Last week, I had the chance to have a nice chat with Jonathan Riddell, Canonical employee and Kubuntu maintainer.
For years, Jonathan was paid to maintain Kubuntu. In a recent move, Canonical announced that Kubuntu will become a community-only project. As a way to start the conversation, I poked him about that:
— What happened? Is Canonical dropping KDE support?
— Well, we are doing with KDE exactly what we did with GNOME.
— Indeed. But what is the reason?
— Canonical seems to think that none of them managed to reach a non-geek audience.
And, sadly, I had to agree with that.
What is a desktop environment?
Since I’m involved in GNOME, I don’t remember being able to explain what the GNOME project was to any non-geek. Ubuntu is something people understand. Linux is an harder concept but still manageable: it’s the heart of the system, something invisible and magic that handle everything. But a desktop? People just cannot tell the difference between a desktop environment and an operating system. And, when you think about it, the whole definition is somewhat arbitrary.
Who cares about the desktop environment?
In fact, the only people I know who really care about what is GNOME are… GNOME developers and fans. What might illustrate this better is the astonishing lack of reaction which followed Ubuntu’s move to Unity.
Yeah, sure, bloggers have discussed merits of Unity vs GNOME 2 vs GNOME 3. There was some buzz about Linux Mint. But, outside the geek microcosm, what have we seen? Nobody really cares. Ask any casual Ubuntu user: most didn’t even noticed it was a big change. Some say it’s better, some prefer the old way but don’t make it a big deal. In fact, most of the blog posts even agree on this: Unity and GNOME 3 are, well, different. No winner.
People want an application launcher, that’s all.
Shifting out of the desktop paradigm.
The world is increasingly shifting away from the standard desktop: iOS, Android, Metro. There’s currently a quest going on to make the computing experience very similar on small devices or on bigger television, including tablets, netbooks, laptops, desktops.
That experience will be the main selling point of OS vendors and it only makes sense for an OS vendor to reclaim control about its own destiny without depending on any external project. Which is exactly what Ubuntu did.
What should a desktop environment be?
When KDE and GNOME appeared, Linux was mostly seen as technical environment. Developing a desktop was logical and highly needed. Lof of what we take for granted today should be credited to those who pioneered KDE and GNOME.
But we reached the point where there’s no clear separation anymore between what is part of the OS and what is part of the desktop. Should the configuration tools be part of the desktop or the OS? Even the package management is now offered as an OS independent layer in the desktop.
This explains why I’ve never successfully explained what GNOME was: what it is and what it has to offer is arbitrary and not clear.
In fact, I see two possible futures for the GNOME project:
1) Being a software catalog. Offering software which have similar design goals and let the OS pick what they want. Sometimes, multiple alternatives for the same need can be offered. This makes the shell mostly irrelevant or anecdotal in the whole GNOME project and it is exactly the way Ubuntu uses GNOME.
2) Offering a complete operating system and controlling everything from the kernel level. This idea is sometimes referred as GNOME OS.
As we said in French, we currently have our ass between two chairs, not really able to take a decision, which is the worst situation.
And what about Unity?
The main criticism about Unity is that it is “yet another desktop”, fragmenting the community. But, in fact, Unity is a pure Canonical project like Android is a pure Google project. There’s no involvement from the community. Canonical wants to be able to control the appearance of its core product and who will blame them for that?
To their credits, it can be added that they tried to play it fair first with the “netbook remix edition”, which failed to gain any attention from upstream.
And is GNOME really better? GNOME-shell design decisions are taken by a handful of designers, most of them employed by Red Hat, which has no interest in smartphone/television. Has the wide community anything to say in the design process? Not much. And that’s a good thing if you want to avoid the bicycle-shed/UserLinux syndrom.
So, all technical qualities set apart, what is our problem with Ubuntu?
Maybe, what we hate with Unity is that it proved us that we were a small circle of geeks, that most users don’t care and didn’t even noticed that they were switching their desktop. The desktop war looks like the window managers war of ten years ago: all the geeks tried to find the best one while, in the end, it appeared that what user wanted was just to move windows. And none really won. WM, those day, are just anecdotal projects that only geeks care about.
The future?
Unity seems to have quite a clear future: it will stay and evolve as the Ubuntu default interface, from Ubuntu TV to any Ubuntu device, offering quite a consistent experience while you stay in the Ubuntu world.
But what is the future of GNOME and KDE? How do you see it? What will they offer? Do you think it’s a good idea to leave Ubuntu in order to keep GNOME at all costs? Should GNOME work on GNOME-OS?
What if, all irrational feelings set apart, we realized that Unity was the right move?
Picture by Carsten Knoch
The What if Ubuntu were right? by Lionel Dricot, unless otherwise expressly stated, is licensed under a Creative Commons Attribution 2.0 Belgium License.
Leave Ubuntu? That would imply that I ever came to Ubuntu in the first place.
Your analysis makes sense, but there is a thing you may have overlooked: the way Gnome apps are able to integrate which each other sets Gnome apart from something like Android. (Probably goes for KDE as well.) I think a ‘Gnome OS’ is feasible but it doesn’t have to be the Gnome project itself who builds it, as long as the technology is compelling enough for vendors to buy into it as a whole.
The world is not moving away from the desktop paradigm and Metro on Windows 8 will fail since the average user does not want change, they will want to know where the Start menu is. Once does not just simply change the entire way we interact with the OS willy nilly. Metro, Unity and Gnome3 are just as bad, they assume that we only want to use one application at a time on machines that are capable of far more! At the moment I am just glad that Gnome Classic is available, it being a much saner environment than Unity and Gnome3.
The future must be for Gnome OS
@commnet 2: Back in the days of the great migration to Windows95 I was having a chat with a friend of mine:
- Hey, have you seen the new 95 thing?
- Yeah…
- Is it good?
- Nope! It’s awful – everything you need is concentrated in a single Start menu… a lot of clicking… Oh! And when you copy files, there’s some silly animation of how sheets of paper fly from one folder to another…
- Sheets of paper flying!!?? You’re kidding me. That’s evil!
Unless history’s wrong, Windows 95 was far from a failure even though the average user of the time was mostly disturbed and displeased with the radical changes.
I wouldn’t jump on the “Metro will definitely fail, because desktop is what users want”. Those guys at marketing don’t get their bucks from the social services you know.
I never understood why there are Gnome applications at all. I think the whole notion of a desktop environment is an error. All we need is a compatibility layer and then let the users assemble their own desktop with the applications that are available. If somebody wants to use Nautilus, good for them. If somebody wants to use something else, good for them too.
Hi,
I want to try to give a answer to two of your question:
What is a desktop environment?
Who cares about the desktop environment?
Today a desktop environment is a crucial part of every desktop OS. It’s that what the OS presents to the user and which enables the user to interact with the machine.
So every user cares about the desktop environment. Especially casual users care really much about the desktop. In fact for most casual users the desktop _is_ the OS. If the desktop changes than their OS has changed. The casual Windows user identifies Windows by the desktop an not by the kernel or some system libraries and the same is true for the casual Mac user and the casual Ubuntu user. I’m quite sure that the casual Ubuntu user noticed that his desktop changed but for him it was the OS which changed (through a upgrade to the new version of the OS) and thus he didn’t care much about the changes. He installed a new version of Ubuntu and get a new version. In fact, because for most casual user the OS is what they see. They even expect that something at the desktop changed after an upgrade, else nothing changed at all for them and the OS stayed the same.
Think about a simple experiment. Take to computers one with Fedora and a default KDE and one with Ubuntu an a default KDE. Now present both computer to a casual user. He will tell you that both computer run the same OS. On the other hand show tham an Ubuntu system with a default KDE and a Ubuntu system with a default Unity system and he will tell you that both computer run different operating systems.
So to make it short. The desktop environment is at the heart of every OS. Much more than 10-20 years ago where it was just an addition to the shell. Therefore everyone cares about the desktop environment, even that much that today for most casual users the desktop environment is and defines the OS. That’s why KDE, GNOME, Unity,… become more and more important while it is not that much important what runs behind the curtain.
“But, in fact, Unity is a pure Canonical project like Android is a pure Google project. “
I can’t speak for Android, but I ran some rough statistics on the lp:unity branch commits, and this is what I found:
84 total contributors
28 were members of the unity teams
56 were from outside the unity teams
Unfortunately I couldn’t separate the non-unity contributors into canonical and non-canonical, but anybody not in one of those teams is likely contributing the Unity as a member of the community, not as an employee of Canonical.
I dislike Unity for a number of reasons, but the main one is that I get the feeling Canonical is building a house of cards on top of moving sand. Unity has a ton of outstanding bugs, lack of recent releases, serious stability issues, and the lead developer has already admited his failure in getting the project back on track (https://smspillaz.wordpress.com/201…).
Things are likely to get even worse with dimished community involvelment now that distributions are dropping Compiz altogether. That I know of, OpenSUSE, Fedora and Gentoo have already dropped it.
In the long run that leaves Canonical maintaining both Unity and the underlying compositing manager on their own, with comparatively little community involvement, at least when compared with GNOME and KDE. And it shows, I lost track of the number of bugs I found and filed for Unity after the 11.10 release, which was markedly worse in this regard than the previous one.
Pulling an Android, like you described, is a valid strategy, but only if you have resources to throw at the problem, which Canonical doesn’t. Their recent forays into the world of TVs (really?…) isn’t helping. In my opinion they completely lost their focus, shunning the advantages of the Linux ecosystem (the community and re-use of upstream code).
“Unity has a ton of outstanding bugs” –> Ooops, I meant Compiz.
@CMD : Why do you think you can only use one application at once with Unity? I am hardcore-keyboard-shortcuts-multitasker. Even my non-geek wife multitasks fine with Unity, including a Vbox Windows XP VM in a windows…
I really agree with your 2 possible futures for Gnome.
Personally, I prefer future 1. Gnome should remain an open platform.
Gnome was and should remain a bunch of apps that interact well. It’s all about interoperability.
I would be glad if more people start considering Unity to what it was attended: just another shell for Gnome. This does not mean that there is no space for Gnome-shell and a “Gnome-shell-OS”, but I think they should be built on a same flexible and easily extensible “core” (which include system-settings, session…).
It makes me sad when I see the amount of time Canonical spend on patching Gnome. I know they tried to send some of these patches upstream, but it’s not that easy (cf. indicators, notifications…).
Some of these changes conflict with the Gnome vision, either in a technical or a political (=design) way.
And this is really a shame… Why can’t everyone sit down and think of a common and customizable (in a low-level way) platform that may benefit everyone?
I thought Fedora is about to become GNOME OS, with GNOME being developed by RedHat employees in the same way that Unity is being developed by Canonical (i.e., some contributions from outside, but decision making mainly from inside).
With GNOME tending to rely on services like systemd, it could become as “Fedora only” as Unity is “Ubuntu only”. It used to be seen as “Canonical left GNOME for its own baby”, but it might as easily be seen as “two major linux desktop vendors making their own thing each”. And all others can choose to pick one or the other.
Though I don’t prefer Unity over GNOME, one area where it really excels is that it chooses the best default application for the job.
Currently, many of these are not GNOME applications. Totem, Epiphany, Evolution and Empathy are striking examples of desktop software that’s absolutely “not there yet” and each of these has had years to mature.
GNOME’s insistent focus on integrated apps must either:
a) Shift to choosing fully functional defaults
or
b) Shift to ensuring these apps reach a reasonable level of maturity before including them in the default desktop.
Currently many GNOME applications are no better than technology preview status, and that’s a shape, because they sure are pretty.
I think Canonical made the right move when they decided to go the Unity way. Depending on other projects would make Canonical’s uphill battle even harder.
I find Unity very easy to use, it’s intuitive and having the Launcher always visible makes things even easier for average users.
My family uses Ubuntu. My mom is a horrible computer user, she never got around Win7, maybe because her first computer experience was with Ubuntu 10.04, which I modified with Docky because she never understood menus at all, it was too hard for her and my siblings. When 10.10 landed, I gave her Unity, yes the really crappy Unity. She got around it very quickly since all her favorite apps were right there in the launcher and there were no confusing menus anymore.
Now I installed Ubuntu 12.04 and she, like most people didn’t even noticed it.
I’m glad she never calls me for support.
I didn’t install Gnome Shell in their computer since the launcher is always hidden and that proved confusing and required one too many steps to get to for them.
I can handle almost any desktop, I like Shell, Unity and Pantheon, but Unity works for me.
@Bjoern : “In fact for most casual users the desktop _is_ the OS”
Agreed. What us geeks regard as the “OS” is of no interest whatsoever to the average computer user. The important part is the system they interact with, and that’s the desktop and applications. The average person don’t use the operating system Windows – they use the desktop interface Windows.
There are difficulties with both possible futures you present.
The problem with being “a software catalog” is that — as Brad Laue alluded to above — none of the Gnome applications are close to being the best application in the world for anything. This matters because it means Gnome developers are less likely to understand what toolkit and library features would be necessary to make an application the best in the world for something. (One simple example: if your applications are so feature-starved that they don’t need menus, you’re less likely to do a good job of providing good toolkit infrastructure for menus.)
The “GNOME OS” idea, though, is even worse. First, it assumes that “GNOME” is a brand that any customer would be interested in, when it’s more likely they’d giggle at it. (Latent examples of this are the Gnome developers who mutter that “if you aren’t using Gnome Shell, you aren’t using Gnome” — as if any non-geek cares whether they’re using Gnome or not.) And second, it assumes that a mass-market operating system consists of software alone. That stopped being true in the early 1980s. Nowadays an operating system consists of software, a distribution channel (possibly including retailers and resellers), a hardware certification program (if not all-out hardware integration), marketing, various online services, an application marketplace, an SDK, administration guides and training programs if the OS is used in business, and a carefully nurtured network of application developers. A few of those components could develop in an all-volunteer or vendor-neutral way. But as Debian and others have demonstrated over the past two decades, most cannot.
I think Gnome should continue as an umbrella project for developing some applications, but not as a software catalog or a gatekeeper. Instead, it should concentrate on those applications that can drive development of lower levels of the stack. What UI toolkit features are necessary for a beautiful color-coded calendar application? What networking API changes are necessary so that applications stop sucking so hard when connecting to the Internet through a proxy? What XML and animation libraries need to be in place to make it trivial for System Settings to let you animate an RSS feed as your screensaver? What networking and discovery features are necessary for a file manager to solve the problem described in http://xkcd.com/949/ ? And so on.
It’s always satisfying to see a member of the Gnome community reach those conclusions, since that makes you think some folks are not as impermeable to the obvious as seems to be the concensus@gnome.org
I thought that, in the days of Gnome2, the community was basically doing everything right. Excellent core applications, simple desktop layout and sane defaults. Make the computer useful, and get out of the way.
And then, The Shell.
Between the 180 degree turns in philosophy (differentiation as a motivating factor), opaque development (getting ignored on IRC as soon as folks realized I’m not running Rawhide), bizarre technical choices (JS? really? because “web developers will be brought on-board” – except the extension framework changes every release, and aesthetically-minded designers were confronted with Cantarell), it’s not hard to see why everyone is forking.
That, and the gnome.org inability to accept any proposed changes that came from outside RedHat, instead clamoring how GREAT our community is, how INCLUSIVE and MERITOCRATIC it is, while it’s looking more like every maintainer runs a personal fiat over their projects and blocks out any dissenting opinion without feeling they have to justify the strange choices they foist on users, well, that makes a guy want to write sentences that grow wayyy too long.
Of course users don’t care about the desktop. They shouldn’t. It’s like a real, physical desk – you only notice it when it’s getting in your way. Users don’t care about systemd or D-Bus or GStreamer either, and they only care about GTK vs. Qt when the themes clash.
The current round of fragmentation has left people more entrenched than ever. Unless Gnome As A Whole starts reaching out to those who left, I think this will keep happening.
It’s been said before: if Ubuntu has been successful is because they solved all those annoying things related to installing software, solving in a very smart way the installation of privative (but necessary) software and having everything working out-of-the-box. This doesn’t happen with Fedora, OpenSUSE or other distributions.
GNOME OS should be the answer. Get Ubuntu, solve the problems packaging gnome and create GNOME OS. Is it that complicated?
“But what is the future of GNOME and KDE?”
I don’t know about GNOME but I can say that KDE has a future like Unity with Plasma Workspaces and especially Plasma Active, because there hasn’t been a Environment specifically tailored for tablets ignoring Android. And there was news recently about Plasma Active images by the Kubuntu project being made.
The fact that LInux is for the Geeks has nothing to do with the general usability of GNOME2 or KDE. Hence, the conclusion that these two DEs “didn’t make it” to achieve mass adoption is IMHO wrong. Both of these two DEs were similar enough to … eeehm … “the leading desktop environment” to be adopted by the masses (in fact the masses did adopt Win7, coming from XP, they could have easily switched to KDE or GNOME2 any day, usability wise).
LInux is for Geeks because of the many reasons that LInux is for geeks
… it’s the totally anti-competitive environment that is created and maintained by two companies. In all other areas of computing, these two companies do not have under their thumb are happily adopting LInux en gros …
“What if, all irrational feelings set apart, we realized that Unity was the right move?”
amen to that.
@mpt :
> The problem with being “a software catalog” is that — as Brad Laue alluded to above — none of the Gnome applications are close to being the best application in the world for anything.
It’s the very geekish opinion.
Although I agree that there are no good alternatives for some applications which were builded with another toolkits.
But the question is do we need this?
E.g. Thunderbird, I don’t use it even on Windows. Because I don’t need all its features and extensions, I use Evolution and even look forward to Postler will become better. On Windows I use Live Mail and this is enough.
E.g. Firefox. Although my mom uses Firefox I don’t see any reason why she can’t use Epiphany, the only reason that Firefox was installed by default on Ubuntu, and I was lazy to change it.
And so on…
Many applications created by Microsoft or Apple sucks in different ways, but the fact is people use it, because these applications are good enough.
“Has the wide community anything to say in the design process? Not much. And that’s a good thing if you want to avoid the bicycle-shed/UserLinux syndrom.”
That’s a good thing if the designers get it RIGHT for most users, which can only happen if you take user feedback into account.
At the present moment, what was whole and good for the community (both users and devs), i.e. the Gnome 2 desktop and environment (or ‘software catalog’ to use your definition), has been fragmented two ways:
- the ‘old UI paradigm’ vs ‘new UI paradigm’, i.e. people running from GS/Unity to other environments (Mate, Linux Mint MGSE, Cinnamon, XFCE, keeping good old Gnome 2.x running);
- within the ‘new UI paradigm’, the split between Gnome Shell and Unity
In other words, the already scarce resources are even more scarce now. There COULD be a solution:
1) merge GS and Unity to produce a ‘Gnome Unity’ at some point in the future;
2) make ‘Gnome Unity’ customizable, so that ‘conservative users’ (yeah, sure) can adapt their work environment to their liking instead of the other way round.
Sadly I can’t see either 1) or 2) happening, 1) since both Gnome and Ubuntu are fairly focussed on their pissing contest aiming at the big pie in the sky (good luck with that), and 2) because it would require the humility of admitting “ok, we didn’t get it right, at least not for all users, let’s try to fix it” that neither camp has so far shown till now.
In conclusion: while in the current situation the Gnome OS thing would make sort of sense (for Gnome to be visible and recognizable: not that I think that it will do them a lot of good in the long run), I surely hope that the ‘software catalog’ Gnome will stay for all devs like Clem of Cinnamon fame to build upon and bring good stuff for the ‘old UI paradigm’ side.
Rehdon
“What if, all irrational feelings set apart, we realized that Unity was the right move?”
Then we would just rewrite some of our code to mimick it and tell planet gnome it’s so much better than the evil Unity.. — oh wait what?
I really don’t understand why everybody seem to refer to choice as “fragmentation”, these days.
A desktop environment is how an OS and the application present themselves to the user. How they do that should be the user’s choice, ultimately.
Today users are offered “sets” of these presentations: iOS, Android, GNOME, Metro, KDE… Some of them are more customizable than others, but effectivly they are sets of “looks like this and works like this”.
I think you are right that developers aim to give users similar experiences on different devices, but I think that they overdo it and kind of force users to the _same_ experience on different devices.
What GNOME should become is a standardized interface that developers can use to describe in an abstract functional way how they invision the presentation and workflow of their software, but ultimately let’s the user decide individually how she or he wants it to look and work.
That way the user could design the UI and the workflow on her or his desktop PC to be quite different from the UI and workflow in her or his smartphone, but still know that a GNOME-App will integrate seemlessly in both environments.
A graphical desktop environment exists to give users something to play with, so they can pretend that they are doing work. The ultimate is the iPad – where people pay large ammounts of money to have a device that lets them open and close apps by waving their fingers around – but you cannot do any real work on it.
It is not only about desktops. The big underlying reality is that we have become individual people with a great sense of what we like and what we don’t like. So when the paradigm of working on a desktop changes by ‘some one else’ we see that as an interference with our own style of working.
The answer could be making the workspace more like an ‘app’-space where you can put your ‘own’ apps in place (then you would have the feeling that you make your ‘own’ environment then getting one from ‘somebody else’. ).
You like a toolbar, then you can activate an ‘app’ that gives you one, you like a panel or rather a ‘dash’: just install what you like. What gnome should do then is just give the API to make apps and give the environment in which programmers can ‘play’.
Let the user decide how he wants his ‘app’-space. And for those who can’t decide: make some bundles with ‘preinstalled’ apps: gnome-with-panel, gnome-with-dash, gnome-windowmakerstyle, etc.
“What if, all irrational feelings set apart, we realized that Unity was the right move? “
So the question is do we have to re-invent the wheel alone from our side (as Canonical did), or do we have to cooperate to develop a broadly adopted DE ?
From this perspective I fail to see how the Canonical approach could be the right move. Sure it’s easier and quicker to deal with, as you can impose your own vision… But clearly it doesn’t benefit from the FOSS development model.
Very well written blog post, with a thought-provoking message. I can only imagine how long it took you to write this down and get it into shape. Well done!
.Canonical seems to think that none of them .managed to reach a non-geek audience.
Who is them?
KDE?
Please dont make me laugh with that.
Jon is paid by Canonical so he will follow the company line but are you REALLY going to make anyone believe that canonical tried to promote Kubuntu or KDE?
If it did, it did a pisspoor job of it because most people would tell you no.
I get taht they pushed Gnome but KDE? Yeah, ok… they ‘tried’ and failed is how the official BS line is going to stand it seems.
I get why Canonical went with their own DE, to actually differentiate themselves from every other Gnome desktop. Because lets be honest, the distro is NOT the choice that matters, its the desktop.
Give three different distros to someone that uses the same DE and you will see that the first question will be ‘Isnt this the same thing but with different icons?’
How can you sell Ubuntu as being ‘the bestestest’ in the world when Distro X uses the same DE and looks the same?
There needed to be a visual difference between Ubuntu and the others. Something to be able to say “see, THIS way is better-easier’ when in reality its only different.
The reason Ive been able to switch so many people to Linux IS the fact that we have desktops to suit everyones tastes (and the ability to not be insulted like some when ‘my’ distro or DE is rejected by newbs) and that while YOU or I may not like something, doesnt mean we own a monopoly on subjective questions like “which DE is the ‘best’” or ‘which ice cream flavour is the ‘best’”.
Stop it with the stupid memes like “Linux is for geeks” or “Linux isnt as easy to learn”.
I have more than two handfuls of seniors that could tell you otherwise and just as many kids under 12 for whom ive solved computer problems (and myself time from having to fix the same Win problems-virus-crap).
Unity is no better than Metro, Aero, KDE, XCFE, Cinnamon or any other newer desktop.
Its different.
Some will like it and some will hate it.
But luckily at Baskin-Linux desktops, we have a flavour that will suit YOUR tastes because we dont leave anyone behind.
Had we had only one choice my wife would have seen her new laptop with Ubuntu 8.04 a few years ago and went back to WinXP right away instead of having her try something else.
@vovkkk : Please note that you’re advocating the inclusion of mediocre apps because they’re “good enough” – while you and others may not use Thunderbird, the fact is that it’s significantly more stable and easier to use than Evolution is.
These alone are reasons to prefer it, and I don’t think that’s a geekish opinion.
This should be required reading for Gnome devs, as it explains very well why Gnome Shell has made Gnome irrelevant:
http://www.tomshardware.com/reviews…
Please stop using the line “it’s no better, or worse, just different.” – it’s much much worse.
@Tom’s hardware :
Very insightful conclusion on the Gnome3 fiasco.
Is this all that Canonicalites can muster these days? All I see coming to the defense of Unity are frothing fanboys armed to the teeth with condescension. Look, I recognize that the amount of people who can make the distinction between an operating system is a lot smaller than the group of people who can tell the difference between their elbow and their nose. I understand that I am part of this small group, and the (probably) even smaller group of people who actually care. But don’t talk down to me and tell me Unity is “right’ like there’s only one good answer for a user interface. What if I am wrong and Unity is the “right” direction? Oh for fuck’s sake, go suck off Shuttleworth a bit more.
I agree with your post. People are focusing on apps, the desktop environment or the OS is a gray area where nobody non-geek cares.
Unity vs Shell is a non problem if you can run the same apps.
@Ernest : There are Gnome apps because they comply with Gnome’s HIG and developer recommendations (GTK+, Vala, …). The new planned Gnome 3 tabs will support title-bar-less full screen, have a specific toolbar arrangement, will utilize an overview, will drop double-click support and instead use a selection mode, etc.
You’ll still be free to use KDE apps, they just won’t fit into the DE because they won’t follow the new Gnome HIG.
@Tom’s hardware : I actually really like Gnome Shell, except for the fact that it hides the Power off button by default. My non-geek sister finds it the best UI I’ve shown her so far and uses it as her main PC interface.
I feel that the author is basically saying that he prefers KDE. And Unity. Fine. The argument seems to be that Gnome features empty space. Fine. I prefer Gnome Shell and its empty space (feels nice and roomy), enough to install it over Unity on Ubuntu. Best thing is, the UI works really well on touch screens.
Interesting post and comments. I think there are three competing visions that inelegantly share our “bigger than 40cm” computer screens: 1) A windowed desktop running multiple programs that operate on a local file hierarchy; 2) A tabbed browser where we interact with web pages that mostly ignore the local file hierarchy; 3) Smartphone-like full-screen apps that hide the local file hierarchy.
Nobody has figured out how to meld these, so it’s not surprising that Linux desktop environments are in a state of churn along with Windows 8, i/Mac OS, etc. And average computer users are bewildered; most of the time they do everything in the browser, until they’re faced with a “Save as” dialog or they plug in a camera, at which point they just fill random directories with crap they’ll never find again.
I think the inexorable trend is towards browser-based software (which thanks to HTML5 *doesn’t* imply always needing an internet connection, or storing all your data in the cloud). But Apple/Google/Microsoft resist this because they can monetize an app store and because they’re big enough to get platform-specific development. (Note how *all* would-be smartphone and tablet competitors start with “write HTML apps” until they get big enough to sing a different tune “Write proprietary apps for our app store”.) But Linux desktop environments aren’t in the game for $3.99 apps. So I think they should *give up* on applications written for their toolkit and just work on a lighter-weight OS that provides a fantastic browser environment for great HTML5 applications. The raw source code is out there with Chromium OS, Boot to Gecko, Tizen, etc. The innovation lies in figuring out how to present a useful coherent vision for bookmarks, files, application icons, app tabs, recent documents, and all the other paradigms that currently conflict and confuse us.
“What is a desktop environment?”
A desktop environment is a computer analog, but not a simulation, of a physical office.
“Who cares about the desktop environment?”
Almost no one who has one. The better the desktop environment, the less you care about it, and that’s as it should be because you’ve got work to do.
“People want an application launcher, that’s all.”
I have a thing called a “lihin”. You don’t want one because you’ve never heard of one, so you don’t know how awesome it is. You may have a “neir” or an “adan”, which are similar, but if you ever saw my lihin, you’d never want those things again.
“The world is increasingly shifting away from the standard desktop …“
Only trivially. The “personal computer” is what happened when the workstation started to be used outside of its niche. Something else was a better fit, but that something else wasn’t available. The niche for workstations still exists and MacOS X is taking over it.
“There’s currently a quest going on to make the computing experience very similar on small devices or on bigger television, including tablets, netbooks, laptops, desktops.”
Yeah, well, we’ll see how that works out. I think there’s more than enough difference to warrant different interfaces.
“What should a desktop environment be?”
Even better than the best one available now, which is Mac OS X. (Ignoring software freedoms, that is.) Think of all the things you can do in an office, a real office, an expensive office, with people who work for you, with space and the budget for everything you need. Think of all the things people do in their offices, or studios, or shops, or studies, or examining rooms, or laboratories. Think of the desktop environment as the place around a real desk (or two, or an easel, or a drafting table, or a workbench) and not as a computer environment that simulates a desk. After all, who in his right mind would put a trashcan on a desktop?
One of the great advantages of Linux is that you can have different application toolkits, applications, window managers and desktop environments and choose the best one that fits how you work, be it Gnome Shell, Unity, KDE, Enlightenment, XFCE or another environment. KDE applications fit in as best they can with Gnome and Gnome applications in KDE.
Try that with Windows 8.
A desktop environment is a collection of things — applications for music, documents and web, a window manager, a file manager.
I don’t just want an application launcher, I want a desktop environment/window manager that supports (in a coherent, integrated way):
1. snap left/right/full screen at a minimum; snap quad is a bonus.
2. arbitrary window positioning and resizing, including a resize hit test area that is not ridiculously small.
3. quick launch for applications I use frequently.
4. not get in my way when I am using and interacting with applications (the Unity launch bar is bad at this).
5. being able to see and switch between active applications easily.
6. being able to run multiple instances of an application and manage those in a sensible way.
7. being able to switch between multiple virtual desktops.
8. being able to get to a usable desktop if a full screen application has locked up/crashed.
9. being able to easily terminate hung applications.
10. being able to browse, organize and manage documents on the file system and open the files in an application I have installed (including setting the default and opening in another application).
11. keeping me aware of application notifications (new email, browser updates, etc.) without getting in my way.
From a user choice point of view, I see having KDE, Gnome Shell, Unity, Cinnamon, Enlightenment and other shells/window managers as a big win for users.
Also, I see the choices of big applications as being a win for users — whether you choose Epiphany, Firefox or Chrome; Nautilus or Dolphin; Rhythmbox, Totem or Banshee; etc.
The main challenges are:
1. Sharing core technology and libraries where possible
this is where I see “fragmentation” being an issue, as each environment is maintaining its own libraries to do specific functionsto avoid duplicating developer effort where it can be avoided.2. Using standardized (or at least common) protocols and file formats, e.g. music applications can use a standard protocol to expose play/pause and other functionality so that volume tray icons can show you what is currently playing and allow you to control it from the tray irrespective of what environment you are using.
Having multiple desktop environments keeps everyone honest and allows for better integration in the different environments.
@vovkkk :
“E.g. Firefox. Although my mom uses Firefox I don’t see any reason why she can’t use Epiphany, the only reason that Firefox was installed by default on Ubuntu, and I was lazy to change it.”
I see, for example segfaults in “Web” (epiphany)
@Melvin : Can I know how old is your mother? My grandfather (he’s 82) just bought a computer and don’t know how to use it (it was sold with W7)… If he’s able to use Unity (and Ubuntu in general…), maybe I will give him it…
@Reece H. Dunn :
> 4. not get in my way when I am using and interacting with applications (the Unity launch bar is bad at this).
I installed Ubuntu 11.10 with Unity for my wife and this issue, along with having the window buttons on the left made the whole thing completely unusable.
She’s using many www-sites that have interactive elements on the left and launch bar popping constantly over them made the whole Unity unusable. There isn’t even any setting to keep it always on screen like panels on gnome 2 had.
The window buttons being on the left is just obnoxious differentiation. She doesn’t (cannot) have Ubuntu at work and window close button being in different side of the window is just huge redundant cognitive dissonance between the machines she needs to use. (In Unity she kept hitting the Browser bookmark star icon when intending to close a window. After a while, extra bookmark windows get boringly tiresome.)
A smaller issue was the top panel being used as fullscreen app menu bar. As her laptop screen has 4:3 ratio and nowadays a 1024×768 resolution, apps with more menus didn’t fit to the top panel, some of the app menus were hidden below the panel items.
I installed Gnome3 for her and that’s just fine, it actually seems to be designed for a people who need to use computers for work, not just “immerse” themselves with “Unity”.
On another, older machine she’s using XFce and that’s OK for her too. Unity isn’t.