For example, a common scenario of needing reference material while doing or creating something:
Project manager writing a budget estimate (excel or email) while repeatedly looking at a project proposal (word or pdf).
Secretary discussing a meeting schedule (email, skype) while looking at calendars of the involved people.
Any newbie learning a complex task or app while having a tutorial video or book open next to it.
Journalist writing an article (CMS, etc) while viewing multiple sources and previous articles.
In many of these cases multiple monitors would be useful as well, but modern monitor sizes easily allow writing/editing a document on one side of monitor while viewing the reference part on the other side. The historical way of doing this is to print out the reference document and attach it next to the monitor - there are tools for this in any office supply store. Now that screen space isn't so expensive, it's more efficient to keep both of these documents in the computer.
If you are a pure consumer of information, you may get by a single window. If you actually do something, then it's different.
So a natural use case is to edit a full page document on one side, and have another full page document next to it that you can read or reference.
And I like it.
I've run most my apps full screen for years, trying to get back to simulating something like the Amiga "screens".
Now, I do need to see multiple windows at times. But I also prefer not to most of the time. On the Amiga this was well catered for:
- You started with the Workbench screen, equivalent of sorts to a normal desktop + explorer.
- Applications could open private screens. On a private screen they could open a "backdrop window" covering the whole background, and treat it as a full screen UI. And/or they could also open floating windows on top of it.
- Applications or a utility could open public screens or place windows on existing public screens. These acted as extra desktops. Say I wanted a shell and a text editor in it's own workspace per project I'm working on.
Most larger applications let you choose between working in a window on the workbench screen, "fullscreen" on a private screen, or in a window on a public screen.
Screens can be "pulled down" to reveal the screen behind, if you for example want to have a quick look without losing your train of thought.
I'm 95% of the way back there now, with Ubuntu with Unity (even got the global menu bar back - yay!), but there are still annoying niggles in my workflow..
Another 10 years and maybe the rest of what I miss from AmigaOS will have made it back into the mainstream (datatypes, more pervasive unified scripting support of applications, user interfaces that don't grind to a halt when the machine is under a bit of load, to mention three...)
What you describe sounds very like the way I use dwm/dmenu (suckless tools on Linux) on a small screen. Each application has its own 'tab' (workspace in dwm), those are the private screens. Your 'public' screens sound like the dwm behaviour where you can 'merge' tabs on a temporary basis, then unmerge. There is also 'monacle' and 'floating' mode.
xfce4 can also provide 'private' screens by using the alt-F11 keyboard short cut to remove window decorations, so the program uses the whole screen (no panels or window decoration visible) but the program's menu bar still shows. Recent Firefox seems to break this, but LibreOffice works fine.
All that is on 1024 by 600. On 1920 by 1080 x2 I just use a lot of stacked windows.
I was actually planning on installing dwm or another tiling window manager until I had Ubuntu 12.04 installed on my new laptop and decided it was nice enough that I'd try it for a while, and I've stayed with it because enough of the apps I use on it handles very well to strip various chrome when maximized, and most of them are nicely integrated with the new global menu bar.
I do use alt-F11 or equivalent for many apps sometimes, but I find it's poorly tuneable for too many apps. E.g. sometimes I want some chrome, like tabs or location bar for a browser visible but nothing else, but the full screen modes for many apps are more "presentation modes" (Chrome on OS-X at least interestingly separates "full screen" and "presentation mode" - few apps do that)
It's definitively a way of working that came about due to small screens or low resolutions (the default Workbench screen on the Amiga was 640x256 for PAL and 640x200 for NTSC... 640x512 or 640x400 if you could stand the flicker of an interlaced display...), and that makes more difference on small screens, but these days I tend to use it for larger screens too.
Though on larger screens I do sometimes think I'd like a tiling WM again (I used to use Ion way back), so I might just give it a shot. I wonder if any of the tiling WMs would integrate ok with Dash and the global menu bar in Ubuntu, as I actually really like those parts, though... Perhaps it's time to do some testing again.
On my bog standard cheap 1080p monitor, I quite like the dwm main left and right stack tiling behaviour. Focus follows mouse with keyboard shortcuts for cycling focus. Mod-m brings the currently focussed window into 'monacle' mode, i.e. full screen. Mod-t sends the window back to its original tiled position.
Warning: trying out the dwm/dmenu repo package on 12.04 with Unity lead to problems with incompatible .desktop files, so odd behaviour in a dwm session with no obvious way back to log-in window. I had to stop X, use a tty to uninstall dwm and then got back to Unity. I've not fiddled around to find what the problem was. May have been sorted now.
Nice to have all this choice isn't it?
"If you build something so simple that even an idiot can use it, only idiots will."
Nobody will use an OS that meets the needs of the most common user types and stops there. Companies want to deploy the same OS across their entire install base. A modern OS must meet the needs of all its users. That includes power users. The question of whether or not most people really need multiple windows is therefore moot. Some do, so any OS that hopes to be more than a toy OS must support this feature. Full stop. End of discussion.
There are, of course, workarounds in Windows 8 that power users can use, so there's no need to panic just yet. Microsoft is certainly aware of how leery enterprise is of their tablet-desktop fusion. I'd be surprised if Windows 9 doesn't focus on being more desktop friendly.
"Imagine you were doing research and instead of spreading a number of books across your floor all open at the correct page you had them open stacked on top of each other and switched books when you needed to... Now tell me how frustrated you feel. Stuart Wakefield"
Although mostly single window, multiple screens kind of person myself I know the frustration of having to argue over the correct way for me to do my work that my paycheck depends on..
I'm the same way myself. I will say that even with multiple monitors "Metro" (modern) apps don't really perform well.
For whatever reason many of them aren't very information dense. So even if you move each one onto its own monitor for whatever reason much of the available space is wasted with solid colour or nonsense.
Take a look at this for example: http://i.imgur.com/TMSwc.png
Note how little of that is actual information and conversely how much of it is solid green background.
I'm all for having a multiple window UI accessible, or getting a rapid overview. But I find myself much preferring to unclutter my workspaces and have a single focus most of the time, and resort to the overview only rarely, especially because switching is so easy - I don't have to move big physical objects around in a stack, I can rapidly flick through a bunch of open windows on a carousel or get a display with a zoomed out overview.
I do find it important to be able to line up multiple windows though. Every now and again a single task will demand more than one window. I think that's where a lot of the disconnect comes from: Multiple windows on screen at once representing multiple tasks is a recipe for mental clutter. Multiple windows on screen that correspond precisely to a single, cohesive task is not.
The idea of only being able to look at one application at a time - and of not having a dead-easy and intuitive way to switch between them - seems like an utterly asinine way to screw up a productive workflow for the mere sake of fashion.
But in general I often have multiple things going on concurrently, be it video-chat + work, debugger + IDE, responding to an e-mail (having two e-mails open concurrently is useful), or just any task involving assembling information (e.g. searching with one window while building a "list" in a second).
The Metro/tablet style of interface is useful for content consumption, but as soon as you move into the realm of content creation then it starts to get tricky.
I really don't care what the average user wants, if the system doesn't support my usage pattern then I'm not going to use it.
We've had cheap client hardware up to now and for a few years more because the standard UI will support more complex needs. I think that as the bulk of 'casual' computer use moves to smaller portable devices, and as UIs suited to these 'casual' tasks are simplified, we might find ourselves paying extra for more capable devices.
Sad admission: I found myself using my phone to check a definition the other day while typing a complex paragraph in the wordprocessor on a laptop. It just happened.
Now, it depends on the screen (resolution). Isn't it pointless to maximize everything on a say 23" 1920x1080? I don't. on the side of my browser I often have chat, terminal, even spreadsheet windows. They even overlap sometimes, which I don't mind because its convenient for switching. However, on a 1280x800, I maximise my browser, IDE, etc.
For a while there was a windowing environment "twin" (Textmode WINdowing environment) available in some distros which behaved somewhat similarly visually. Looks as if the project's still alive on Sourceforge: http://sourceforge.net/projects/twin/
It can be somewhat useful on XVGA consoles with framebuffer enabled.
Single-window workflows work great on tablets and phones, but when you have a large monitor it's not always necessary to scale windows to the full resolution. Most apps are 1-dimensional — your code scrolls vertically, your video timeline track horizontally, and so on.
However, things are changing. Consumers have realised that it is more productive to use a tool to carry out a particular task and get the job done.
They are fed up of having to wait for their PC to boot, do any updates, fix any problems and try and remember the original task they wanted to do. It's no longer acceptable. That's even before we mention the confusion of having multiple windows (think grandma).
People now expect things to happen instantly so they can use a device as a tool to carry out a single task. (ie: I need check my bank account - switch on device, open banking app, check account, turn device off).
This type of paradigm shift means that not only do people no longer need multiple windows, they have realised they no longer want them either.
Not only that, but as we get more and more smart devices, we don't need or want multiple windows as we'll have multiple devices instead.
Like if you're recording your screen, you want to be able to see the controls, the viewport, and the recorded region all at once, typically on many monitors.
On small screens, space is a premium, so I can understand the motivation and want for full screen / less chrome.
It's also useful when using the computer on a TV. Ideally you want a lower resolution, or to increase font - sizes, windows controls etc.
On a larger monitor up close, I have no need to maximise the browser window. I get lost in a sea of white most of the time. In fact sometimes I need to resize the window - to make text more readable (fluid layouts - which I ultimately have more control over). If I'm using tabs, the resizing interferes with the other open web pages though. If I flip back to another page, I may need to then resize again - which is awkward. In this case I'm better off using multiple windows rather than tabs. Which I'd prefer if I had better window management.
I think that tabs were really just a hack around crappy window managers. And the real subject here is that of window management.
Yes you can use multiple monitors, but you can't assume that people do, or assume the size of their monitors.
I got so annoyed with crap window management over multiple monitors - that I prefer to use workspaces. I want to be able to use the keyboard to flip between monitors/workspaces, and move windows between monitors/workspaces. Launch windows on particular monitors. Turn external monitors off and on when I want to. Not having to have one monitor devoted to a task bar etc. Not having windows pop up where you don't expect them.
The last time I tried Gnome 3 I had trouble with multiple monitors. I can just about use ARandR under Xfce with an external monitor, but can find multiple workspaces don't work that well with multiple monitors.
I'd have thought that you'd actually be able to come up with something very usable with a little thought and common sense.
Back to having multiple windows. You pretty much need them for dragging and dropping. Which actually is a pretty awkward manoeuvre in itself. Other times having multiple windows on screen at the same time, is useful for things monitors/notifiers. Or when on messenger.
I personally alt-tab lots. But it can some times be jarring. And would be no use if I was wanting to watch something while chatting on IRC.
If I had a CCTV application, and wanted to watch four cameras at the same time. Would you leave that to the application, or the window manager? It seems to me that it makes more sense using a window manager rather than doing it in the application. I mentioned tabs already, and currently this UI is totally inconsistent across web browsers, let alone other applications.
I'm a web developer and I never do this. Why would I? One window is always sufficient for me.
With both open at once and the ability to auto-reload you see instant feedback by just moving your eyes a foot or 2 over (assuming dual monitors here) instead of having to alt-tab cycle into the correct window or move your hands and use the mouse.
I usually have a few things in view while developing and it's just annoying (read: poor user experience) if I have to fumble around switching between full sized windows.
Why would I need a full screen terminal. It's a million times more productive to use something like tmux and have 3 or 4 terminals open which are effectively multiple windows.
I certainly don't want the distraction of stuff auto-reloading while I'm still working on a change (e.g. switching to the next file I want to update).
But, then again, the article the question are referring to are kind of wrong anyway. The Windows 8's single window thing is meant for tablet and other consumption that is suitable for that kind of thing. His argument about the Windows 8 becoming less usable for power users are completely wrong, the metro interface is meant for the things that you would do on tablet, i.e. not power user-ish things.
In windows 8 the only way to create an ad-hoc network is via the command line.
In windows 8, you cannot configure a VPN to redial a dropped connection.