Welcome to the Post-WIMP Era

It is hard to say exactly when the WIMP paradigm came into being. The mouse and pointer (or ‘bug’ as it was originally called) came out of the work of Douglas Engelbart and his research staff at the Stanford Research Institute during the mid to late 1960s (his Mother of all Demonstrations is still amazing to watch). It was later, with Xerox Parc’s Alto and Xerox Star, and with the Apple Lisa, that the WIMP approach solidified. Which device got there first is somewhat irrelevant; between them, these early devices established the central features of what we now know as desktop environments.

Xerox Star (1981)

The ‘WIMP’ paradigm is made up of four key components: windows, icons, menus and a pointer. The approach was highly file-centric when it was conceived, and mimicked the physical world of the office. There were files and file systems, and a desktop containing a variety of objects, such as a wastebasket. (The Xerox Star desktop even had an inbox and outbox for mail.) Multiple files could be worked with simultaneously by having them on the screen at the same time, with windows forming the basis of early multitasking functionality.

Apple Lisa 3 (1984)

For decades the WIMP paradigm dominated the design of graphical user interfaces. It is remarkable how long it was followed for. Contemporary systems like Windows 7, OS X, GNOME and KDE still retain many of the key WIMP characteristics, including windows, desktops, file browsers, menu bars and pointers. Though there have been some minor design innovations, the basic recipe has remained the same for over thirty years, despite major advances computer hardware and UI design approaches.

GNOME 2.32 (2010)

As computer software and hardware have developed, problems with the WIMP model were discovered. It did not make efficient use of screen real estate. Objects on the desktop got lost underneath windows. Large numbers of menus encouraged GUI complexity and made it difficult for people to find actions when they needed them. The file system proved to be a labour intensive and error-prone method of storing and retrieving data. Windows themselves produced additional management tasks for users, and pointer-based input was found to be inappropriate for mobile devices.

So it is unsurprising that, dominant though it was, the era of the WIMP model is coming to an end. Some parts are inevitably being retained, but the basic formula of windows, icons, menus and pointers is being replaced. Maximized windows have become the norm, largely replacing the simultaneous display of several windows. Menus in the form of menu bars are increasingly uncommon. Pointers and pointing devices (such as mice and touchpads) are finding themselves coexisting with touch interfaces. File management is (quite rightly) being replaced. The desktop is becoming a thing of the past.

There are a number of reasons for the paradigm shift that we are witnessing in the field of GUI design. One of these has been the arrival of touchscreen mobile phones and later tablets. This new class of computing device required a different approach to GUI design, and forced designers to create alternatives to the WIMP model. Free of its constraints, they found the freedom to develop new approaches.

Apple iPhone 1 (2007)

Apple iPad 1 (2010)

Motorola Xoom (2011)

Mobile, touch-based devices like smart phones and tablets have been hugely successful, providing functionality that people value and experiences that they love. This has not merely been a consequence of their form factors or use of touch input. The design of the software found on these devices has been a major contributing factor to their popularity. Mobile, touch-based applications are typically better designed than so-called desktop software. They have better focus, are easier to use and deliver superior user experiences,

OS X 10.7 'Lion' (2011)

Given the success of mobile, touch devices, it is not surprising that they have started to influence the more traditional WIMP environments. OS X is the prime example here. The latest update to Apple’s ‘desktop’ operating system included numerous changes that were clearly influenced by mobile, touch-orientated design, including Launchpad (a facility for launching applications), overlaid scrollbars and fullscreen applications. OS X has demonstrated how design approaches from the touch world can improve that of the pointer world.

WebOS Laptop (Unreleased)

WebOS provides another example of mobile, touch orientated design being used on traditional WIMP devices. Though we are yet to see WebOS laptops or desktops in the shops, its designers and developers clearly think that it offers advantages for these kinds of machines, as well as the tablets and mobile phones on which it has already appeared. Now that the code is going open source, I’m sure we will see WebOS being used on laptop and desktop devices in the future. It shows how an operating system that was designed for touch input has much to offer machines that rely on pointing devices for input.

Windows 8 (ETA 2012)

Finally, the release of Windows 8 next year will be a major milestone in the move away from WIMP interfaces. The new Metro interface it will introduce is a variant of that found on Windows Phone 7, and will mark the end of the WIMP approach as the dominant interface paradigm for Windows. Windows 8, it should be emphasised, will not just be targeting tablets. It will also be found on laptops and desktops.

It is clear that the WIMP paradigm has started to give way to new and improved design approaches that have emerged in the mobile space, therefore. However, there is another process that is challenging the dominance of the WIMP paradigm: the increasingly blurred distinction between touch and non-touch devices.

Touch input is coming to a range of devices that have been the traditional habitat of WIMP environments: there are already many models of netbooks that have touchscreens, for example, and standalone monitors with touch input capabilities are increasingly common. At the same time, devices that function both as a tablet and a laptop (such as the Dell Duo, pictured below) have been available for some time, and tablet devices are themselves growing additional input capacities through keyboards, mice and touchpads, which sometimes feature as part of docking stations and cases.

iPad With Case and Keyboard (2011)

ASUS Eee Pad Transformer (2011)

Samsung NB30 Touchscreen Netbook (2010)

Dell Due Netbook (2011)

Acer T230H Touchscreen Monitor (2010)

These new types of devices that are emerging require operating systems that can handle both pointing devices and touch screens. An OS that can be used with only touch or only a pointing device will not be appropriate for them. Importantly, these new devices will become increasingly common after the release of Windows 8. Almost every PC sold with Windows 8 preinstalled will have a touchscreen. Many of these devices will also have a keyboard and pointing device.

Where does all this leave GNOME? GNOME 2 was firmly in the WIMP camp. With GNOME 3, we improved on that by starting to move away from the classic WIMP approach. There is no ‘desktop’ in GNOME 3, for example. The new GNOME 3 applications won’t typically behave like normal windows, either. They will be maximized by default and won’t have titlebars when they are maximized. A lot of the time you will not actually be able to tell that there is a window there at all.

Work to make GNOME 3 touch compatible is ongoing. Current design work is focusing on a number of areas, including scrolling. We are also developing touch compatible approaches to application design, so that new GNOME 3 applications will be effective with touchscreens. Here you will find click targets that are the right size for fingers, and we are making use of drag actions for key functionality. We are also planning ahead to make use of multitouch capabilities once they become available.

But touch does not come at the expense of pointing devices. GNOME 3 will remain a highly effective environment for pointer and keyboard input, which we will continue to optimise for. Many of the features of touch design, such as the use of drag actions and kinetic scrolling will be effective and enjoyable when using mouse input. Multitouch can also be used to positive effect with both touchpads and touchscreens. As this work progresses, those of us who contribute to GNOME design are finding that it is possible to create coherent designs that are effective using a variety of input devices.

tl;dr version

The time of traditional desktop GUI design is over and a new era is beginning. This offers the opportunity to make software that is better than what we had before. Touch input will play a major part in this new era, but it will exist alongside pointing devices and physical keyboards. Touch capabilities are already coming to laptops and desktops, and almost every new Windows 8 PC will feature a touchscreen. GNOME needs to be ready so that it can be used on these new devices. Thankfully, we’re already on the path to create great new user experiences that work with a variety of input devices.

This entry was posted in gnome, gnome-ux. Bookmark the permalink.

49 Responses to Welcome to the Post-WIMP Era

  1. ding dong, the WIMP is dead 🙂

  2. Vincent says:

    Yeah, try doing any real work that involves multiple applications. WIMP is here to stay, whether you like it or not.

    • Allan says:

      Hey Vincent,

      That really isn’t the case. First of all, you can do ‘real work’ using a post-WIMP UI. You might need a pointing device for certain kinds of activities, but there’s nothing about the UI design that will hold you back. Application design is yet to catch up in some cases (and we do have a way to go there with GNOME 3), but when it does I think you’ll be more productive than you were before.

      Second, this really is happening and you can’t stop it. Windows 8 is coming. There will be Metro style versions of Windows productivity software, including Microsoft Office. We’re probably going to see Android being used in more and more ‘production’ settings, too. That really only leaves OS X, and while that has a healthy niche for now, even it is slowly moving in a post-WIMP direction.

      Maybe we just have to agree to disagree though. 😉

      • robert says:

        Have you tried to chat with a coworker while both people need to see for example a document on an iOS or Android device? It is intolerable, switching between full screen apps back and forth. The W in WIMP is still needed to not waste time when multitasking. Probably not the full customizable experience we have now, resizable/stackable Windows, but a simpler WM, that does not means the metafor changed. they will still be Windows

        You can call these new interfaces not WIMP, but they are WIMP. stacked fullscreen windows, menus, icons everywhere and a desktop that have application launcher (Windows 95 style where every application installer droped an icon on the desktop). The only change I see is that the filesystem is hidden to the normal user, and that is a good thing

      • Allan says:


        So what you need is some way to chat without the need to change windows? Oh, wait, we have that! 😛

    • john says:

      WIMP gets in the way of doing real work. You have to force your workflow to conform to it’s tedious metaphors that only made sense in the context of accommodating people that were used to using paper.

      • Kirk M says:

        I can assure you that WIMP has been getting a tremendous amount of work done since it was first readily available to businesses and the general public (and it still is). And that covers nearly 3 decades. And by the way, paper, home and office, is still in wide use. I really don’t believe that the WIMP, task centric UI is going to be replaced by what makes up a touch screen, application centric interface. Rather, the world of personal computers (and that includes everything from desktops to smart phones, all makes and models) will evolve to include both types. Which one you’ll use depends on the task aat hand.

  3. Michael Jonker says:

    IMO there is an increasing majority of users and devices firmly focussed on consumption of content, the majority of this majority consuming entertainment and communicating.

    This is super-evolving a bubble economy of faster, slicker, simpler devices (which are cool – don’t get me wrong).

    I just wish that before throwing ‘WIMP’ out with the bathwater there is a content production paradigm that can be summed up in four or less letters to replace it.

    (For and on behalf of the elves who work in the background)

    • Allan says:

      Hi Michael (and the other elves 😉 ),

      Touch devices often get characterised in terms of content consumption rather than production. I personally think that’s a mischaracterisation. Though you do get consumption apps for tablets, you do find production ones too. Just look at how many really awesome word processing apps there are for the iPad, for example. There are also some amazing possibilities for music and image production.

      I’m actually pretty excited about what this will mean for production-focused apps. They’ve tended to suffer from some of the worst UX design around. It’s time to shake things up and do something better.

      • Michael Jonker says:

        I know what you mean, and all this is true under a certain paradigm.

        I understand how nice it is to move around a good production app, and also how banal the world becomes when production apps do too much, cater for the masses and template the visual identity of a generation.

        When I work, I rapidly and seamlessly move between code, production apps and consumption container. I prefer many things that do one thing well instead of one thing that tries to do everything. All I need is the most efficient way possible to switch between my activities (with the maximum screen real estate for my activity).

        I predict that as the human hourde loses the art of handwriting, it will gain the literacy of coding in a generation or two. By that time I hope that FOSS (and Gnome) will have evolved into something unique and let let the propriatory fellows entertain the mass market.

      • For reference, here’s an example of a powerful photo post-processing app for the iPad that sports a delicious touch- and gesture-based UI yet it still feature-rich: Snapseed -> http://www.snapseed.com/

  4. Mouse and keyboards are great for work, you use them on a passive position your arms are in rest position and takes really minimum physical effort.
    The touch world is great fun but don’t expect it to “fix” the desktop paradigm specially if there is no fixing needed to start with….
    I find that the driving force most of this “revolutions” is corporations needing to sell us new silicon toys. That make no mistake I love. But…. The mouse is way more convenient for several hours of work than pointing your finger to a screen in vertical position. the level o precision of a mouse arrow is several times grater than of a finger.
    Typing in a keyboard that is virtual and has artificial light produces strain and uncomfort after several hours of use…(its not natural to humans to have light coming from the ground).

    Windows 8 metro style is only optional, on the desktop. And i guess aimed at devices that can be tablets and laptops at the same time…
    Don’t fix what it is not broken.

    P.S. Note i think there are many ways to improve the experience and range interaction methods in the desktop. We abuse the need for specific action point of entry locations, and we could use other way less physical intensive interaction methods to communicate with it…. BUT what I’m seeing so far is not better any way.

    Nuno Pinheiro KDE-Oxygen Designer.

    • Allan says:

      Hey Nuno,

      I think you have missed my argument, so let me be clear: I am not saying that touchscreens and tablets will replace mice and keyboards. What I am saying is:

      1. Computing devices are often going to have touch and a keyboard and a touchpad or mouse

      2. We need to design for all these input devices, not just one or two of them

      I would also say that there is a lot that is broken about the desktop, but that is maybe a different discussion. 😉

      • Ok 😉 Maybe i jumped into conclusions to soon…Making our current desktops more usable wen different interactions methods are available is a great idea.
        I’m just concerned that we are destroying what works, to make way to what may not work… I sure wont be spending 8 ours of my day pointing at my vertical 27 inch screen.

        Making my non existing foldable laptop work on tablet mode for touch on screen interaction, great 🙂

      • Fri13 says:

        Yes, we need to design GUI for those devices…. and what we have learn from usability standpoint… it is smart not to do a single GUI what tries to fit all those, but to do separated GUI’s for all those different devices.

        That means,
        A GUI for desktop use
        B GUI for tablet use
        C GUI for smartphones

        They can overlap littlebit by having same technologies, but the GUI needs to be designed for them separately. Not by trying to bring a single GUI to all of them.

  5. In all of this there are no improvements/solutions for work that frequently crosses application-boundaries. Same for any complex applications. 3D modeling and animation, CAD, music production … even the quite common word processing tends to lead to a number of features that cannot be handled with approaches taken from small handhelds.

    And if you do use a large monitor, a system optimized for one maximized window at a time turns from a solution into a problem.

    • Allan says:

      Hey Thorsten!

      My view is that we shouldn’t automatically maximise windows on large displays – I think we can be intelligent there.

      You’re right, there might well be some apps that are only used with specific pointing devices. The difference, I think, is that they will run on a touch friendly environment and will exist alongside touch friendly apps. And just because an app isn’t touch friendly does not mean that it can’t incorporate design patterns that are also found in touch friendly contexts. Touch isn’t hostile to non-touch, in other words.

      • Alvaro Romero says:

        If you “force” the user to use the “touchscreen approach”, you are hostile to the classic keyb+mouse combo. There are many differences between the two models, and different use cases.

        Most workstations have keyboards and mice – will remain true at least for one decade or two -, and tablets and mobile devices aren’t suitable for many tasks – programming, accounting, CAD design, …, meanwhile mobile devices are better for some kind of jobs like goods delivering, restaurants, sellers, and other jobs where users don’t need to do multitasking and complex information input.

        I think that KDE approach is better, with separated Desktop and Mobile enviroments with a common technology behind. I don’t like the “Jack of all Trades” approach; you don’t get the best performance because you have to do some compromises to mix the touchscreen approach and the classic approach.

    • Simon says:

      Yes, that last point is important. On a 9″ device where screen space is scarce, maximizing windows makes sense, which is why Ubuntu’s old Netbook Edition was great. But, on a 24″ screen, not so much – running a web browser full-screen wastes a huge amount of space, that could be showing something more useful.

      • Meower68 says:

        I really get tired of this notion of an app needing to have the entire screen, even on smaller displays.

        How often have I wanted to refer to information on one screen while inputting information into another? On my smartphone? Weekly, if not daily. It would really nice if I could pull up at least two different apps simultaneously. I don’t need a full desktop with overlapping windows and the like; just give me the ability to split the screen between two or more apps.

        I don’t care how fast your device/OS will switch between full-screen applications. If I’m needing to refer, repeatedly, to the content on one screen while interacting with another, switching back and forth is a non-starter.

        It’s not just for 24″ screens. It’s for phones. It’s for tablets. It’s for any device upon which you need to do real, honest-to-goodness work.

        The ease with which modern GNOME will let you have an app take over the screen, or split the screen horizontally/vertically between two apps means that SOMEONE at GNOME gets this.

        Until such time as phones and tablets can do this (yes, Android, I’m lookin’ at you), they will be used for leisure stuff and very light amounts of actual work. They will NEVER be an option for replacing the desktop/laptop machines of today.

        I want to eliminate the need for desktop machines. Contrary to popular belief, keyboard + mouse is NOT that efficient. At least, that is, as long as it takes two hands to work the keyboard and one hand for the mouse. Until we can evolve tails with fine motor control to run the mouse (Zimbu the monkey on Dilbert, for those who don’t get the reference) or make one-handed keyboards more prevalent (trying to find one which doesn’t cost a fortune AND doesn’t suck), a touch interface has the ability to become MORE efficient than keyboard + mouse.

        We need the ability to:
        * point to pick a spot for the cursor
        * pull up a transparent overlay with the on-screen keyboard on it
        * do our text input
        * quickly switch back to pointing mode

        If the keyboard is full-screen, we have plenty of area to work with. While I can’t touch type on a tablet (no tactile/haptic feedback), I can input text quickly on a half-screen keyboard on a 10″ tablet. If I had a full-screen area to handle the keyboard and the text prediction functionality, I could really fly. I have to use the phone keypad with text prediction on my smartphone to get anywhere near that kind of speed (I have no fear of writing long messages, such as this, with my smartphone). A full keyboard, with smaller buttons, is much slower. Larger buttons are simply more forgiving of fat fingers, and forgiving = faster.

        If the keyboard is transparent, we never lose sight of the app with which we’re working. This is critical Switching, visually, between two apps requires our mind to do a task switch, as well, and that’s part of what’s wrong with modern smartphone apps requiring the whole screen.

        Having a button on the screen which switches, extremely quickly, between pointing and working with a transparent keyboard would allow you to do some major text input on a phone or tablet.

        On my WinXP machine at work, I have multiple desktops set up (been using VirtuaWin since Win2k; free and gets the job done). Java programming happens on one desktop. Web browsing, including testing what I’ve written, happens on another. Email happens on another. Database querying happens on another. LDAP querying happens on yet another. I have 9 workspaces to choose from, but only a couple windows open on each desktop. In short, even though I have a full overlapping-window WIMP interface to work with, I’m already migrating to the task-oriented approach. Been doing that for years.

        Eclipse, my main Java coding environment, has all kind of little sub-windows to it. Any one of them can completely take over the screen, with a quick double-click, or they can go back to showing me a complicated window with lots of different things, simultaneously. Again, sometimes you simply need to be able to split the screen into multiple windows.

        When we can do this, elegantly, with tablets and smartphones, they will be used for much more than media consumption. They will be used for work. Until then, they are fancy, expensive, somewhat useful toys.

  6. Corporations try going different directions. In all actuality, consumers decide the future of technology. For example, the way the keyboard feels makes it easy to use once you learn it over a few years. It’s what I’m used to. Touch screen keyboards, on the other hand, make it hard for me to type because I can’t feel it. Worse, for the blind it’s got to be impossible (as is a pointer).

    The good thing about touch keyboards is they get out of the way and come back when needed (at least they’re supposed to). The keyboard can turn from a US qwerty structure to a Chinese structure without having to unplug one keyboard and plug in another.

    If there were some way to make the flat screen feel or transform 3 dimensional, that would be a huge leap for the user interface. Touch sounds great, but it loses out on the fact you still can’t “feel” the software implementation on hardware. In my opinion, until you have a device (A) that can emulate every other kind of device (B), you’re never going to completely rid of device B. We still have controllers with buttons, dials, and many other user interfaces that touch simply cannot replace.

    I know this isn’t about touch “replacing” the WIMP, but there’s so much room for improvement, and I think touch can play a part in replacing WIMP interfaces. We need to apply to all the senses, so far we only use 3. Site, sound, and barely feel. Feel needs to be improved to be much more arbitrary, and eventually someone needs to introduce arbitrary scent and taste to hardware/software.

    • Calum says:

      Have you ever watched a blind person use and type on an iPad or an iPhone? With the right assistive technologies, which Apple have built right in, they become stunningly proficient very quickly.

    • Fri13 says:

      “Corporations try going different directions. In all actuality, consumers decide the future of technology.”

      That is naive argument. As consumers are not in the control, but the corporations are.
      Consumers could be, if all would be compatible and everyone could choose from all alternativies, what would mean, there is no space for competition but only for teamwork and alternative development.

      Example. When Microsoft comes and rolls out Metro for Windows desktop. It causes that all OEM’s start pushing PC’s with newest Windows. Every new OEM PC what consumer buys, comes with Windows preinstalled. They can not return software system and keep hardware system, but they need to return whole package. Here Microsoft has a domminant market position.

      Then Microsoft gets consumers to re-learn everything.
      Then corporations, what are slow to adapt, will starts rolling a upgrades (new computers usually) in few years, and that area is as well saturated. Basicly by the demand of the workers who are learn to use newer Windows at home.

      Then at that point, WP starts selling very well, if ever. As people are custom to use Metro GUI. Alternativies do not have so great change, no matter how great they are.
      (And at that point, Nokia has a change to get something back, but every other phone manufacturer rise with WP as well).

      And then Microsoft will release tablets…. Again, people are custom to Metro, they want Metro. They dont want to learn anything new, they want things to be same.

      Consumers dont have the power to decide the future of technology, it is corporations who does it.

      And in this point when there is going a re-design, it would be very important that every alternative gets a marketing so people would see it. Many users want to stick at old one, but when there is no easy way to revert from new to old, they stick on new or what ever is default.

      Android (one of the many distribution for mobile of Linux OS) will have difficulty times ahead… As in future there is no similar GUI on PC’s like what WP has because Metro. IOS will remain because Apple has the foothold on markets because Macs.
      So now if ever, every open source application and desktop should give a alternative view to have same look as Android offers, to desktop use. And that is what we already have. WIMP is better than touch interface for desktops and laptops. WIMP isn’t going anywhere. But the point is, Android and iOS has the common ground, they both follow WIMP design but fitted to small screens. We have icons, menus and pointer (finger) but not windows. The higer ground for Android is the applications. At somepoint if Microsoft succeed to get enough momentum to WP (it has endless cash because Windows and Office sales, littlebit same thing Canonical had and it abused that position) so it gets the needed apps. Then it is lost case for Android.

      The Open Source has not gone anywhere, even having actually more features, better usability, almost same applications and possibilities (execluding rare areas like CAD, Games… etc), but if the Open Source community can not push the teamwork and alternativity to markets trough OEM, so that we could get rid off the competition. It will remain small or have very slow growing.
      What we would need, is to ruling in EU that every PC need to be sold with empty HDD/SDD. Then any software system for PC, needs to be bought separately from the shelf (unless having own license and copy). BUT only if it is a OEM + software system manufacturer/maintainer. So if it is company what does both, a software system and hardware system, then they can sell them bundled.

      That would mean, every PC needs to come without Windows.
      Apple can continue selling Mac (Mac is not PC).
      PC OEM’s can start own Linux/BSD distribution and support it.

      But it would end up to situation where there are Linux distributors who sell their distributions next to Windows in store shelfs. And people need to make a choice what they buy with the new PC.
      At that point, the installation process comes important, it is easy already. But hey, that was the thing how it was done at 90’s. You bought a PC, OS/Software System, Internet connection, applications and you spend few hours at first evening to install everything.

      At those days, customer had change to vote. But it was then even bigger efffect by Microsoft, that they got their OS to PC standard in the first place. So it was demanded to get OS from Microsoft if wanted to have PC-compatible computer and run PC applications.

      Tablets and Smartphones are again in bad situation, those devices are just personal computers, even more than desktops or laptops. And it would be very important to apply same rule. Every smartphone needs to come without software system, so user can install the one what chosed, unless OEM is developing/maintaing their own. So Microsoft would need to manufacture own hardware system to sell WP preinstalled.
      Nokia could sell Symbian preinstalled in their phones.

      And at that point, consumer could just choose hardware system, choose software system and then live with it.
      And then every software needs to follow standard, what gets updated once a year or two, from file standards (ODF, OOXML) to network protocols etc.
      And every year, there are dozens of alternative developments for standards. But after the time period, together everyone choose the best one to be a standard and offers only it for markets.

      The GUI is not the problem, it is the compability. GUI should be changable by the user and by the device. But every file should be possible open any same class application. Every device needs to be possible connect to other, were it with wire or wirelessly.

  7. Greg says:

    If I had never used OS/2 Warp 3, seen what X11 window managers can do, and been aware of Mac OS, I might agree with you. I’d also probably use a computer in a very different way, and probably not so much. I’d probably have many more books, notebooks, and scraps of paper lying around than I already do. I might even have a file cabinet and an ordered bookshelf again; maybe even my own card catalog. I’d probably be successful, happy, and have other people deal with those infernal machines for me. I was going to say you’re right about touch and wrong about WIMP, but now, having considered the alternative, I say have at it.

  8. I’ve been using Gnome “3” for about 8 months now on a 1920×1200 monitor. I’m a software developer and pretty much live in vim and with a couple instances of Chrome. I really enjoy living in a maximized and horizontally split screen world. Being able to drag a window to the side of the dekstop to go 50/50 feels wonderfuly efficient. I often wish I could drag one of the windows wider and the other narrow automatically (e.g. 60/40, etc). My desktop feels very organized because windows aren’t overlapping. I make use of several workspaces and find ctrl-alt-{up,down} is the new alt-tab. I say bring on the non-WIMP desktop.

    One complaint I have is that while chat *is* integrated with the desktop it often behaves less than ideally. I end up with the regular chat window plus the shell integrated chat popup. That alone is weird. But the chat content between the windows sometimes is not synced making it feel less consistent. Also I sometimes miss chat notifications when they get hidden away in the notification bar if I don’t catch them in time.

    While on the topic of the bottom notification panel…it can also annoyingly get activated inadvertently and then steal window focus. I’d prefer to see the bottom panel removed and the top panel for notifications . Perhaps a glowing orb on the panel that aggregates all app notifications. Think gmail notification button in Chrome: http://www.intech-bb.com/wp-content/uploads/2011/12/google-mail-extension.png. Or heck maybe the glowing orb is the “activities” button and clicking on it somehow lists notifications. The orb could throb when the user is “available” or sit quietly when “busy”.

    • Florian Müllner says:

      “I often wish I could drag one of the windows wider and the other narrow automatically (e.g. 60/40, etc).”

      There’s actually hope that we’ll get this feature this cycle[0].

      [0] https://bugzilla.gnome.org/show_bug.cgi?id=645153

    • Fri13 says:

      W in WIMP means windows… You want to get non-WIMP GUI but still you love the change to have 50/50 view and you wish to have 60/40…

      The Windows means applications are separated and you can move them, you can scale them and you can see them same time.

      Non-WIMP (non-W actually) means that you have only one application at time and you need to do lots of alt tabbing or use a task manager to switch.
      Even using a netbook with 1024×600 resolution or older computer with 800×600 resolution makes everyone to notice that it is terrible idea when using keyboard and mouse. It works for tablets, but not for keyboard and mouse.

      The WIMP would work for big touch screens as well, it just would need some modification. Like bigger menus, icons and possibility to move and resize window from anywhere. Like with three finger to move and three finger tab and drag to scale like KWin Alt+Right button does or Alt+Middle button with Metacity.

      WIMP has nothing wrong, just blame the designers who abused its possibilities.

  9. Jake says:

    Yes, Apple and Microsoft are introducing touch flavoring to their desktop OS, but notice the difference in approach. OS X is still OS X. Sure, they introduced some full screen apps, but it’s an optional functionality and to my knowledge is not the default. And that’s different than maximizing all windows. Other things introduced are complementary to their traditional desktop experience. Windows 8 will contain a traditional desktop environment that users can easily jump into.

    It is actually unfortunate for GNOME and Linux in general that this is the case. Apple and Microsoft are not doing what GNOME did. That is, they aren’t trashing their working interfaces and giving their user bases the middle finger. They can’t, because it is their position that their software and user bases matter to them. It’s unfortunate that GNOME (along with most Linux distros) decided their own software and user base don’t. I’d like to suggest reading up on both “inferiority complex” and “self-fulfilling prophecy” on Wikipedia for possible insight.

    Since Apple’s and Microsoft’s software and user bases are the ones that matter, and since their next generation desktop software still firmly offers WIMP interfaces, we are not in a post-WIMP era. Not even close.

  10. Lapo says:

    Very good analysis, brilliant post. As usual I’d say.

  11. Benjamin Otte says:

    I think what you have experienced and most of the people commenting here have not is the transition from WIMP to the new world actually working.
    It’s kinda hard to understand how the work currently done with Inkscape, a word processor, a spreadsheet or a terminal will be done more efficiently (or even as efficient) with a touch device. All we see are people using the new “apps”.

    So we all excitedly wait for the Youtube videos of people developing complex C++ apps in Visual Studio Mobile and being more efficient at it than we are today.

    • As someone who lives in Inkscape pretty much all day (together with my text editor and web browser), I don’t think mouse+keyboard is as good as it ever gets for developing graphics. Having briefly used things devices such as the Wacom Cintiq [1] it’s a lot more natural to use a stylus and draw on the actual surface itself than using the mouse.
      Regarding the menus and widgets, they tend to mostly just sit there and eat away space from the actual drawing I’m working on. I find that less than ideal. Fortunatley Inkscape is moving more and more of it’s actions to be directly manipulated on the canvas itself, so it’s going in the right direction. Inkscape also have a couple of overlapping dialogs left (neatly covering the objects I want to manipulate), but it looks like those are on their way out too.

      For text editors I have do I agree that a physical keyboard seems like a better fit, and I could see the snap-on-keyboards featured on the Prime as being a good fit there (or even a bluetooth keyboard?).

      1. http://www.wacom.eu/index2.asp?pid=90&lang=en

      • Fri13 says:

        But tablet belongs to “pointer” class in WIMP. The pointer does not mean mouse but as well a pen or any other device what allows to move cursor (balls etc).

        Would you be happy to use inkscape with just your finger? WIthout pen, mouse and keyboard?

      • (thread is not going deep enough it seems)
        Fri13: Sure, however you end up classify a pen that I use to touch the screen with, it is such a device I mean. I just wanted to clarify that overlapping windows and a mouse is not as good as I think we’ll get for doing graphics. Fingers I would only use if I wanted to achieve something less presice (just as I resolve to finger painting on canvases in the analog world only very seldom)

    • drago01 says:

      Well what most people don’t get in those discussions is that touch devices like tablets are simply *consumption only* devices. No matter what kind of apps you find in the appstores that are trying to do be productive apps all of the are simply cumbersome and innefficent to use.

  12. mudfly says:

    I can’t stand any of the UI’s in the above examples. My android phone is quite a bit more capable then previous generations of phones. In addition to using it as a phone, I use it for GPS, Texting, Photos, and reading reddit or drudge while in waiting rooms. None of these tasks are superior to dedicated devices for these tasks, with the exception of texting. However I still prefer a slider keyboard to touch, I am much faster, and yes I have tried swipe. I simply find android/ios/metro completely inferior in UI and capabilities to any “wimp” OS that was released in the last 15 years. I also have a Nook Color running Cyanogenmod 7, that is a million times better then the stock Nook firmware. I pretty much only use it for reading books, because it is inferior to my workstation at regular tasks, The screen is nice for reading, but the touch to change the page irritates the hell out of me. Its always changing pages when I adjust holding positions, and I accidentally touch something. Then what gesture do you use to go back, all of the apps are different, and I can only wait to see the patents issued on “gestures” so every app is forced to implement different inputs. Lion to me looks cluttered, I hate messy desktops, and that is one messy desktop, not to mention that the people who have upgraded to lion at my work are less productive now, and have told me personally that they regret the upgrade. This is coming from the apple religious people. WIMP will not go away, there is no way corporations will install this crap on their workstations. Nearly every employee at my place of employment runs dual head, with multiple apps running at any given time, citrix sessions, email, jabber clients, telnet sessions, web applications, spreadsheets, etc. Please explain to me how “one app at a time” makes someone more efficient who is used to having 6 visible open windows at once? Yes I understand the attention argument, I don’t buy it.

    • vovkkk says:

      Funny, how almost every opponent of post-WIMP thinks that’s all about “one app at a time”…
      Why so? because your phone can only “one app at a time”, huh?
      So if it can two apps at a time, will you buy it?
      Have you ever tried tiling window manager?

    • The type of user that runs concurrent “citrix sessions, email, jabber clients, telnet sessions, web applications, spreadsheets” is not the average user. However, I guarantee you there will be ways power users can continue being productive in a post WIMP world. Gnome seems to be taking a hybrid approach and I support that.

  13. Craig100 says:

    Totally agree with Mudfly. I run Ubuntu 10.10 with both panels at the bottom of one of the screens in my dual screen set up. All windows I have open or minimised I can see and get to in one click. I have my most used apps pinned to the top of the two panels. It’s lovely, quick and efficient. If there’s a need for another version of the OS for touch capable devices, great, go ahead, fill your boots, but please don’t spoil our hard won progress on what has been up till now something that has been a pleasure to use. OK, it’s could be improved with less terminal and more gui, and if you want to make it prettier, most wouldn’t object as long as work flow isn’t changed, but that’s about it really.

  14. vzades says:

    one era ends another era begins

  15. Fred says:

    It’s important not to forget that GNOME should focus not only on one but many target devices. The PC with mouse input as well as touchscreen, tablets, perhaps someday phones. What I really like about GNOME3 so far is that I can use it equally well on my workstation as I can on my tablet, I can even use exactly the same GUI metaphors for the most part. Pure flexibility. It’s great that GNOME as one of the first Desktop Environments got rid of many (imho unnecessary, legacy) WIMP features.

    However on my development machine I’ll always need the ability to have a lot of windows open at the same time to get work done. I need to access files and folders quickly, best via shortcuts. Designing applications for fullscreen mode per default might not be such a good idea, but of course there should be an option to switch to that mode.

  16. mpt says:

    Many people nowadays see different designs for different form factors, and think that it’s actually a change in design over time. So I can understand how you might have cited the iPhone, the iPad, and let’s be generous, even the lackluster Xoom as “hugely successful” examples of “post-Wimp” interfaces.

    But you’re smart enough that by the time you cited OS X Lion, webOS laptops and desktops, and netbooks as examples too, a mental alarm really should have been going off somewhere. OS X’s Launchpad is uneasily redundant with its Applications stack. Its overlay scrollbars are used with trackpads, never touchscreens. Full-screen applications aren’t post-Wimp, they’ve been common in Wimp GUIs for quarter of a century now. webOS laptops and desktops, as you admit, don’t even exist. And as for netbooks … Hey, does anyone remember netbooks? Boy, those were the days.

    Your logical low point was where you asserted that the webOS “designers and developers clearly think that it offers advantages for” laptops and desktops. I’ll leave aside whether it was really the designers and developers who thought that, or whether it was executives trying to talk up its prospects … Because, come on, who the heck cares what any of them “clearly think”? Get back to me when they’ve sold a few million.

    It does matter, though, what people like you think who are designing interfaces that actually ship. Because by the time they realize that no, a single interface can’t delight both mouse and touchscreen users, they will have lost several years. Witness Microsoft, which plonked mostly-desktop Windows on Tablet PCs for nearly a decade — and a too-desktop-like Windows Mobile on phones for even longer — before realizing both were mistakes. Or take your own example of the Xoom, which committed the seemingly minor sin of shipping a phone OS on a tablet, but even that mismatch was too much.

    Now, it’s possible Metro will prove me wrong on all of this. But Gnome won’t; it doesn’t have the resources. Pick a form factor to ship on, and optimize for that. If you achieve a hit application, then consider porting it to something else. Better to be successful somewhere, than ignored everywhere.

    • Allan says:


      Well gee, thanks for finding my logical low point! That’s nice of you. 😛

      The Xoom might not be the most successful device. WebOS for laptops and desktops might not have originated with its designers. Windows for tablets and Windows Mobile were certainly inadequate. Those details are uninteresting and somewhat misleading. The point I was making is that Android, iOS, WebOS and Metro all feature superior approaches to application design in comparison to your traditional desktop OS, and they’ve got a lot to offer.

      Whether GNOME 3 can be successful on the kinds of devices that I’ve written about is obviously an open question. Maybe you are right and we should just pick a form factor and stick with it! The problem is that what we used to think of as stable form factors are starting to blur. The point of my post is that what we think of as laptops and desktops are changing. The question is – what are we going to do about it?

      • Greg says:

        “. . . Android, iOS, WebOS and Metro all feature superior approaches to application design in comparison to your traditional desktop OS . . .”

        I don’t know about Android and WebOS, but the makers of iOS and Metro have made a point of pushing a design approach. On almost every desktop but the Mac, it’s not hard to have a superior approach to design . . . all you really need is to have one. All four benefit greatly from having years of accumulated design knowledge without years of accumulated user and developer expectations. The typical smartphone is also more powerful than the typical computer desktop was when those were growing up; my iPhone has more processing power and memory than the laptop I used to use when I worked on the GNOME 2.0 HIG. Besides wanting to avoid patent lawsuits, non-Mac desktops also faced a broader hardware and user market. Make what you will of IBM’s CUA, but realize it was better than the chaos that preceded it. Maybe the same holds for GNOME 2 (and KDE around the same time): it was never as complete or cohesive as Windows or Mac OS, but it was a big step towards that.

  17. Craig100 says:

    “The problem is that what we used to think of as stable form factors are starting to blur.” Isn’t quite true. There are still many more “traditional” laptops and desktop machines sold. The issue is that there are now ADDITIONAL form factors available which are causing MMI (Man Machine Interface) issues with the existing OS’s. So, don’t throw the baby out with the bathwater by changing the OS to suit the new form factors so much that you now cause MMI problems on the desktop/laptop. I know it’s really a development resource issue that’s causing this. Really, the new form factors need a separate OS version to do the job right. What we’re getting are unsatisfactory compromises because either there aren’t enough developers interested in maintaining more than one version of an OS or someone’s come to the mistaken conclusion that you can successfully bridge all form factors with one OS version. I know it will all come out in the wash but to some of us, it’s so obvious.

  18. Nomad says:

    This totaly ignores the vast development tiling window managers like xmonad and awesome have made. Wimp always neglected the keyboard, and gnome did it do. With gnome 3 this has gotten better, but it’s still pretty badly usable if you want to avoid the mouse.

  19. Stefan says:

    WIMP is a paradigm. What will replace it (for the average consumer at least) is not really a paradigm. It’s “every vendor trying out another way to implement a slick and usable touch interface”. You can switch from one WIMP system to another easily (windows to osx to ubuntu). But what about Windows8? I only know the screenshot you have posted and I’m completely stumped what I should do with it. I only see a basket of text in a 3 x 3 pattern with a huge border of empty space around it.

    I hope, a new paradigm evolves, that makes it possible for companies to compete on the same playing field. But since apple patents all the stuff that makes sense, other smartphone and OS vendors are forced to so things differently just for the sake of doing it differently. sad…

  20. mpt says:

    (Minor correction to my previous comment: The Xoom shipped with the tablet-only Android 3.0. So for “shipping a phone OS on a tablet”, read “shipping a phone-derived OS that was, PR notwithstanding, not yet properly adapted for a tablet”.)

    If your point was only “that Android, iOS, WebOS and Metro all feature superior approaches to application design in comparison to your traditional desktop OS”, your post would have been much shorter. You had two points: (1) form factors are blurring, and (2) this doesn’t matter, because the kinds of interfaces used on mobile devices are an improvement for PCs anyway. The more form factors do blur (for example, if millions of people start daily docking their phone into a large-screen-and-keyboard terminal), the more people may indeed want those interfaces to be similar.

    I think, though, that both are mistaken. First, I think netbooks will be only the first prominent example of PC manufacturers trying and failing to blur form factors. Five years from now most people will be using phones, tablets, and/or TVs exclusively, not devices that try to split the difference between those and PCs. The kinds of people who still need to use PCs at all (e.g. office workers, software developers, engineers, some scientists, writers) will be highly correlated with the kinds of people who won’t mind learning a second, non-tablet interface (file hierarchy, menus, keyboard shortcuts, small click targets) for more features and/or greater productivity on the PC. If the majority of PCs are still sold with touchscreens in 2017, it will be because they’ve become so cheap that a non-touch screen isn’t noticably cheaper, not because people actually touch them much.

    With your second point, you’re right that “windows themselves produced additional management tasks for users”. Window management remains an under-solved problem. But your other examples fall into two categories. First, those that have already been common in Wimp interfaces for a decade or more: full-screen interfaces, maximized windows, and filesystem-eschewing applications. And second, those where a mobile-style interface is just as bad or worse than Wimp. An application lost on one of several phone home screens is no better than an icon lost on a desktop. And even if a Wimp interface does “not make efficient use of screen real estate”, mobile interfaces — to allow finger input — have to be much less efficient still.

    If your aim is to improve PC interfaces, there are many more fruitful opportunities. For example, Gnome could solve the problems where people forget what they’ve copied to the clipboard, or destroy one clipping with their next. Gnome could solve the problem where people have trouble sending files to each other. Gnome could establish a much simpler interface for document versioning than the Mac’s mélange of Undo/Redo + Versions + Time Machine. And so on.

  21. Mathias Hasselmann says:

    > There will be Metro style versions of Windows productivity software, including Microsoft Office.

    Allan, Microsoft’s lead designer not necessarily shares this opinion:
    http://www.theverge.com/2011/12/16/2640634/steve-kaneko-microsoft-design-metro-office-interview: “the large Metro style interface, designed for touch interaction, doesn’t scale in an obvious way to software like Office that has a lot of dense information”

Comments are closed.