A Disturbing Dialog About Ubuntu and Unity
linux-magazine.com
linux-magazine.com
There is an enormous class of design problems that have several appealing solutions, one of which is better than the others, but not so much better that it's worth the time to find it. "Where should the launcher be?" is very likely a member of this class: There probably actually is a best place for the launcher to be, but it's better to just have it in the wrong place than to have to argue about it on a bug tracker, and certainly better than having to make a configuration drop-down for it.
The entire reason you have a design team is because design-by-complaint reliably produces shitty software. Unity exists because every other Linux GUI has fallen victim to this process. Will it succeed in not being shitty? I have no idea. Ubuntu will sink or swim on the merits of its leadership. But the only way to avoid this one major failure mode is for Shuttleworth to ignore, as politely as possible, the complaints of his users.
> Unity exists because every other Linux GUI has
> fallen victim to this process
That's very debatable. According to me, Gnome 2.3 is by far the best designed UX for a desktop operating system. Like most people, I generalize based on my own limited experience, and I have seen people grow up and learn how to use a computer in Gnome who were disgusted by Windows and OSX user interfaces. Like all animals, humans can be conditioned and there's simply no such thing as "best UX for all".Lets look at Alt+Tab functionality, which now in Unity (and Gnome3) switches between "apps" instead of windows. The problem with Unity/Gnome3 is that they're doing exactly what Shuttleworth claims to avoid. Quoting Mark:
> we're not that interested in matching functionality
> that was in Win95, especially if we think that
> functionality will get dropped in Windows 8 or 9 or 10.
Unity (and Gnome3) are both busy copying obsolete UI concepts from OSX, which itself is stuck in the 80s. I would say that they'd be better off copying Win95 instead, at least Microsoft recognized (correctly) back then that a concept of an "app" makes no sense on a modern desktop, and windows is what users want to manage, not "apps".Mr. Shuttleworth admits to be heavily influenced by Android and iOS. That's weird to me, since desktops are ultimately very different beasts. The concept of an "app" came back from the grave only because the 1st generation of mobile devices were very weak, DOS-like if you will, at multi-tasking and IPC. So... Mr. Shuttleworth picked the wrong corpse to invigorate.
In their mindless pursuit of copying everything from OSX they didn't even bother recognizing OSX own bugs like broken virtual desktops. See, Apple themselves failed to copy them properly from Unix UIs (windows from different desktops in OSX are mistakenly combined in the same useless giant Alt+tab list). But the thing about Apple, though, is that they can afford to have an obsolete window management circa 1985 because their windows are shiny and fly around without lag, while Unity just recently learned how to scroll text smoothly.
In the end, Mr. Shuttleworth really shouldn't mention Win95 at all, which now seems to emerge as a clearly superior desktop UX of all 3, nearly 20 years after its introduction.
It appears as though nowadays it's almost considered to be shameful if computer don't try to imitate mobile interfaces, even though it feels as if the apex of usability was reached not that long ago with usable and slick Gnome 2 releases.
See, "apps" made sense in the prehistoric times when single-threaded desktops with limited RAM were primarily used to do one thing at a time. But things changed. The concept of an "app" have not been making sense for a long time: you can run Gmail and Facebook in two browser windows (which together live in the same process) or you can have a single window (Chrome) launching one process per tab, or you can have a terminal with several processes running in different windows, or you can have a copy of Microsoft Word embedded inside of an Excel spreadsheet. Similarly you can have an iframe with embedded Google Maps inside of your Wordpress blog. So what's an "app" then?
Microsoft correctly recognized in early 90s that the concept of a "program" or an "app" was obsolete, and modern UI should be using "window", "document" or "folder". An app is just an implementation detail: you can have 3 apps involved in rendering a single document or drawing the same window. Early smartphones couldn't handle these tricks (like multi-tasking) well, so this temporarily reincarnated idea of the "app" came back from the grave, to be happily copied by Unity just because "iOS has them".
Boo.
I think in terms of apps. "I need to switch back to my browser". "Alt-tab". "I need to change windows in this app", "Alt-`". I guess owning a Mac got me prepared for it, but I love this new behavior in Gnome3.
> You say this, as every major platform migrates towards apps.
That's the way it goes with these damn computers. In circles. I recommend you unpack a bag of floppies, get yourself some Windows 3.1 and take a good look at Program Manager.It goes in circles because computer technologies are bursty. With Win95, managing by window was reasonable because having >10 windows would usually crash the box. That's also why every window was a taskbar entry, as there was actually room for all of them... By WinXP, when virtual memory became more robust and RAM cheaper, you needed to sometimes collapse taskbar buttons by app to make them fit... And now in Win7, it's the default with just big icons. People leave way too many windows open now because closing them is a chore, and people are crappy/lazy memory budgeters. Apple realized this with iOS, and ported some relevant concepts back to Lion--apps should be able to close and restore at any time, and their open-ness is irrelevant to the user. So yeah, that's why we're back to thinking about apps; windows may be what you most often interact with, but they are no longer the mental unit of desktop management. It's now all about task-->app-->windows.
As an aside, let's consider how effed up MDI is, the inconsistencies in window/toolbar management between Win Office apps, and how the Mac "one window per document, palettes appear on app activation" scheme has managed to stay remarkably consistent for over a decade.
Stick them on different virtual-desktops and they will not show up in the Alt-tab.
I find virtual-desktops a great way to organize, that has been available for 20 (?) years
right click task bar > properties > taskbar buttons: never combine
...to switch back to a window centric approach, and this combined with the ability to use virtual desktops (even on Windows with a 3rd party app) with the taskbar only showing windows from that desktop, and going even further if you like, using a multi display setup where each display has a task bar showing only the windows from that display (like having "# of virtual desktops" * "# of display" actual virtual desktops) (note: I HATE this but know people that swear by it), you can easily manage even ~50 windows without becoming visually or mentally crowded, with no need for the app-centrick view that makes you click around all the time......the point being that unity should try and learn more from the nice and clean but yet easily costumizable interface of Windows 7
Disclaimer: I'm mainly a Windows user, touching a Mac from time to time, and using Xfce when on *nix because it just works and doesn't stand in my way...
There's no one right way, and governing by pure democracy simply won't work, because everyone has a different opinion.
And sure, you can make it configurable, but that's not the point - is it? If you're a power user, you'll find a way to manipulate the system to your will no matter what. But 99.9% of the users (well, when dealing with Linux admittedly that percentage is much smaller) will NEVER CHANGE THE DEFAULT.
So: apps vs windows? Side launcher vs bottom launcher? Doesn't matter. Just choose one, and for the love of all things holy, be consistent.
>windows from different desktops in OSX are mistakenly combined in the same useless giant Alt+tab list
Completely agree. In particular, I used Gnome (2.14 through 2.18) for a few years, then OS X (Leopard through Snow Leopard) for a few years, and the way "tasks", applications, windows and workspaces are managed on Gnome strikes me as objectively better than the way they work in OS X -- with the possible exception that the OS X might be a little easier to learn for users that have very little patience for learning much at all about their computer.
But OS X is better at sparing the user from tedious admin hassles and has better aesthetics. Gnome 2 on the other hand allows the user to set the text size of almost all of the text on the system (the most notable exception being Firefox, which does not pay any attention to Gnome's settings), which OS X does not allow. (Some of the text on OS X, e.g., in the menu bar, window title bars and certain Apple applications like Calendar, cannot be changed with reasonable effort.)
Also, the "annoying glitches" on Gnome are worse. The annoying glitches on OS X consist mainly of the UI becoming unresponsive at unpredictable times (usually signified by the spinning beach ball cursor) whereas the annoying glitches on Gnome 2 consist of inattention to detail on the part of Gnome 2's designers and implementors, lack of consistency including "gotchas" like Ctrl-W having minor mostly-reversible effects in Emacs and in bash while having drastic irreversible effects in other Gnome apps (namely, closing the window). Also, Gnome would freeze, crash or otherwise need to be restarted an order of magnitude more than OS X does. Note though that Gnome and the graphical system on Linux might be much more reliable with a different video card: I only tried it with two old cards.
This is THE reason I will not touch Unity or gnome shell.
I disagree; I think this is a subjective thing based on one's usage habits. I tend to think in terms of apps, with multiple windows attached to each. (This is partly because I keep a lot of windows open, and tabbing between 5-10 apps is more reasonable than tabbing through 30-50 windows.)
I agree with your main point, however; Ubuntu should be blazing its own trail, not mimicking the past.
And if you're talking about conditioning, perhaps you should considering not using Alt-Tab, a task switcher that has never really worked properly, because it cannot be both memorable, linear and context-adaptive.
Those shiny OS X windows that fly around without lag allow me to task switch visually, using touch gestures. It works great and uses all parts of my brain. I recognize windows by their shape, proportions and contents, not by a single icon. Swipe up, see all windows grouped by type, swipe up, focus on all windows of the same type. Swipe left/right, switch desktops.
Except for the Finder in its default configuration. (Opening a folder does not open a new window unless the user has changed the Finder using Finder > Preferences > General.)
But more importantly, the Dock is app-centric whereas the task bar in Gnome 2 and in Windows 95 through Window XP is window-centric. I can live with the Dock, but I'd rather have a task bar or more precisely a task bar plus a one-click way to open heavily-used apps such as is provided by both the Gnome Panel and the "panel" or "bar" at the bottom of Windows 95 through Windows XP.
I can't live with Apple's Expose as the normal way to switch between windows. (And Spotlight is too slow on my 2006 Macbook as a means to switch to or start apps). To switch between windows, I use OS X's facilities for app switching, and rely on every app having only one or two windows open. To switch between apps, I use the dock or I use a function key -- four of my function keys' having been bound (using Spark) to my most-heavily-used apps, namely Emacs, browser, IM client and Finder. I have not tried OS X Lion yet.
I definitely can't live with Spaces after having used the simple and sane workspaces of Gnome 2.
So I agree with old-gregg that these are areas where OS X is weak compared to Gnome 2 on Linux. (I prefer OS X overall though at least on laptops.)
ADDED. When I ran Gnome 2, it was no trouble to have ten terminal windows open. Since I started running OS X, I have worked around OS X's less capable window-management facilities by upgrading my Emacs skills so that instead of using terminal windows, I now use Emacs buffers to interact with shells and other "text-mode command-line" programs. (There's no problem having many Emacs buffers open at the same time, and I only ever have one Emacs window (or "frame" in Emacs terminology) open at a time.)
ADDED. Moreover, when I ran Gnome 2, it was no trouble to have 3 or 4 or more Firefox windows open. One thing that makes OS X a little more comfortable for me is that when I switched from Firefox to Chrome, I stopped having dozens of tabs open with the result that I am able to manage most of the time with only one browser window open. (Fewer tabs are required in Chrome because Chrome does not have Firefox's bug in which invoking the back button or the history menu frequently takes many seconds to complete.) In other words, I got by in OS X by rearranging my life so that I did not run any apps with a lot of windows open.
ADDED. None of these difference between Gnome 2 and OS X is a huge deal, but it does shed light on the quality of decision making at Canonical these days to note that they decided to transition from the Gnome 2 way to the OS X way in one of the few areas in which Gnome 2 is better. (Hmm, if Canonical switches to the OS X way of doing cut and paste, in which you have to press Command-C or Control-C to get the selection into the clipboard, that would be further evidence that they do not know what they are doing.)
I do not disagree with you (long live OpenDoc ;-)), but as a counterpoint: users buy _apps_, not _windows_.
As long as that is the case, apps do make sense on a desktop.
The problem is Shuttlworth can't keep his mouth shut. Personal attacks only make the situation worse. He should take a page from Apple here - "keep your mouth shut" - and produce.
And if you can show me a single personal attack from Shuttleworth here, I'd be very interested. I don't always agree with him but in my experience he's always been extremely professional. Claiming anything else is incredibly unfair and mean-spirited.
For the lazy, every comment on the linked bug by him (with an aside that 9 comments on this bug hardly seems to be ignoring the complaint):
https://bugs.launchpad.net/unity/+bug/882274/comments/9
https://bugs.launchpad.net/unity/+bug/882274/comments/19
https://bugs.launchpad.net/unity/+bug/882274/comments/20
https://bugs.launchpad.net/unity/+bug/882274/comments/25
https://bugs.launchpad.net/unity/+bug/882274/comments/26
https://bugs.launchpad.net/unity/+bug/882274/comments/27
https://bugs.launchpad.net/unity/+bug/882274/comments/29
After following open source projects for some years, it seems they're all eventually accused of some dark corruption - usually after making some UI change.
"You're betraying the ideals of the [project / community / founder]," the accusers say. "It used to be about giving users [freedom / choice / slightly less RAM usage]. Now it's just about [main leader of project / 'the developers' / the corporations]'s whims! Why not just give us [a setting / the old behavior / a large-scale feature]!"
Take a look at Firefox forums after every release, or Pidgin's bug tracker every time they tweak the GUI...
It reminds me of an older Less Wrong article - (http://lesswrong.com/lw/uu/why_does_power_corrupt/).
Now, it could be that all projects are slowly decaying into [ego / corporate / dictatorial designer]-centric tarpits, but I kind of doubt it.
Firefox is a good example. I have used it since "Phoenix 0.1" came out. Already during "Firebird" times, users have started to complain that it is getting slower and bigger. Mozilla ignored it. It needed Google's Chrome browser to ring the bells at Mozilla and show them that the user complaints were valid after all.
Compare that with well managed projects like the Linux kernel. There its usually the BDFL who does the wake up calls, after having listened to user complaints and suggestions. Not saying that everything is perfect, but the kernel has been hugely successful for 25+ years and is still well focused.
Ubuntu OTOH has completely lost its original focus of providing an easy desktop Linux. They started pushing server versions and now suddenly smartphones. As if the desktop would disappear in the next years.
Now someone with more money does a shitty job to make the UI like he likes, and don't have the decency of adding Conf parameters because it will have to suffice for everyone. After all, it's good enough for Him.
He copied the anchored menu bar, moved the window controls to the left, copied exposé, copied some elements of the OSX panel, tried to improve typography, etc. Some of these were actually good developments (the Ubuntu branded font set is nice) but most of the time it just feels like a cobbled together interface.
It also seems he is trying to copy some other elements of Apple. He is eschewing customization (ostensibly) for uniformity and simplicity. And he has made Canonical's designers the top decision makers. Programmers implement. Community gets what they're fed. Usability is their ultimate trump card (as we see in this case).
This doesn't appear to be working well, though. Apple hired world class designers and payed them tons and gave them a massive amount of power. And they had Steve Jobs who would say "this sucks" when no one else would. Canonical (probably?) has pretty good designers given near-absolute power. I'm thinking there's quite a difference between "world-class" and "pretty good". Not to mention that Shuttleworth, an amazing man in his own right, is no Steve Jobs. He probably doesn't say "this sucks" enough and maybe even doesn't know what actually does suck when it comes to design and usability.
Anyways... I wish Shuttleworth well but I'm not very optimistic about Canonical's future. And I think they'd do better if he brought his own personality, creativity, and intensity to the project instead of chasing Jobs' shadow. Or maybe he should find a new leader to take his place.
Well, it's not even a good distro at work. We're using the last two version of ubuntu, because everytime we upgrade, it breaks.
I used archlinux and I liked what it does. I expect archlinux to break and I expect archlinux to be up-to-date. Even when archlinux breaks, it's no biggie because I backed it up and it's easy to reinstall archlinux.
Ubuntu? What do I expect? Stability and up-to-date programs. Major upgrade does not implies major breakage. Now, they want me to expect Mac OSX like experience? They are now even more unsuitable for work than archlinux is.
I have really come to like the AUR and consider it one of the strengths of Arch, despite the AUR's shortcomings. Doing recent work on Ubuntu I was really missing my AUR packages and the ease of contributing to the AUR.
IN my personally opinion ubuntu completely destroyed linux Because rather then there being many great distros to choose from and play with there is now really only one. In fact OpenSuse was my favorite but that became just an also ran distro that people would compare to ubuntu and but say sorry but it doesnt compare and then stick with ubuntu. Which sucks. Of course OpenSuse also had the kde problem but thats a whole other topic of discussion.
There are lots of excellent distros. The world does not revolve around Ubuntu.
I really don't understand this comment.
There are very many other distributions, covering a wide range of needs.
- Fedora, Mageia, Sabyon, and Mint Debian : nice desktops, easy to use - Arch, Gentoo, LFS : suitable for power users and tinkering - Slitaz, TinyCore, Puppy : minimal, configurable
etc etc.
None of these are based on Ubuntu.
I do not mean to say design and UX and usability is not important, but still, it is a matter of priorities. If a designer decide to have transparent flapping buttons and if this make the code behind it extremely complex and if it makes impossible to customize and configure manually any parts of the system, then I would say it does not belong to Linux.
Maybe it is a case with Ubuntu. Maybe Ubuntu is a Linux for designers. Then I am wrong, and will switch to another distro. For me Ubuntu was an easy to install Linux with few driver issues, and my first step after installing it is to remove the useless Visual Effects in preferences.
But authority only works if you use that authority to do the right thing. Unity isn't doing that. I am in part frustrated that Unity got _worse_ with the latest Ubuntu release, and I'm frustrated that it has lots of bugs, and the interface is neither traditional nor discoverable. But even worse than these problems is the problem that Unity isn't pursuing a worthy goal. It's a rethinking of how you manage applications and files on the Linux desktop, but the Linux desktop has never been particularly functional or filled out and it falls behind further with time. The major trends we see is the desktop spiraling in on a single application, the browser, and files and media moving to the cloud. Unity is messing around with something that will never win the hearts and minds of new users, while it alienates what users there already are. I've never met someone who loves Unity! There's a couple people who accept it, and a much larger group that hates it. (Personally once I figured out to install gnome-shell I can like Ubuntu again, but Unity really did almost push me over the edge to get a Mac.)
It's not to say that a rethinking of the Linux UI is an entirely useless endevour, but Ubuntu and Canonical should not be pursuing a fantasy of a rich ecosystem of GUI apps. There are maybe a dozen relevant applications on Linux now, and that number will only get smaller. But the beauty of this is that if you just pursue an experience that encompasses those dozen applications, and you dive deep into those applications (which you can because they are all open source!) then something really neat could come out of that. I think people could get behind an idea like that, even if those same people simultaneously felt annoyed with particular changes that came from that. But it's hard to look past those problems when it feels like present usability is being compromised for a misguided fantasy.
Far more likely they want to cut in on corporate IT, where saving $50/seat in OS licensing and $100/seat in hardware adds up to real money.
While it would definitely be nice to please everyone, even merely trying to do so will seriously derail you from your own vision and design. That direction can either be yours or nobody's.
People also overestimate the effect of this kind of soloing. At worst, they will just create something nobody will use.
And if the majority of users will cease to use Unity, there will be 3rd-party packages for Gnome3 or Gnome2 or whatever it is that people want. As a critical review, if Ubuntu ever get to that point it might indeed be a good idea for Ubuntu to accept defeat and return to the mainstream. However, it is no earlier than that when we will see if they truly are stupid. And if they are, then there's a fair chance of forking and there will not only be Kubuntu and Lubuntu and ${WHATEVER}buntu but also Gubuntu.
Some participants in open-source software development underestimate the costs and the disadvantages of openness, democracy, etc. They seem convinced that if everyone would just uphold those values, everything will turn out well, and they seem impervious to evidence to the contrary.
It might be the case that if Canonical's designers and leaders get to boss the programmers around, then not enough programmers will continue to contribute to Ubuntu. But that is an empirical question that might turn on considerations other than openness and meritocracy. Maybe many programmers will continue to volunteer to help Ubuntu even if they understand that programmers and users do not have an equal voice in the direction of the project.
I switched back to Ubuntu for 11.10, hoping some of the things which pushed me to Mint in the first place had been fixed, but I am again bitterly disappointed. Unfortunately, I value having a pretty interface, I find xfce ugly, and simply installing gnome 2 in 11.10 doesn't "just work".
I also tried Arch, but I found I was spending too much time getting it to work how I wanted, and I eventually gave up on it for the same reason I stopped using Gentoo.
The one thing I never considered was switching to Debian - which might be the best idea yet!
Actually, I do wish I had just stuck with Mint, but I dont have time to go back now, and If I am going to install another distro, then why not Debian...
I'll give XFCE a try. I don't want to leave Ubuntu, because I like the packages and I don't have to read a manifesto sized manual to install it or learn another package system.
Perhaps that's not who they're targeting.
That's not a good thing? I haven't used Unity yet, but might try it out just on that comment alone.
When I updated to 11.04 I tried Unity for about 3 hours, got annoyed by certain things hurting my work style (the all-workspaces Alt-Tab mentioned above being one), and logged out and back in via "Classic Ubuntu".
I felt no antipathy towards Ubuntu at this point, and I was impressed they were trying new things. They just weren't for me.
Then 11.10 came out, and immediately my machine starts bugging me to upgrade. Apparently when 11.04 came out Canonical had said that Classic wouldn't be available in 11.10, but I missed this.
The upgrade process doesn't prompt you about that fact at all - it just takes the GNOME configuration that you were using, uninstalls it, and gives you only Unity.
After trying Unity again (<1 hour this time), I went to do the same old "Ubuntu Classic" switch and there was no option. That's when I got cranky. I had to Google to find out how to get back the desktop environment I'd had before I upgraded.
I would have felt good about the Ubuntu upgrade if any of the following had happened:
- Upgrade process had said "Please note the environment you are currently using (Ubuntu Classic) will not be available after you upgrade. You can install the gnome-panel package to re-enable it after you upgrade."
- Upgrade process had seen that I was already using Ubuntu Classic and didn't remove the gnome-panel package as part of the upgrade process, letting me keep the Ubuntu Classic after my system upgraded. How would that have hurt?
- (Best One) Upgrade process would have asked me "Would you like to try the new and improved Unity environment after you upgrade? If you change your mind later on, you can log back in as Ubuntu Classic" (alternatively, it could have said "install gnome-panel package and then log back in as Ubuntu Classic" if they couldn't bear to leave those packages installed.)
I agree that to some extent this is me whining like an indulgent nerd about an OS I got given for free, but it's also bad UX on the part of Ubuntu.
During a standard upgrade, your operating system shouldn't decide that the way you've been using it is wrong, and give you an unexpected and not immediately reversible transition to a different way of using it. That's hubris on the part of Canonical.
I genuinely find this attitude confusing.
I gave a presentation to a user group at a senior center last night on Ubuntu, during which one person made a similar argument. (The only real difference was that he used the word "arrogant" rather than "hubris".)
I don't know your background so I can't speak to it, but this person I spoke with had contributed no code, artwork, translations, etc. to the project. He received a product entirely gratis, which he got value from. But then the product moved forward and changed. He still had what he had been given freely, of course. But he demanded that the people doing this work for free either never make changes to their work, or if they do to continue to maintain the old system as well. This work should also be given to him without charge or without him having to put forth any effort.
And for those of us actually spending our lives building the thing for him to do otherwise was arrogant.
I'm literally speechless when confronted with these sorts of arguments.
I do see your point, I really do. My concern, and the part that I specifically labelled hubris - which in hindsight is a regrettable term - was neither the desire to make changes nor the desire to not maintain old configurations. Those things are understandable. When I tried Unity in 11.04 I was impressed that Canonical were working to advance the state of the art.
The part that irks me is when I try to imagine the assumptions behind the choices for the upgrade behaviour in 11.10. Users are proactively prompted to make that upgrade, and I perceive a disconnect between their expectations of a simple upgrade and the resulting user experience.
In the post above I mentioned some ways in which this could have been made much smoother (simply warning users that their entire desktop environment was about to change, for instance.)
Canonical doesn't have to answer a single thing to me for the operating system they worked hard to give me for free, but I still wonder why they chose that path.
I agree, though, that this distinction is not clear enough by reading Xubuntu (or Kubuntu, Edubuntu, etc) websites.
Me, I usually just install a very basic Debian and pile crap on top as I go :)
Ubuntu / Linux has far bigger problems than Unity. (Jump to point 5 for conclusion).
1. There are huge regressions.
With the upgrade to 11.10 my laptop's battery life has been cut from 5-6 hours to 2-2.5 hours. This is on an Asus UL machine that on Windows gets 10 hours.
This may be a kernel issue, but Canonical could add a lot of value by at least warning, or better, fixing/minimizing the problem. There are all kinds of boot parameter hacks, etc. to try to fix this, but it requires a lot of reboots and fiddling with internals. How about including a script to optimize these settings that is run after the upgrade?
"Linux on the desktop" is a dangerous misnomer, because on desktop computers power is not a concern. But nowadays power consumption is almost everything. Canonical already has its sights set on mobile devices, but hasn't even addressed power consumption for notebooks properly.
2. Poor communication of core functionality changes
New additions are presented and lauded in glossy detail, but removed features are not as clearly presented.
Nautilus (file manager) used to allow drag and drop copying/moving of files onto the left pane "bookmarks" folders. This was simply removed without warning, crippling quick drag and drop filing. Once again, this is "upstream", because of the Gnome/Nautilus teams' decisions, but it affects users in a noticeable way.
The initial update to Unity (11.04) removed all gnome applets, and made the time tracking software I used to use inaccessible. The upgrade to 11.10 made me lose my skype app/status indicator.
I don't care if my dock is on the left or bottom, or my open-close buttons are on the left or the right. Most of Unity's "coolness" is Compiz anyway - zoom, desktop switching, etc., so all that was available before Unity. For launching applications, gnome-do is still much faster, light-weight, and flexible (e.g. allowing creating/opening individual tomboy notes).
By communicating feature additions and subtractions better, people won't be so negatively surprised. Set proper expectations.
3. Key consumer software weaknesses
No good media creation suite. I was trying to make a photo slide show with a soundtrack a while ago, and I went through two days of installing/testing, setting up ppa's, compiling sources, etc. to get the latest versions, and nothing really works well. I mostly code, write, and use the web, so it's not mission-critical for me, but for a "consumer desktop OS", the absence of an official and well-functioning suite of applications akin to Apple's music and movie makers is a weakness.
Other apps like Evernote and a more up-to-date Skype would be nice as well.
4. What hardware does it run on flawlessly?
When you buy a computer with Windows or OS X, you know that everything will work. With Ubuntu/Linux, you don't. It's a huge stress factor before buying a new computer. Ubuntu has "certified hardware", but it's buried on some wiki page or other back page. This information should be front and center on the home page, so that I can buy something with confidence that it just works.
5. Conclusion
Ubuntu (whether Unity or Gnome) is far more usable than Windows (messy config menus, no multiple desktops, no full-screen desktop zoom, inconsistent shortcuts, etc.), and at least as good as OS X (which for example doesn't allow you to change the system's font size globally, and is less keyboard navigation friendly).
I'm considering abandoning Ubuntu again, because it can't compete on power consumption.
I am sympathetic to Canonical being annoyed by the bitchy entitlement complaints over superficial UI features (e.g., open/close buttons left or right). It's bike-shedding to the max.
That said, I think better up-front communication of changes can help set expectations. It forces the designers to reason why they are removing/changing/adding something. This doesn't have to lead to drawn out discussions, but some design decisions seem to be "shot from the hip" without realizing that they may affect/ruin thousands of people's work flow.
Lastly, there are huge issues like power consumption and predictable hardware compatibility that currently heavily weigh against Ubuntu/Linux.
And the interface... I don't mind spending a week or two learning an interface if it will allow me to get the job done, but The Gimp makes me want blood. Decent media software is the ONLY reason I keep a Virtual Box image of Windows 7 on my drive. Well, that and the occasional browser testing, but I find that's less of an issue in recent years.
As for hardware, I've run into Very few problems on a few DIY systems and laptops / netbooks. Graphics compatibility has always been an enormous issue. Especially multi-card for 3+ monitors. I assure you it's possible, as I've done it a couple times, but I tend to fall into a slight depression once the battle of getting it to work is over. Otherwise, I've been incredibly lucky. Install and go.
The battery issue is a major one. I've found that Lubuntu keeps up well on my netbook, and I rarely unplug my work notebook, so I don't run into it much, but you're spot on about the importance of getting that right, or at least making the correction simple.
Fine points, overall.
As for Gimp itself: I must admit that I love it and love it a lot more than Photoshop. I use it to post-process the pictures that I take (example: http://www.flickr.com/photos/valanx/sets/72157627905845341/ ) and while I have no need for anything fancy, I have yet to find something that beats the speed of the Gimp.
Yes, Alt+c, adjust, Alt+i, s - you do have to get down to your keyboard if you want speed, but once you're fluent with that stuff, you just breeze through it. Put differently: The Gimp gets out of my way and provides just the functionality that I need. I love that.
It could be a lack of support for a specific advanced power-saving feature at kernel level, but that wouldn't be my first guess..
That overbearing spouse deserves to be left in the dust immediately, and maybe some of these old users do as well. Not that they would be truly left behind, considering the vast landscape of other equally expensive options available requiring just as much effort to install - or MORE, if that's your thing.
We've been talking about bringing Linux to the desktop for more than a decade. And it's not even close. So a company comes along and says: Listenting to you guys isn't working, so we're going to try something else. And now they're somehow considered the antithesis to Linux. What is "Bringing Linux to the masses" supposed to mean? Making all computer users around the world cantakerous curmudgeons who can change every pixel any graphical interface can ever hope to offer? I'm pretty sure I wouldn't like that world. I like Linux geeks, and consider myself one, but I also like all the people in my life who will never even care what Linux is.
And maybe there is a way to the masses via the old ways. I implore you to invest your time and money to find out. I would honestly love to watch that unfold, and I'd probably install that version as well to see if it fits me. In the meantime, I'm going to be happily using Ubuntu everywhere, as I am now - on my 2 year old DIY desktop with 2 giant 27" monitors (worked upon first install, with no proprietary drivers - 3 screens: not so easy), on my television (DIY home theater running xbmc - hdmi audio and video - also worked upon first install - no proprietary), and on my laptop at work (thinkpad, also worked immediately), and of course, on a couple servers at various stages of deploy.
As unpopular as the opinion may be around here, I really enjoy using Unity. It's new and obviously imperfect, but it's one of the better user experiences I've found on Linux. I felt the old gnome 2 desktop was a step back from Windows 7 when I switched my media center to it last year. It looked as though I was installing software from 15 years in the past, trying to fit into a world that had an inch of dust on top of it.
I like the windows-7-like dock bar in Unity. I like that searching for apps via keyboard is king (although the sorting sucks, sorting by usage is essential). Not a huge fan of the omni-preset menu bar, but it's not really THAT big of a deal (admittedly having 2 giant screens makes that possible). I don't give a damn about screensavers, as I haven't used one since flying toasters were the in thing. Truth be told, I spend far too much making things ON Linux to care much about Linux itself. And that's the way it should be. At least that's the only way I can stand to use Linux on every single computer I own.
I remember Rob Pardo reflecting on criticism World of Warcraft received before initial launch. A lot of the Everquest players thought getting to max level was too easy and didn't require enough of a time sink. Rob pointed out that there was little value in focusing on players who had that much time to spend on a game, and they had to let those players go. It's tough, from a commercial and critical perspective, to say he was incorrect. WoW brought MMOs to the masses in a way nothing else has.
Letting go of users can be hard. Being the user that's getting let go of is harder.
But reading a lot of this what I see is "won't fix. Sorry, against our design" with no attempt to either leave the door open for future changes or figure out what the user wants. Better approaches include a long-term feature requests queue or some other way of preserving feedback for later review during design sessions, or a canned response which details why a category of feature requests is unwelcome and inviting further discussion of what could be done instead. If it is not applicable, the user ideally should be challenged and asked to justify why they need what they say they need before the bug is marked wont-fix with no discussion.
A canned FAQ with, as you say, details of why they're doing stuff would have helped. More importantly, links to instructions for removing Unity and installing other environments would have been good.
But that would have opened another problem: Installing Gnome 2.x is only going to be viable short term. Dependencies will start to break more and more stuff until users are forced to upgrade.
Really, Gnome should have forked and had a stable 2.x release with gentle bug fixing, and a renamed 3 release with fun new features.
"Figuring out what the user wants" isn't necessarily the answer either. Not if your goals are as lofty as bringing Linux to the masses. Un/fortunately, the current user base is NOT necessarily the future.
But if they want to bring "us" along, there are options. Drive the design-centric stake into the ground. Put their designers front and center with blog posts and marketing and workshops. Make it real. Let those who want the same-old to make the choice to go elsewhere - cordially.
And from there, teach the function-first world from which it comes why design is important, by leading through example. Open the curtains on this dark basement of a community to let some light in and start a long and important conversation about personal hygiene, kerning fonts and rounded corners. Help a couple app developers make some real money selling a simple yet well designed app through the new software center.
In my mind, that's real communication and leadership. Bickering over features won't get anyone anywhere, and while their hand-waving "won't-fix" is NOT the way to go about it, neither is wasting too much time on people who probably won't like what Ubuntu is trying to become.
Personally, I like the new stuff, and I'm looking forward to seeing what's next.
In general you can assume that what's not working for current users will not work for a large subset of future desired users.
Sure, that's a valid way to work. He should have started off by saying that, rather than building a large, vocal, community and then ignoring them.
Bitter arguments among side-stream groups (church-goers, linux-users) are a much bigger reason why these groups do not attract mainstream audiences then what it is they're fighting about.
Unity, Gnome, and KDE are all headed towards a more Mac-like UI because Apple is the most lauded software design company in the world.
Like church-goers, Linux users should be supporting and celebrating innovation in design that are meant to increase use in the software. If such innovation was more broadly celebrated for its own sake, then it is much more likely that one of the major players would be willing to risk new design decisions - to really experiment with the UI. Instead, the same changes are all adopted by the major players because trying something too different will get your ridiculed, and that leads to losing early adopters, and that leads to a diminishing user base.
If you want to see Linux become more mainstream so that a generation of computer users can have a little more software freedom, then join the discussion by celebrating experimentation instead of condemning it.
Secondly, is it worth wile to bring the mobile experience to the desktop?
My point was that this isn't going to happen if the community continues all of this back biting.
Shuttleworth is trying to gain some of this advantage for Ubuntu.
It might seem dictatorial, but I think he's got a point .. if Ubuntu is going to innovate and create differentiation from other OSs it needs to change. An element of faith is needed to allow that change to occur.
It might be enlightening to read the comments on the bug dispassionately and see where the name calling and posturing is coming from and where it is not.
However, the direction does not seem to line up with typical open source hacker goals (it's very hard to adjust unity even in the smallest ways, and there seems to be some number of breaks from typical conventions, e.g., breaking xscreensaver), which is one of the key values of the Linux world. That, I believe, is the controversy here.
I use Bodhi at home, and have been real happy with it.
Ubuntu has always been about targeting the mainstream as far as I can tell. They do a good job of it, and there's plenty of other options for those who think otherwise.
"MATE Desktop Environment, a non-intuitive and unattractive desktop for users, using traditional computing desktop metaphor."
Excellent! Just what I want!
I wonder how active this will be and whether it will see any improvements or just maintenance
OK, so which distribution is everyone switching over to in place of Ubuntu?
[1]I think this is the name of the Gnome 2 package. Personally, I couldn't stand Gnome 2 and even preferred Unity; I didn't jump into the world of the tiling WM until compiz started crashing my GPU.
[2] at http://awesome.naquadah.org/ ; I don't like the name, but I love the WM.
Can you show me the data that you've gathered on this? I'd love to see the user survey.
(Is it maybe possible that just because you don't like something, others still could?)
I am switching to Debian after 10 years of Ubuntu I was a big defender of Canonical and their SO and but now I lost it...
So they want Unity to be a GUI for tablets? How far up your ass does your head have to be, to think that a GUI designed for desktop use, and a GUI designed for tablet use, can be used interchangeably? Someone ask Mark, he'll know. Or someone who worked on Metro since they seem to have the same parts of their brain broken. If Microsoft does break into the tablet space, it will be from work on Windows Phone 7, it will not be from Metro. Apple recognized this when they used their phone GUI for tablets despite having a perfectly good desktop GUI around, and Canonical is all about copying Apple like a fucking funhouse mirror, how did they not pick up on this?
If they want to target the tablet space, good for them. I think the ship has probably sailed, but more competition there isn't going to hurt. But don't fuck traditional desktop users, who currently account for 100% of your user base, in the process.
Gnome 2.3 was a very, very good UI aimed at desktops. 3.0 is another kinda sorta tablet UI that only kinda sorta works on desktops (I am playing around with Gnome 3.0, XFCE, and a few others). The fundamental problem is that what works well with one in one input environment breaks badly in another. This was Apple's great contribution with the iPhone and the iPad.
Attempts to merge environments will get you with something that kinda sorta works everywhere and does not work anywhere well.
If you want Linux for tablets/phones just get android.
Basically, what I want out of a distro is a convenient platform to build the system I want. That's Window Maker with a lot of xterms, an Emacs session, a Firefox session, and possibly other applications as I need them.
Any distro can give me that much. The reason I pick one over the other is ease of administration. Ubuntu gives me this in the package management systems, the large repositories, hardware support, and, if all else fails, large, well-trafficked forums where questions actually get answered. (I'm leery of less-used distros because I know, I just know, that I'd stumble on some odd corner-case problem that nobody who's used that distro has seen and be up too late diagnosing and fixing it myself.)
So all these posts that imply Unity is the only way to use Ubuntu sound, to me, like people implying New York City is the only place to live in America. It's big and it's in all the movies, but there's a whole continent out there just beyond.
While it's nice that they try to improve the desktop, it seems like they forget that in the real world, Linux adoption happens with power users recommending (and supporting) new users with their Linux installation. Maybe Ubuntu can change that, but we're not there yet. And if Ubuntu doesn't cater to the power users _as well_, one day they'll wake up to another Linux distribution gaining popularity.
I've switched over to XFCE (Xubuntu) for now, but it seems like I might be better off going back to Debian. After all, I can join the Debian community and actually have a say if I disagree with its direction.
True. I used to suggest Ubuntu myself (actually, I was one of the people who started with Ubuntu, now I'm on Arch).
I've switched to either suggesting to install one of the forks of Ubuntu (mostly Kubuntu or Xubuntu), Linux Mint or even Crunchbang.
Unless it's used as a transition for a clueful, but new to Linux, user who's going to learn a bit and then move to a different distribution. (Probably Arch or Tiny Core, if you're suggesting CrunchBang.)
Exactly to those I would point to #!. By the FSM, I would never suggest #! to a user that doesn't already have a degree of technical knowledge.
A workaround is to use gnome-tweak to have nautilus not manage the desktop. I highly recommend it. As much as I don't like nautlius, I feel that it drawing the desktop is flat out archaic.