What OS do you use?

[Iranon]

You can get rid of the Metro silliness with one simple registry edit. Good, because I found the integration with the traditional environment beyond clumsy.
Even more useless near-duplication of system settings and less efficient interfaces requiring too much mouse movement leave a bitter taste, and there seem to be more services running amok even if a neurotic users best attempts to kill them with fire.
On the upside, the search function seems to have taken a much-needed step forward...

*

Regarding open source DEs... I think it's interesting how we're getting a lore more diversity in desktop environments. I was actually impressed with Gnome 3's attempt to make a pleasant and user friendly interface for non-technical users.
Sure, it sacrificed some efficiency ("let's make misclicks impossible by excessively wide spacing") and doesn't cater to intermediate users to whom friendly means familiar, but those are besides the point.
My only major disappointments are substandard configuration methods and a backwards approach to settings: installing extensions and ticking checkboxes should enable functionality, not get rid of unwanted ones.

Yes some bits are annoying but I find most major desktop environments bloated and limited and intrusive. So far, nothing else has allowed me to tweak my environment to my preferences as well as FVWM, with the bonus that its configurability doesn't result in visual clutter (I'm looking at you KDE). I also prefer open-ended text-based configuration to navigating a 10-page wizard.

[Copper Bezel]

That would only be true if it was only (or largely) going to be used on a touch device. As long as it's presented as a primary UI on the desktop as well, it's eminently fair to evaluate it in that environment.

It's an odd thing, trying to sort where Microsoft actually intends to put Windows 8. PhoenixEnigma is the rule, not the exception, in having skipped Vista entirely, and Windows is different from OSX or most Linuxes in that the majority of its consumers never upgrade to a major new release on the same hardware. I really think Microsoft might be buying into the idea that touch is going to become ubiquitous on new hardware within a few years, and if they do believe that and if they're right, then making Windows 8 a touch-first OS makes sense, because it can be released primarily on touch platforms while non-touch platforms wait it out, and its relative success or failure will indicate the direction for 9. Comparing 8's mouse-and-keyboard usability to OSX now, which is entirely mouse-and-keyboard driven, or Gnome Shell, which is meant to make an even compromise between platforms, doesn't seem like a fair evaluation of the UI as a whole. Admittedly, it's a perfectly fair evaluation of how well it runs on existing platforms, but we don't really know how relevant that is just yet.

Regarding open source DEs... I think it's interesting how we're getting a lore more diversity in desktop environments. I was actually impressed with Gnome 3's attempt to make a pleasant and user friendly interface for non-technical users.

I don't know if I would call myself non-technical, but I don't code and I survive with a netbook (plus an ad hoc docking station at home) as my primary machine, so I'm closer to that category than the alternative. I do like features that move "big" actions to the content area of the screen or application, whether that's a speed dial in the browser or Exposé window selection. Using a home screen for app switching, launching, and closing is convenient to me, and the centrality of search-based launching seems like a good thing for users at any level of technical ability.

Again, that use of the content area of the screen is something I actually like about Windows 8 in theory, though I haven't tried it in practice, and what it's doing with its very different interpretation of a home screen. I might tremble at the hideous clutter of a typical Windows user's desktop, but my impression is that a majority of users actually use desktop icons as the primary means of launching apps. Windows 7's pinned apps choked that back a bit, but Windows 8's home screen seems to me like an attempt to embrace it instead, as, We give up. If you still like your desktop icons, have them, but let's at least make them easy to get to, pretty, and useful instead.

My only major disappointments are substandard configuration methods and a backwards approach to settings: installing extensions and ticking checkboxes should enable functionality, not get rid of unwanted ones.

I use Chrome as my primary browser, so Gnome Shell's handling of extensions makes sense to me. Rather than having an application with a lot of built-in options, you start with a sane default system and a flexible base to plug things into to add and remove features.

The configuration in Gnome 3 that doesn't happen through extensions, that is, configuration for things that aren't the shell, is severely broken; settings features are squirreled away in unexpected places, simply don't work as expected, leave out available values for a setting in one place and then provide them in another place, or are simply absent. My first stop in trying to change a setting is dconf-editor, and when it's easier to dig through a tree and alter a key than it is to click a button, something is wrong.

Finally having an "Up" button alongside "Back" for navigating filesystems is nice.

Gnome actually just took that away in 3's Nautilus file browser. I miss it. (In fact, at a glance, it's like Explorer and Nautilus traded UIs in 7-8 and 2-3.)

[EvanED]

That would only be true if it was only (or largely) going to be used on a touch device. As long as it's presented as a primary UI on the desktop as well, it's eminently fair to evaluate it in that environment.

It's an odd thing, trying to sort where Microsoft actually intends to put Windows 8. PhoenixEnigma is the rule, not the exception, in having skipped Vista entirely, and Windows is different from OSX or most Linuxes in that the majority of its consumers never upgrade to a major new release on the same hardware. I really think Microsoft might be buying into the idea that touch is going to become ubiquitous on new hardware within a few years, and if they do believe that and if they're right, then making Windows 8 a touch-first OS makes sense, because it can be released primarily on touch platforms while non-touch platforms wait it out, and its relative success or failure will indicate the direction for 9. Comparing 8's mouse-and-keyboard usability to OSX now, which is entirely mouse-and-keyboard driven, or Gnome Shell, which is meant to make an even compromise between platforms, doesn't seem like a fair evaluation of the UI as a whole. Admittedly, it's a perfectly fair evaluation of how well it runs on existing platforms, but we don't really know how relevant that is just yet.

Touch is an awful input regime for longer-term desktop use from an ergonomic standpoint. Either you have the monitor set at a comfortable height for viewing -- in which case you have to hold your arms in the air for long periods of time -- or you have the monitors set at a comfortable height for touching -- in which case you have to stare down at your desk for long periods of time.

It's not that I can't think of any applications of where touch would be nice, but I don't think that many of them are covered by desktops now, and I think that where desktops are used now, touch is almost always a bad idea.

[userxp]

I agree with that. I dislike Metro and Unity for the same reasons: the icons are too big and poorly placed. In Windows 8 to get to the "Actual apps" you have to scroll all the way to the right, past the Microsoft-made "cool Metro apps" (and why would I want to run an application full-screen anyways?). In Unity you can either choose between the 5 apps you use the most, or between ALL the apps in your computer, in apparently random order, mixed with random applications that you don't actually have (from the download center or however it's called) which usually don't work anyway.*

For reference, you can download Windows 8 Developer Preview here.
Fun fact: when typing "Windows" I accidentally pressed control instead of shift and closed the tab. Why Opera usually decides not save the text you have (even with an extension designed specifically for this installed) is beyond me.

Touch is an awful input regime for longer-term desktop use from an ergonomic standpoint. Either you have the monitor set at a comfortable height for viewing -- in which case you have to hold your arms in the air for long periods of time -- or you have the monitors set at a comfortable height for touching -- in which case you have to stare down at your desk for long periods of time.

Or both.

*Disclaimer: I have never actually used Windows 8 or new versions of Ubuntu for longer than a few minutes.

[EvanED]

Touch is an awful input regime for longer-term desktop use from an ergonomic standpoint. Either you have the monitor set at a comfortable height for viewing -- in which case you have to hold your arms in the air for long periods of time -- or you have the monitors set at a comfortable height for touching -- in which case you have to stare down at your desk for long periods of time.

Or both.

*Disclaimer: I have never actually used Windows 8 or new versions of Ubuntu for longer than a few minutes.

I have seen that before, and while it might work, I'm still skeptical, at least of that being a significant mode of interaction in the near future.

[Copper Bezel]

Yeah, I don't see anything like that taking off to become the norm. I also like the idea of a motion-sensing space over the keyboard or something, and it will be interesting to see what hardware arrangements develop in response to Windows 8, but I don't see the pointer and keyboard arrangement disappearing anytime soon, and I imagine that the next Windows release will have to make some compromises.

Touch is an awful input regime for longer-term desktop use from an ergonomic standpoint.

Oh, I know. I mean, I think there's a lot of office work that could happen on an easel-type desktop and a tablet, but touch doesn't offer a lot of advantages for real content generation work. Again, I really see Windows 7 remaining as the platform of choice for desktops, and I really think Windows 8 has already positioned itself as a little nichey.

I agree with that. I dislike Metro and Unity for the same reasons: the icons are too big and poorly placed.

Well, it's true that they're both concerned with making the click targets as big as possible. The difference is that Unity doesn't get Windows 8's excuse of being better on touch, either. Unity is such a transitional creature right now; it's not usable on tablets, but I don't really see any real advantage on the desktop, either. (It's much better than using Gnome 2 at defaults, but that's a very low bar.)

I don't really know why the Ubuntu-logo Home button on the Launcher was included at all, though, save perhaps for its future variations for touch devices. When you have a dock and a search-based launcher, you don't need anything else, and I was quite comfortable with DockBarX and Synapse while I was on Gnome 2. Unity (like Synapse) also has much better, cleaner search results than Gnome Shell gets, so it gets points for that.

If you're not using the Metro apps, you definitely wouldn't be getting much of the benefit of the Windows 8 interface; the home screen does a lot of things a Start menu doesn't, but it only does them for Metro apps.

It's a little unfortunate, too, since Metro apps can only come from the Windows store - so, say, Internet Explorer gets to be slick and chromeless, but Firefox has to run in the desktop app. I think I'd really like Metro, but I'm only referring to the interface; Windows 8 as a whole is going to be fairly useless in relation to running actual applications.

[troyp]

Yeah, I don't see anything like that taking off to become the norm. I also like the idea of a motion-sensing space over the keyboard or something, and it will be interesting to see what hardware arrangements develop in response to Windows 8, but I don't see the pointer and keyboard arrangement disappearing anytime soon, and I imagine that the next Windows release will have to make some compromises.

There was the Touchstream/MacNTouch idea of combining a keyboard and touchpad, but I can't imagine a zero-force keyboard being efficient for actual *typing*, although I'm sure it's great for everything else. Personally, i like the idea of a (widely) split keyboard with a touchpad in the middle, or else a keyboard with a touchpad behind it, slanting slightly upwards.

That 10gui interface doesn't seem that impressive to me. I mean, it seems like a very powerful interface, but then it's based on 10-finger multitouch, so that's simply what you'd expect.

The difference is that Unity doesn't get Windows 8's excuse of being better on touch, either. Unity is such a transitional creature right now; it's not usable on tablets, but I don't really see any real advantage on the desktop, either.

That's not really surprising, given that Unity's goal is a single GUI across all devices. I mean, they've set themselves what seems a very difficult goal, there. I don't even see why they think it's a *desirable* goal, personally. I suspect it's more to do with branding than usability.

[Copper Bezel]

Yes and no - the brand identity, not the feature set, is what's supposed to be consistent, and what we're seeing is the desktop version. The idea is to have a consistent design and metaphors, but little differences to suit the form factor. Ubuntu TV doesn't have a top panel, for instance, and the apps are fullscreen (as you'd expect.) The tablet interface will probably be similarly modified. So I don't think it's fair to blame any part of Unity's design on concessions to touch devices, because it's theoretically not making any. At the same time, Gnome Shell does attempt to work both ways, and it's just as usable on the desktop as Unity (or, in my rather subjective opinion, more so.)

A lot of Ubuntu users are using Unity and loving it, and I'm really selling it short. There are just a lot of little oddities I can't quite understand the logic behind.

There was the Touchstream/MacNTouch idea of combining a keyboard and touchpad, but I can't imagine a zero-force keyboard being efficient for actual *typing*, although I'm sure it's great for everything else. Personally, i like the idea of a (widely) split keyboard with a touchpad in the middle, or else a keyboard with a touchpad behind it, slanting slightly upwards.

Yeah, I like the idea of the trackpad up behind the edge of the keyboard and I've wondered why ergonomically split keyboards didn't shove a trackpad up in between the halves, but ergonomically, I really don't mind my netbook's arrangement, either, with the trackpad just under the space bar, easy to manage with a thumb but impossible to accidentally hit while typing. I really think that if we're talking trackpads, the problem's largely been solved.

I'm operating on the assumption, too, that there should be a physical keyboard and that the user's hands are generally going to be on it. I hadn't seen the MacNTouch keyboard, and it's clever and gets points for using straight-keyboard alignment. If it's possible to learn to touch-type at the same rate without any actual tactile feedback, I'm up for the challenge, but otherwise, there's no point. And it would still be operating like a trackpad (moving a pointer,) not like a touch device where a point on the input corresponds to a point on the output, so I don't think it solves the problem, either. (And anything that did operate like a touch device would pose some very silly problems for multiple-monitor arrangements.)

[troyp]

Yes and no - the brand identity, not the feature set, is what's supposed to be consistent, and what we're seeing is the desktop version. The idea is to have a consistent design and metaphors, but little differences to suit the form factor. Ubuntu TV doesn't have a top panel, for instance, and the apps are fullscreen (as you'd expect.) The tablet interface will probably be similarly modified. So I don't think it's fair to blame any part of Unity's design on concessions to touch devices, because it's theoretically not making any. At the same time, Gnome Shell does attempt to work both ways, and it's just as usable on the desktop as Unity (or, in my rather subjective opinion, more so.)

Huh. I had the impression (from one of Shuttleworth's blog posts, I think) the consistency was meant to be a bit more extreme than that, but maybe I took him too literally.

I'm not overly impressed with Unity, either, but it is quite new, so maybe it'll improve over time. I think the degree of customizability it ends up having will be one of the most important factors in its eventual usability. One thing I'm happy about is the emphasis on a text-based launcher, and I hope Unity means Canonical will put a lot of effort into developing it (mind you, when I last - briefly - tried Unity, I found Synapse to be superior to the Unity launcher anyway :-/).

I'm operating on the assumption, too, that there should be a physical keyboard and that the user's hands are generally going to be on it. I hadn't seen the MacNTouch keyboard, and it's clever and gets points for using straight-keyboard alignment. If it's possible to learn to touch-type at the same rate without any actual tactile feedback, I'm up for the challenge, but otherwise, there's no point.

You know, despite what I said before, it really should be possible to type full-speed with practice. I'm sure it'd be harder to learn without feedback, and I just don't like the idea of it, but I can't see why you couldn't learn to type fast on one. I guess the only way to know for sure is empirically, but I've never used a MacNTouch, or known anyone who had one (although I had a mad urge to buy one on eBay the other week). One thing that might help is to have some ridges on the keyboard to indicate home position (and maybe other locations) to help a typist keep their hands positioned correctly.

And it would still be operating like a trackpad (moving a pointer,) not like a touch device where a point on the input corresponds to a point on the output, so I don't think it solves the problem, either. (And anything that did operate like a touch device would pose some very silly problems for multiple-monitor arrangements.)

Well, these keyboards were made to operate with existing OSes, so there was a limit to how much they could innovate, but they did try to take advantage of multitouch: there was a wide range of gestures you could use that mapped to various operations (I don't know all the details). Presumably with an environment that supported multitouch they could do a lot more.

[Copper Bezel]

Huh. I had the impression (from one of Shuttleworth's blog posts, I think) the consistency was meant to be a bit more extreme than that, but maybe I took him too literally.

Yeah, that was all I heard for a while, too, but at least since the pitch for an Ubuntu phone in 2014, he's talked about it as an interface "family", and Ubuntu TV backs that up.

I'm hoping to see Unity do well, too, and I agree that the centrality of search-based launching is a big perk. The more the shell can operate like a browser, the better. And Synapse is still amazing - I've tried to wean myself off and use Gnome Shell's search whenever possible, but I still keep Synapse around for a number of reasons.

I see what you're saying about the MacNTouch. It will be interesting to see what that concept could do with a touch-based OS, and I'd still really have to try it to see whether or not it would affect typing in the long run, but there's no compelling reason to think it would. A Libretto-style virtual keyboard that could do other, non-keyboard things could be smashing.

[Iranon]

Random musing:
In past decades, we went from a CLI to single-task GUIs. Then came multitasking with tiled windows, to be supplanted by free-floating ones.
These days the efficiency-minded go back to tiling, "fullscreen mode" is a hot feature and "distraction-free computing" a hot buzzword, and the most-appreciated increase in usability comes from repacling clicking with typing: search-based launching may as well be a command line with a different take on autocomplete.

[Copper Bezel]

Oh, absolutely. Before finding out about Synapse, I used xfrun4 as my primary launcher, which is just an auto-completing run dialogue, which is just a terminal prompt that doesn't require hitting Tab and closes itself after launching the app. I totally see the continuity there. But what's made search-based launching really viable is what it's learned from smart address bars in browsers and the extent to which users have adapted to those same address bars.

Tiling window managers and stacking ones co-evolved. Xerox PARC actually used stacking, as did the first Mac GUI, while early versions of Windows used tiling and Xerox toyed with it in later versions. They've always been touted as more efficient, and Windows, to its credit, has always had some tiling capacity built into its window managers. The mouse-driven edge snapping introduced in 7 really changed tiling from being an artificial, clunky way of maximizing efficiency into an intuitive, spatial thing. Like smart address bars, it's a way of allowing the user to throw some sloppy input at the system and get a clean, automated result. KDE, Unity, and Gnome were quick to pick it up, and the chromeless, touch-based tiling in Windows 8 is just a further variation on the same idea.

I don't see it as going back to or re-envisioning a older system in either case. They're just different strategies that have enjoyed prominence at different times as each one has evolved and adapted.

[troyp]

Snapping was available for Compiz before Windows 7 came out, although it might have been 7 that made it take off. Before 7 (well, maybe Vista, if that counts), Windows only had "tile vertically" and "tile horizontally", which were very basic. Still, at least it did have those. I remember at one stage on at least one of the major GNU/Linux DEs (KDE, I think?), there was basically no tiling functionality at all. There was, like, some unmaintained extension from years ago which I don't think would even install ^*

[*] I remember finding a thread where someone was asking about it, but he apparently chose the wrong forum and made the mistake of saying "even Windows can do that", because there was maybe one brief response before the moron brigade arrived saying things like "Linux isn't Winders, get it?"

[Iranon]

Yes, poor-man's-tiling in a spatially intuitive way is nice. I have this tied to a few dozen right-click mouse gestures on titlebar - drag left for left half of the screen, up-and-left for first quarter, left-then-up for first eighth, up-then-left for second eighth and so on.
Together with half a dozen tile-everything shortcuts, I find this almost as efficient as dynamic tiling and a lot less intrusive. Also, sloppy focus with automatic raising lets me get away with significant overlap for better use of my screen real estate.

Incidentally, I didn't mean to imply the older systems are always better, more that an implementation that makes sense to the user is more important than what underlying concept we use. However, imo modern desktops give up too much open-endedness and ease of implementing personal preferences manually.
To stay with the example at hand, drag-to-edge snapping in the style of Windows 7 is extremely easy to enable in some old-school window managers. Nobody needed to anticipate that users would want this specific feature, they simply had to make their configuration syntax open-ended and powerful enough.

[Copper Bezel]

Incidentally, I didn't mean to imply the older systems are always better, more that an implementation that makes sense to the user is more important than what underlying concept we use. However, imo modern desktops give up too much open-endedness and ease of implementing personal preferences manually.

Yeah, I'm not quite convinced yet. Unity doesn't allow for a lot of customization, and Gnome Shell's extensions are not always reliable (in that they often conflict with one another or fail to work in unusual situations.) At the same time, there really are things that Unity and Shell can do with an integrated system that other DEs can't, and while I don't know Javascript, Shell's extension base seems to qualify as a relatively open-ended "configuration syntax."

I'm content with Shell for my own use, but there's definitely a limit to how much it can be modified by the end user to keep up with new methods, which means that it might age more quickly than Gnome 2 did or KDE does, with their more modular designs that allow components to be easily swapped in.

[troyp]

On the topic of search-based UI, Shuttleworth's latest blog post introduces the so-called "HUD" (Heads Up Display), which uses fuzzy searching to choose menu commands in applications. Looks promising (although I'm mildly skeptical of their eventual goal of dispensing with menus altogether). I'm really pleased that Canonical is pushing search technology further: I think it has enormous potential that's only been partially explored.

On the down side: As if "HUD" weren't a lame enough name, Shuttleworth has coined the term "intenterface".

[userxp]

For using applications with the keyboard: this seems like a logical next step. Just like you can launch applications in any decent OS by pressing a key and the first few letters of its name (I remember using Katapult and Yakuake, back when I used Kubuntu), it makes sense being able to invoke an action by typing a few letters. However:
-It's useless unless you are using the keyboard AND already know which menu option you are looking for, so definitely no replacing menus.
-I don't like its "Unity-ish" transparent/blurry/glassy look and large size. I'd rather have it more integrated with the menus (a small textbox popping up above the standard menus, similar to Windows' start menu).
-You can't just create a "new revolutionary super-awesome interface" in your head and set it as the default one in one of the most used OSs in the world, no matter how convinced you are that it's awesome. They do know what "usability testing" means right?

Also:

Voice is the natural next step
Searching is fast and familiar, especially once we integrate voice recognition, gesture and touch. We want to make it easy to talk to any application, and for any application to respond to your voice. The full integration of voice into applications will take some time. We can start by mapping voice onto the existing menu structures of your apps. And it will only get better from there.

Spoiler:

NO I don't want a "fully integration of voice into applications". There are so many problems about that:

• Poor accuracy. People often have to repeat what they said when talking each others, and our brains are much better than any computer at understanding speech (probably close to as good as they can be). And keep in mind that a good speech recognition software is not exactly easy or cheap to make, and this is free software we're talking about ("What do you mean your Ubuntu doesn't understand 'New document'? Well maybe, but look, it can read you the GPL license aloud in any language! Ain't that cool?").
• It's tiring. Unless I'm on a very talky mood I often just want to relax with music and use my computer in silence. Bonus: imagine a room with 100 people using a "fully integrated voice interface". Bonus 2: input your password using this interface.
• Languages. Speech recognition is basically nonexistent beyond the first 3-5 most spoken ones, and speaking to a computer in a language that's not your native one is A LOT more difficult than using a text interface, especially since you have to pronounce things the exact way the computer wants them. And even if you're English, try telling your music player to play "الغواصة الصفراء" next.
• Windows Vista/7 has one of those already (so much for "Ahead of the competition"). I've never seen anyone actually use it, ever (other than for testing/fun). Last time I tried to, I spent one minute (literally one full minute) trying to get it to understand the letter "f".

I think that's enough ranting for today. I believe that speech recognition, like touchscreens or motion control, might be very practical in some cases, but it's not for standard computer usage.

[Copper Bezel]

Voice is cack at the moment, and all the more so on Linux. I don't see that going anywhere. If it can work successfully, then it would be very, very nice for pausing media from across the room, like the XBox implementation, but not much else.

I knew of at least one person who used Windows 7's speech-to-text, but it was because she was a bad person.

The HUD, on the other hand, I like. It's meant to use fuzzy, semantic matching the way Google does, which should make menu items more discoverable; type "scale" and get entries that happen to be named "resize" instead and so on. Add that to the Zeitgeisty habit-forming stuff, and it really does start to sound better than poking through menus and trying to remember where this or that app puts this or that function; theoretically, the menu system becomes custom-made for your habits without any work on your part. And it warranted a BBC article, so that's cool.

Of course, it could also be terrible. But it's being presented as an option, of course, not a default.

Edit: And yes, "intenterface" makes me cry, and it should make you cry, too.

[EvanED]

I'll probably come back and edit or repost a bit more, but from that article...

Mr Shuttleworth admits people may find it harder to discover the full functionality of a program if they do not have a menu system to hunt around, but he says other benefits make up for this loss.

With all due respect, Mr Shuttleworth is a bit of an idiot. He's forgetting the principle benefit GUIs have for most tasks, which is easy discoverability by playing around with it. (Typing random strings into the search to see if anything comes up is not easy discoverability.)

I think the idea sounds great (it sounds a bit like the Windows Vista/7 start menu and control panel searches, which is great, applied to menus), but removing menus is outright stupid.

[Copper Bezel]

Yeah, but they're not actually removing the menus, either.

I agree that a method to browse available categories would be nice - the bookmark extension I use in Chrome for things I don't add to my speed dial works this way, as a collapsed tree of folders that can either be clicked through or searched (at which point the folders in the tree containing matches simply expand.) At the same time, heavily nested menus aren't any good for discoverability, either. (I wouldn't be surprised if a more mousy navigation style popped up in a future version - something intended for browsing, probably echoing the Dash categories - and I'd honestly expect that to happen before something like this became a default.)

If this were actually being presented as the default in its present form, I'd see it as a slight efficiency advantage if you know what you're looking for. Even with a new application, you could theoretically find the action you want more quickly than you could in a typical menu system, so I think "discoverability" might actually be an inexact term - what's not here is browsability. That's possibly the most important function of menus, the answer to the question, "So what does this application I just installed actually do?"

[EvanED]

So first of all, I think the longer term voice recognition has promise. Like touch I doubt it has much of a place on the desktop. However I think that for a variety of applications it could come in very handy. And it doesn't seem to be all that bad. For instance, this post is being typed with a voice recognition in windows seven. I am using the mouse and keyboard occasionally because I don't quite understand how the interface works and what commands are available. However the actual text is unedited. And there are only a couple mistakes in the first paragraph. Insert

[Typing now]
I wouldn't expect something like dictating documents or emails to become mainstream (as vital as it is to certain kinds of disabled folks), but there are a variety of situations where you don't have your hands free. And as computers become even more widespread, that will become more common.

Yeah, but they're not actually removing the menus, either.

You may be right; I was going off of troyp's post that said it was a long-term goal, but I don't see anything like that skimming over Shuttleworth's blog post a couple times. I retract my "idiot" claim under the condition that making the menus inaccessible is not, in fact, in the works. (Though I think even setting the default to hidden, at least on desktops, is enough to qualify as being an idiot.)

At the same time, heavily nested menus aren't any good for discoverability, either.

Way moreso than something command-line style though, even if its a smart command line.

I think "discoverability" might actually be an inexact term - what's not here is browsability. That's possibly the most important function of menus, the answer to the question, "So what does this application I just installed actually do?"

That's fair.

[troyp]

You may be right; I was going off of troyp's post that said it was a long-term goal, but I don't see anything like that skimming over Shuttleworth's blog post a couple times. I retract my "idiot" claim under the condition that making the menus inaccessible is not, in fact, in the works. (Though I think even setting the default to hidden, at least on desktops, is enough to qualify as being an idiot.)

So what he said at one point was...

There’s still a lot of design and code still to do. For a start, we haven’t addressed the secondary aspect of the menu, as a visible map of the functionality in an app. That discoverability is of course entirely absent from the HUD; the old menu is still there for now, but we’d like to replace it altogether not just supplement it.

So he does want to get rid of menus, but he's not going to do it unless/until he's got a viable alternative. That quote from the BBC article completely misrepresents what he said, period. (If he had said that, I'd call him an idiot, too). I'm not sure if they'll succeed in creating something to replace menus entirely, but I'm glad they're trying, at least. I don't feel as though menus are the last word in application UIs.

As for voice-based UI, I think it has a lot of potential as long as it doesn't lead to other means of interaction being removed.

[coyotebush]

Incidentally, OS X, for some time now, has had a (relatively basic) menu-item search integrated into the Help menu.

I've used it approximately once, and can't say I've seen anyone else do so.

[J the Ninja]

Incidentally, OS X, for some time now, has had a (relatively basic) menu-item search integrated into the Help menu.

I've used it approximately once, and can't say I've seen anyone else do so.

I use it pretty often. It's nice in media-production apps like Photoshop, After Effects, Maya, etc, where you have a LOT of branching menus. Hell, I've had the prof in an After Effects class specifically mention it as a way to help find things.

[Copper Bezel]

Implementation makes all the difference. I don't use OSX, so I don't have any experience with the menu search feature, but it doesn't sound like it's designed to be a first step. (Of course, we still don't know how successful the implementation will really be for the HUD, either.)

[userxp]

OK so Windows 7 was acting bad again so I finally made the switch to Linux Mint (best one I could find). Two things:
1. None of the distros I tried had any support for power profiles or anything similar. This is VERY important. On Windows 7, I can put my laptop in "Power saving" mode, which gives me >6 hours of battery time but does not let me play any games, or "High performance" mode which gives me 1-2 hours but is significantly faster. I am trying various programs now to see if I can get power saving in linux too.

2. Unless anyone has any arguments against, I'm going to make my own user a superuser. I mean, I'm the only one using my computer, and I trust myself enough to not do anything stupid. Anyone that criticized UAC but defended Ubuntu is an hypocrite.

[Copper Bezel]

UAC is obtrusive for the enduser, although it's actually superior as an architecture. Anyone hating on UAC needs to at least acknowledge that, but the implementation leads to a lot of unnecessary hangups that don't exist in Linux. That's where the criticism comes from.

There are very good reasons for both. I would strongly discourage using a graphical admin session for daily tasks. It opens you up to a lot of security problems, because any particular application you run also has admin privileges (including, say, the Flash plugin in your browser) and that's seriously bad security. Account control really can prevent accidental damage, too, especially if you're new to Linux. Breaking things and fixing them is a great way to learn a system, but this isn't something you want to start with. In normal, daily use, you don't need to elevate at all, except through the Software Center.

Mint was a good choice, however. It's Ubuntu with a few tweaks, and the result is more or less the same as what you'd get if you had a more experienced user set up an Ubuntu install for you. Most of the differences have to do with legal issues (things that Ubuntu can't install by default for license reasons, particularly non-free codecs and plugins.) For good or ill, it also uses the more mainstream Gnome Shell over Ubuntu's Unity interface.

The lack of power profiles can be a problem. I've written a shell script in the past for toggling settings like that (though it wouldn't work with the newest Gnome.) The best CPU scaling app right now is Jupiter, which allows you to set your usage for on- and off-AC and toggle it from the system tray. Gnome allows you to set backlight settings for on- and off-AC in Screen settings.

[EvanED]

UAC is obtrusive for the enduser, although it's actually superior as an architecture. Anyone hating on UAC needs to at least acknowledge that, but the implementation leads to a lot of unnecessary hangups that don't exist in Linux. That's where the criticism comes from.

Like what?

A bit after Vista came out, so I would have ammunition in /. arguments about UAC, I kept a log of every UAC prompt I got for basically an entire month. The significant majority of them were triggered by actions where the analogous thing in Linux would also have required sudo.

The primary exception to this was changing environment variables. If memory serves, at the time the dialog where you do this had both per-user and system-wide environment variables, and so to edit your own you needed to elevate. Even this was fixed in Win7.

The lack of power profiles can be a problem. I've written a shell script in the past for toggling settings like that (though it wouldn't work with the newest Gnome.) The best CPU scaling app right now is Jupiter, which allows you to set your usage for on- and off-AC and toggle it from the system tray. Gnome allows you to set backlight settings for on- and off-AC in Screen settings.

Thanks for that for me too! I recently put Ubuntu on my laptop and have been frustrated by the lack of good controls here. I did find some program (cpufreq? maybe) that gives me the ability to control the scaling, but I can't set a different profile for plugged in and unplugged. So I'll have to give that a try.

[userxp]

Oh don't worry, I've used Linux before. That's why I wanted to become superuser: because I often think "I just need to rename that file in /etc/ and comment the last line", and it takes three times longer if you have to add [gk]sudo before each command.

In fact, my idea was to have three OSs: Windows, a "normal" Linux distribution (Mint in this case), and a "Lightweight" Linux for hacking (and because I believe that a computer thousands of times more powerful than my older computer should be faster, not slightly slower). I'm probably going to go with Crunchbang, even though it doesn't recognize my graphics card, so I'll have to re-learn how Xorg drivers work again .

Update: well I had to reinstall it since installing the Nvidia driver broke it (reinstalling was faster than fixing it). Apparently Optimus is not good for Linux. Thank god for Bumblebee.

[Copper Bezel]

Well, fair enough, but it's still a very bad idea from a security standpoint and might actually cause some other unexpected problems, particularly on Mint, which is built on Ubuntu, which is written around the assumption that you're not using the root account for ordinary applications. It's probably a simpler solution to just make launchers for elevated Nautilus and gedit and work from there.

A lot of the difference in performance and system requirements you get between #! and Mint really comes from the desktop environment, so you might actually just use Mint with LXDE or a custom login session. Crunchbang uses Openbox and tint2 panel, IIRC.

A bit after Vista came out, so I would have ammunition in /. arguments about UAC, I kept a log of every UAC prompt I got for basically an entire month. The significant majority of them were triggered by actions where the analogous thing in Linux would also have required sudo.

My exposure to Windows 7 has been very limited, and my impression of the UAC comes largely from watching over a tech guy's shoulder at work while looking up tutorials for him on how to do his job. We were trying to remove application icons from the desktop on a kiosk-type machine and add something to Startup Applications for a particular non-admin user, there was a lot of logging in and out, and I think I blacked out at some point. I assumed that when people complained about the UAC, they were experiencing something rather like that. If we were really just experiencing a mix of his gross incompetence and my complete lack of familiarity with Windows, that wouldn't surprise me a jot. But it was a Windows 7 machine. In any case, my main point was the first one - that like it or no, UAC is actually a good idea.

[Iranon]

I concur. The problem isn't UAC, it's windows' tendency to require a little dance to show you are worthy of making your own decisions.

This attitude shows through pretty consistently. Examples: In a lot of locations, you can turn features the designers thought were a good idea (mostly security-related) on but not off.
Many setup screens feature greyed-out options, with a non-gimped but otherwise identical version hidden a little deeper (without any need for authentication along the way which would justify it).
Then there's labels being inconsistent and misleading, assuming you're just after the most pedestrian use. Harmless but annoying example: most appearance settings are only accessible after you click on "customize colors" somewhere along the way in some of the reasonable paths.

[lynkyn]

OSX and Other Distro (Gentoo). I use OSX a lot more, but I can't sleep at night without Linux installed.

[balthasar_s]

Computer #1(PC ~900MHz 512M ram 36GB HD):
1. Windows 98 SE
2. FreeDOS 1.0
3. Windows XP Professional SP 2

Computer #2 (PC Compaq 386/20e 386 20MHz 8MB ram 110MB HD):
1. MS-DOS 5.0
2. Windows for Workgroups 3.11

Computer #3 (Amiga 500 68000 ~7MHz 1MB chip ram 8MB fast ram ~1GB HD):
1. AmigaOS 1.3.3
2. AmigaOS 3.1

[AvatarIII]

Toshiba laptop: Windows 7 Home Premium

I have that.

maybe I'll get Linux one day, but for now Win 7 is ok for my needs

[sam_i_am]

at home i use windows 7. at work windows 8

[Giant Speck]

Toshiba laptop: Windows 7 Home Premium

I have that.

maybe I'll get Linux one day, but for now Win 7 is ok for my needs

Yeah, I don't use my Toshiba laptop anymore. It just sits in a drawer.

I use Linux Mint on my ASUS now. It's not bad. I certainly would not recommend Linux to everyone, though. I'm sure a lot might like it, but I'm just as sure that a lot of people would prefer Windows. Use what you know and what you like. I'm content with Linux because I don't do a lot with my computer; the only things I ever boot up Windows for anymore is iTunes and Adobe Photoshop Lightroom.

[Derek]

Computer #1(PC ~900MHz 512M ram 36GB HD):
1. Windows 98 SE
2. FreeDOS 1.0
3. Windows XP Professional SP 2

Computer #2 (PC Compaq 386/20e 386 20MHz 8MB ram 110MB HD):
1. MS-DOS 5.0
2. Windows for Workgroups 3.11

Computer #3 (Amiga 500 68000 ~7MHz 1MB chip ram 8MB fast ram ~1GB HD):
1. AmigaOS 1.3.3
2. AmigaOS 3.1

I'm afraid to ask which of these is your primary computer.

[cjmcjmcjmcjm]

Computer #1(PC ~900MHz 512M ram 36GB HD):
1. Windows 98 SE
2. FreeDOS 1.0
3. Windows XP Professional SP 2

Computer #2 (PC Compaq 386/20e 386 20MHz 8MB ram 110MB HD):
1. MS-DOS 5.0
2. Windows for Workgroups 3.11

Computer #3 (Amiga 500 68000 ~7MHz 1MB chip ram 8MB fast ram ~1GB HD):
1. AmigaOS 1.3.3
2. AmigaOS 3.1

I'm afraid to ask which of these is your primary computer.

Seriously, MS DOS 6.22 is the best one. Upgrade, already!

[boXd]

I'm thinking of going back to Arch Linux again. Don't want to do a dualboot, but so far every time I switch back to good ol' Linux I find out later that some important piece of software I use often no longer works. What kind of software, besides games, generally only works on Windows (or OSX even) again?

[hotaru]

What kind of software, besides games, generally only works on Windows (or OSX even) again?

metro apps only work on windows (and only on windows 8). that's all i can think of at the moment other than games that won't work with wine.