phoronix.com/scan.php?page=news_item&px=GNOME-Val...
Vala isn't in a good state today, it virtually has no maintainer. Gtk isn't in a good shape either. API keep changing, bugs are never fixed and maintainers seem to care very little about stability. Actually the whole GObject ecosystem is in a difficult situation.
IMHO, using Python is right now the only viable way to write Linux desktop apps without using C or C++ .
> The built-in official Gnome programming IDE ("Builder") seems to support "C, C++, Python, Vala, Rust and Ctags" too. [2]
Gnome Builder isn't financed by the Gnome Foundation AFAIK.
I think the only Vala maintainer right now is from Elementary OS, where they use primarily Vala.
Why should we do that? Both Qt and Gtk+ looks nice in gnome, so you could choose more suitable for you.
(On a related note, I was reading up on Apple's switch from PowerPC to Intel earlier today, and it feels very familiar to this.)
Gnome3 won't be a good move from Plasma5.
https://www.reddit.com/r/linux/comments/5l39tz/linux_distros...
Doesn't do as much, granted, but isn't that kinda the point?
It's also a bit more of a pain to run multiple monitors with RatPoison and I've become well acquainted with reorganizing my screen configurations using the console commands to xrandr.
I also should have noted that I switched back to Xubuntu from Linux Mint.
I'm surprised there isn't more discussion here about that part. I'm really saddened to read that. When the project was first unveiled, it looked like exactly the sort of experience I want out of a smartphone.
Neglecting the core product to pursue a phone OS that was very unlikely to succeed was bad for the company.
Not only for small companies. Microsoft threw billions had at it, bought a phone manufacturer, had a pretty nice phone OS, had a name in business, and was years earlier then Mozilla and Canonical. And they failed miserably.
(Of course, the whole WP7 -> WP8 transition was handled badly, but it shows that even a company with extremely deep pockets will fail to capture significant marketshare.)
I think Windows Phone failed primarily because Microsoft did not take early steps to cultivate a competitive app store, which they probably could have done back then -- these days, it would probably be impossible. They also didn't have very many desirable flagship phones, especially compared to the iPhone and the Samsung Galaxy series.
Microsoft's failure was partially self-inflicted, but Mozilla and Canonical would have had no chance even if they did everything right. They don't have the scale, and they are now up against app stores with millions of apps. It doesn't matter how nice the OS is if it can't run Facebook, Snapchat, and a bunch of popular games on day one.
A dedicated desktop OS is capable of rich and complex interactions. Could a mobile 'desktop mode' ever match such capabilities?
I'd rather have a dedicated desktop OS that maximises features suitable for the desktop hardware it runs on. Not an OS that has to lower it's capabilities and interactions to match the smallest device it runs on.
* dnf's basic commands are all obvious: install, search, update and info - unlike the seperation between apt-get and apt-cache
* dnf automatically fetches and updates the repo cache
* dnf can rollback entire transactins
* dnf supports delta-upgrades, which speed up things a lot on slower connections
* dnf makes unattended upgrades easy (just add -y), apt may still ask questions even with DEBIAN_FRONTEND=noninteractive
The new apt tool improves on some of these issues, but dnf is still a much better tool in my opinion.
In the meantime, apt-get has been there (and "good enough") from before yum even existed. I really don't like re-learning for the sake of it; the less I need to know to run a box, the better.
People are used to apt. Switching to Fedora means learning an entire set of commands. While dnf's feature set is better in a lot of ways, that's just too much to ask when apt is there and it works for the majority of users, newcomers to Linux or otherwise.
This is the second duplicate I have seen today on HN.
I'll still hold out hope that someday an OS that is similar enough to desktop Linux - one that can ship updates independent of the carrier or manufacturer - will rise and fill that void.
I've been using mostly devices supported by third party roms from the start, but that doesn't always mean the latest and greatest hardware. Currently on a Nexus 6P and will probably move to Lineage when it's no longer supported.
I decided to stick with Unity because hit the sweet-spot for me when it comes to the functionality I actually want in a desktop environment. I was not happy with GNOME 3, nor KDE, nor any other desktop environment I tried. Now Canonical is pulling this again?
I just hope that there will be a UUbuntu flavour for the 18.04 release.
Next up ... will Mir be killed off too? ;) (Whoops ... didn't see that they are killing Mir too... :))
Hopefully they can do the canonical magic with Gnome 3, without too much reinvention this time.
What I do care about is good hi dpi support, good font rendering, nice consistent controls etc. Somehow it seems that all these Linux window managers are just variations of superficial stuff (multi desktops, launchers, search...).
Other than hi dpi support, none of that is under the control of desktop environments. And, at least for me, it's mostly fine these days.
What they do get opinionated about are things that you mention (multi desktops, launchers, search...) so they work on those.
Not that that is a bad thing as such... but it leads to what you describe.