Developers who scream at UI People "I don't believe any of that cognitive psychology crap. That's what you want, that's not what I want!"
Users who do the same.
The feeling in the linux development community that designing user interfaces is somehow less work, less important, and less worthy than writing technical code.
The belief by the linux development community that if you don't have patches to submit regarding your UI ideas, you're merely whining and refusing to contribute.
The linux community's outright refusal to believe that linux software generally has severe usability problems. The first step in solving a problem is to admit you have a problem, and that's not going to happen anytime soon
Linux zealots who run around screaming that linux is perfectly ready for the desktop and anyone who says otherwise is merely spreading FUD about linux being hard to use.
An incredibly high tolerance for badly designed interfaces you will naturally find in most linux geeks. In fact, they pride themselves on being able to suffer/adapt to bad UI's. This partially explains the point above.
The "It's so Purty" Problem:the false belief that an aesthetically pleasing interface is one that is usable. You often see this fallacy crop up around discussions of the Zaurus UI, where the fact that the Zaurus UI is prettier than that of the Palm overshadows the fact that a Palm requires a minimum of one stylus tap to enter a date, whereas the Zaurus requires a minimum of four to do the same.
The attitude that if a person has trouble understanding how to use a UI that is poorly designed, ambiguous, and confusing, it is because "they don't want to learn".
I often say that Bill Gates doesn't have to lift a finger to crush linux on the desktop because so many people in the linux community do his job for him.
The single menubar at the top of the screen is faster to access because of an ergonomic principle known as Fitts' Law, which states that the time to access a target is a function of the target's distance and size. A at the top menubar is infinitely large because there is no possible way to overshoot it vertically (i.e. you can slam the mouse up to the top of the screen really fast and don't have to correct for any vertical error because you're running into the top of the screen). On the Windows/GNOME/KDE interfaces it's possible to vertically overshoot the menubar, which makes that layout far less usable than the layout on a mac. For a more in-depth explanation of this phenomena, check out this article by UI guru Bruce Tognazinni.
If its size doesn't satisfy my secretary, I'd have to go to great pain and expense to get it enlarged. And adding another 6 inches of plastic to what's already there would look kind of weird.
Teenage intruder: See? I run nmap 234.34.53.5 and I get a list of all the ports that are open on their machine. I can then do some other stuff with libpcap...
Brokaw: Wardrobe!....dammit, get this kid a large sleek trenchcoat, combat boots, and a pair of those $300 designer sunglasses. They're expecting neo, not urkel. Audio!...cue that "techno" music they listen to. (to "hacker")Okay, kid, your motivation is to disrupt The System, bring down The Corporate Machine that runs the government, and then make it with Carrie Ann Moss in a hovercraft.
Teenage intruder: But I just thought I would show you how I learned about this network vulnerability in my quest for knowl....
Brokaw: (to cameraman) Start rolling in five, four, three, two...
If you caught spreading a computer virus, the candian government would cram you in a jail with 11 other annoying "big house"-mates, put you on canadian televion for 24 hours a day, and the last person to get anally violated would get $500,000.
The mac menubar does have faster access times, because of Fitts' Law. Check out this article by UI design guru bruce tognazzini. Basically, interface elements located on screen edges have faster access times with a mouse because you can't overshoot.
Why do so many interfaces have the Windows-style menus? GNOME has them because that's what KDE did. KDE did them because that's what microsoft did. Microsoft did them because of their usual incompetance at designing UI's (M$ office adaptive menu's, multi-row tabs, need I say more) and because in they didn't want to get sued by Apple. Fat lot of good that did them, as Apple filed a lawsuit over the UI anyways.
Recently, Tim O'reilly wrote a piece on the growing politicization of open source. The software industry has already been politicized by Microsoft. We already have an IT purchasing system where merit has been passed over for political expediency. Quoteth The Who, Microsoft "decided the shotgun sings the song". With government IT spending already politicized, Open Source is merely playing by the rules of the game.
On the palm's datebook, in the single-day view I can tap my stylus on a line representing a time, and with a single tap I've focused the line and selected the time and I can immediately start writing down my appointment. One single tap and immediately I can do my work.
On the zaurus' datebook, in the single day view I can't click on the lines to immediately start entering the time. And in fact, they're not even lines, they're thinly disguised Windows/KDE text fields (Which is stupid, because the palm's use of lined text fields makes the whole interface feel more natural, like paper. Paper is the most vicious and ruthless ane efficient competitor of any PDA, but I digress).
In order to create the new appointment on the Zaurus I have to go upward (BTW, did I mention people in western cultures read from up to down), to the document icon to create a new date entry, which only further proves that trolltech tried to carbon copy a desktop interface into a PDA one. This is far less intuitive then merely placing your stylus on the line and starting to write like on the Palm. I'll admit the palm does have a button that brings up a time entry dialog (if you choose to do it that way. But you're not forced to, like on the Zaurus), but the button on the palm to do this is not some ambiguous-looking unlabled icon like on the Zaurus but a run of the mill button with the labell "new".. No guessing or risk-taking required. And the "new" button on the Palm is located under the time slots, not above like on the Zaurus, so the Palm keeps in step with the way people read things.
Guess what happens once I get the Zaurus' modal dialog for a entering a new date? The text fields for 'Description' and 'Location' are already filled with "(None)" and "(Unknown)". Why are they already filled? I haven't even decided to write anything in them yet! And if a field is left blank, isn't it already obvious that the value is unknown? What's worse, I have to erase the crap in these text fields in order to write my own stuff in them. Users always expect text fields to only contain stuff that they have actively written in them, and Trolltech has greatly violated that expectation by automatically stuff into them.
To focus the "Description" text field for a Zaurus event I have to tap on it yet again, whereas the palm focused the description field when I selected the time.
To finish off entering a date on the Zaurus I have to go up to the right hand corner of the titlebar (again, the bizarre 'down to up' fetish) and hit "Ok". On a palm, there is no hitting "Ok". You don't need to confirm an entry, the entry is merely just as you left it. The PalmOS UI is stateless, which is the direction that many of the foremost HCI professionals say we should be going in. And stateless UI lends itself very well to devices where everything in stored in ram.
Oh, I almost forgot, on the Zaurus' datebook, not everything fits within the width of the screen (including the widgets). I have to scroll horizontally to manipulate widgets on the edge. On the palm, no horizontal scrolling required. Every single widget fits precisely on the screen.
Tap summary:
At minimum, entering a date on PalmOS requires only one tap of the stylus to select a time, focus the text area for writing Total minimum # of taps requred: 1
At minimum, the zaurus requires 1 tap to create new time, 1 tap to select starting time, 1 tap to focus the event description text field, and 1 tap to confirm the date entry. Total minimum # of taps required: 4.
I can understand why you like the Zaurus' keyboard--it doesn't handle the stylus half as efficiently as the Palm does.
About my disclaimer:
The first step in solving a problem is to admit that you have one. My point: Right Now, most of the linux community absolutely refuses to admit that they have usability problems. According to them, we don't have usability problems, we just have people spreading FUD about linux being hard to use. If anyone with any interface design background points out these problems, they are immediately met with incredulity and some really nasty denials.
When I say "this product has severe usability problems because we've got so many linux zealots denying that there is a usability problem" and then some linux zealot starts screaming "Bullsh*t, there is no usability problems!", the linux zealot has effectively proven my point far better than I ever could.
I got a zaurus about 3 months ago. In some ways, it really points out both the incredible strengths of linux PDA's and the incredible weaknesses.
The really awesome thing about the zaurus is that you have the power to do things like create GUI based apps in python, which allows you to rapidly create useful mobile apps and not have to worry about things like memory management and cross compilation. It really takes the edge off of doing PDA development. I wrote a program to keep track of what I eat making use of python + xml + qt. There is no chance in hell I could have done that with a 33mhz dragonball.
The problem with the zaurus is that there are very serious usability problems with the zaurus UI. Mossberg was absolutely 100% right when he said the zaurus is hard to use. Trolltech (with Qt embedded) and sharp (with their hardware design) has given absolutely no thought to making their products usable. Most of the glowing reviews of the zaurus interface you will see are given by entrenched techies who pride themselves on being to stumble their way through badly designed interfaces and decry anyone who finds a UI confusing or ambiguous as "not wanting to learn". Or they equate the zaurus' aesthetic beauty with usability (again, the "purty == usable" stupidity we see in so much Free Software these days). The reason why the Palm UI gets such a glowing review is that palm creator Jeff Hawkins designed the interface (both hardware-wise and software-wise) before he ever wrote a line of code or manufactured the hardware. He crafted a block of wood in the shape of the palm, whittled down a chopstick for a stylus, and carried the mockup around with him everywhere he went to ponder how the Palm UI should be designed. In the HCI world, we call this "preliminary task analysis", and it's obviously something that sharp and trolltech haven't done in the slightest. To read more about why the PalmOS UI still kicks the crap out of the zaurus UI, read the Zen of Palm and then try to find a document of equivalent enlightenment and quality relating to the zaurus.
For developing custom applications very quickly, the zaurus kicks booty, but it would be irresponsible to suggest to someone the zaurus as something you could use for phone numbers or schedules, no matter how many other cool features it has.
Note: Right now, the linux community is in hardcore denial about usability problems in general, and any attempt to deny the truth of this post only further proves the truth of it.
The reason why the mouse requires so much effort on Windows is that Microsoft (and by extension, most windows programmers) make UI's that take a lot of power away from the mouse.
The example most relevant to your post is the pull-down "File" menu. When they copied apple (or tried to), microsoft changed the location of the pull-down menu bar from the top of the screen (like on a mac) to the window of each respective application. With Apple's way, you can't possibly vertically overshoot the menu bar; with Microsoft's way, not only is it possible to overshoot the menubar horizontally, but you have to watch out for overshooting the menubar vertically as well. Putting it simply, a menu at the top of a screen has faster mouse access times than a menu on a window. I know it sounds counterintuitive, but it is a result of something called Fitts' Law, which states that the time to access a target is a function of the target's distance and it's size. For more information on Fitts' Law, check out this article on usability guru Bruce Tognizzini's website.
I think it's a neat idea, but you have the problem of people's haptic abilities (i.e. sense of touch) worsening as they get older. A touch-driven interface might really suck for some elderly person already trying to get a grip on computers in general.
Not that interfaces that use sight or sound will be invulnerable to aging-related isses, but it is something to keep in mind.
Wrong. Microsoft is incompetant at designing GUI's
on
GUIs for Everyone
·
· Score: 2, Insightful
If the user is typing something really important in an IE text field, and then all of a sudden the text field loses focus and they hit the backspace button (thinking they're doing a backspace in a text field) guess what happens? They go back to a previous page and in many cases the text field they have been typing in get's completely blanked out when this happens. We UI designers would typically call this an "unexpected action". The user expected hitting backspace by itself would do a backspace in the text, and instead it brought them to a previous page. And wiped out all the valuable work they had done in the process.
Other examples of microsoft incompetance include window-in-window MDI, multi-row tabs, and their latest shennanigan, the adaptive menus that constantly change position on a user (which screws up the users motor muscle memory for where the menu selections are). All these "features" have been harshly criticized by many in the HCI community.
For further reading, check out the Interface Hall of Shame, of which Microsoft is the most frequent inductee. To see Microsoft usability get slammed by one of the most prominent members of the UI design community, check out AskTog.com
Microsoft is so successful in the UI biz despite their poor usability for precisely the same reason they are so successful in the server biz despite their poor security: they've got a monopoly, a proprietary file format, and the ears, hearts and minds of every pointy-haired boss and every clueless IT manager in America.
What I think a lot of people don't understand about desktop linux is that there are many different stupid factions fighting over equally stupid ideas.
One camp of idiots in the linux community can't seem to understand how the basic file/folder/desktop paradigm works. They says things like "don't call what people put files into a 'folder'. That's too much like windows. Call it 'directory'."
That any HCI professional would say "folder keeps consistancy with the paradigm" is irrelevant to idiots such as these.
Another camp of idiots, the polar opposite of the first, wants to blindly carbon-copy microsoft. They can't seem to understand that many designs that Microsoft came up with were dead wrong from a GUI design perspective and that Microsoft has been constantly criticized by the HCI community for implementing them in the first place. Multi-row tabs, window-in-window MDI, billions of tiny, cryptic, unlabeled toolbar buttons that are too small to have fast access time with a mouse are just some of the many skeletons in microsofts UI design closet. That an HCI professional would say "adaptive menus like the kind in office 2000 can be easily proven to be a stupid idea" is irrelevant to idiots such as these.
Keep in mind that the mac development has had a long tradition of making usable, user friendly software. Many of the programmers in the mac development community are meticulous about designing their interfaces, and they understand that the will catch hell from their userbase if they put out a crappy, inconsistant UI that was developed for the software as an afterthought. Finally, the mac until recently has not had a command line (excepting some stuff in MPW), which has forced mac developers to think in a totally graphic way. They are coding for a desktop, but it is not a "Unix Desktop", it is a "Mac Desktop that Just Happens To Use Unix".
Linux programmers, on the other hand, come from a development community that has had a 30 year long tradition of calling end users stupid and telling them to go read the fine manual (which usually is anything but fine). There's an attitude that if an end user cannot perform a task that is confusing or ambiguous, it is because they "don't want to learn". Much of the linux development community thinks that HCI is a BS field of study and that usability design principles are sheer nonsense. Unlike in the mac community where deriding unusable software is widely accepted, in the linux community criticizing software usability is scorned by developers. It's referring to as "bitching" or "whining". Users who complain about badly designed interfaces are either told to shut up an code a better one or to "quit whining about what you get for free" or to "Stop spreading FUD about linux being hard-to-use". And finally, linux developers are severely tainted by their command-line heritage. Many interfaces they cook up have command line ways of thinking that just don't work on the desktop. Desktop linux will always crawl miles behind OS X if it's developers can't come out of their shell (pun intended).
Nothing prevents linux from taking over the desktop. Other than its entire developer community.
The first step in solving a problem is to admit that you have one
Slashdot Mirror
I often say that Bill Gates doesn't have to lift a finger to crush linux on the desktop because so many people in the linux community do his job for him.
Now Bob Zubrin will cite unlimited bandwidth as a reason to go to mars.
The single menubar at the top of the screen is faster to access because of an ergonomic principle known as Fitts' Law, which states that the time to access a target is a function of the target's distance and size. A at the top menubar is infinitely large because there is no possible way to overshoot it vertically (i.e. you can slam the mouse up to the top of the screen really fast and don't have to correct for any vertical error because you're running into the top of the screen). On the Windows/GNOME/KDE interfaces it's possible to vertically overshoot the menubar, which makes that layout far less usable than the layout on a mac. For a more in-depth explanation of this phenomena, check out this article by UI guru Bruce Tognazinni.
If its size doesn't satisfy my secretary, I'd have to go to great pain and expense to get it enlarged. And adding another 6 inches of plastic to what's already there would look kind of weird.
Two people in the dairy section of a chinese supermarket....
Chinsese shopper #1 (looking at a milk carton) I didn't know google had their own brand of milk.
Chinese shopper #2: It's not a brand logo, it's a "missing" picture.
The venture capitalists and wild stock market speculation beat them to it.
Does the warranty expire on December 21, 2012?
Teenage intruder: See? I run nmap 234.34.53.5 and I get a list of all the ports that are open on their machine. I can then do some other stuff with libpcap...
Brokaw: Wardrobe!....dammit, get this kid a large sleek trenchcoat, combat boots, and a pair of those $300 designer sunglasses. They're expecting neo, not urkel. Audio!...cue that "techno" music they listen to. (to "hacker")Okay, kid, your motivation is to disrupt The System, bring down The Corporate Machine that runs the government, and then make it with Carrie Ann Moss in a hovercraft.
Teenage intruder: But I just thought I would show you how I learned about this network vulnerability in my quest for knowl....
Brokaw: (to cameraman) Start rolling in five, four, three, two...
You probably have to patch your home and rebuild it before you can install that bidet module. I don't think many consumers would go for that.
If you caught spreading a computer virus, the candian government would cram you in a jail with 11 other annoying "big house"-mates, put you on canadian televion for 24 hours a day, and the last person to get anally violated would get $500,000.
The Army suddenly realizes that the string of text "b3 411 7h47 U c4n b3" on its recruitment site was not, in fact, an error message.
The mac menubar does have faster access times, because of Fitts' Law. Check out this article by UI design guru bruce tognazzini. Basically, interface elements located on screen edges have faster access times with a mouse because you can't overshoot.
Why do so many interfaces have the Windows-style menus? GNOME has them because that's what KDE did. KDE did them because that's what microsoft did. Microsoft did them because of their usual incompetance at designing UI's (M$ office adaptive menu's, multi-row tabs, need I say more) and because in they didn't want to get sued by Apple. Fat lot of good that did them, as Apple filed a lawsuit over the UI anyways.
If I had a cut figure, was well rounded, and stayed hard for days, I'd be a girl's best friend too.
But a 3G wireless phone might one day replace your microwave oven. Assuming you have enough talk time to cook both sides of the terriyaki.
Recently, Tim O'reilly wrote a piece on the growing politicization of open source. The software industry has already been politicized by Microsoft. We already have an IT purchasing system where merit has been passed over for political expediency. Quoteth The Who, Microsoft "decided the shotgun sings the song". With government IT spending already politicized, Open Source is merely playing by the rules of the game.
On the zaurus' datebook, in the single day view I can't click on the lines to immediately start entering the time. And in fact, they're not even lines, they're thinly disguised Windows/KDE text fields (Which is stupid, because the palm's use of lined text fields makes the whole interface feel more natural, like paper. Paper is the most vicious and ruthless ane efficient competitor of any PDA, but I digress).
In order to create the new appointment on the Zaurus I have to go upward (BTW, did I mention people in western cultures read from up to down), to the document icon to create a new date entry, which only further proves that trolltech tried to carbon copy a desktop interface into a PDA one. This is far less intuitive then merely placing your stylus on the line and starting to write like on the Palm. I'll admit the palm does have a button that brings up a time entry dialog (if you choose to do it that way. But you're not forced to, like on the Zaurus), but the button on the palm to do this is not some ambiguous-looking unlabled icon like on the Zaurus but a run of the mill button with the labell "new".. No guessing or risk-taking required. And the "new" button on the Palm is located under the time slots, not above like on the Zaurus, so the Palm keeps in step with the way people read things.
Guess what happens once I get the Zaurus' modal dialog for a entering a new date? The text fields for 'Description' and 'Location' are already filled with "(None)" and "(Unknown)". Why are they already filled? I haven't even decided to write anything in them yet! And if a field is left blank, isn't it already obvious that the value is unknown? What's worse, I have to erase the crap in these text fields in order to write my own stuff in them. Users always expect text fields to only contain stuff that they have actively written in them, and Trolltech has greatly violated that expectation by automatically stuff into them.
To focus the "Description" text field for a Zaurus event I have to tap on it yet again, whereas the palm focused the description field when I selected the time.
To finish off entering a date on the Zaurus I have to go up to the right hand corner of the titlebar (again, the bizarre 'down to up' fetish) and hit "Ok". On a palm, there is no hitting "Ok". You don't need to confirm an entry, the entry is merely just as you left it. The PalmOS UI is stateless, which is the direction that many of the foremost HCI professionals say we should be going in. And stateless UI lends itself very well to devices where everything in stored in ram.
Oh, I almost forgot, on the Zaurus' datebook, not everything fits within the width of the screen (including the widgets). I have to scroll horizontally to manipulate widgets on the edge. On the palm, no horizontal scrolling required. Every single widget fits precisely on the screen.
Tap summary:
Total minimum # of taps requred: 1
Total minimum # of taps required: 4.
I can understand why you like the Zaurus' keyboard--it doesn't handle the stylus half as efficiently as the Palm does.
About my disclaimer:
The first step in solving a problem is to admit that you have one. My point: Right Now, most of the linux community absolutely refuses to admit that they have usability problems. According to them, we don't have usability problems, we just have people spreading FUD about linux being hard to use. If anyone with any interface design background points out these problems, they are immediately met with incredulity and some really nasty denials.
When I say "this product has severe usability problems because we've got so many linux zealots denying that there is a usability problem" and then some linux zealot starts screaming "Bullsh*t, there is no usability problems!", the linux zealot has effectively proven my point far better than I ever could.
For developing custom applications very quickly, the zaurus kicks booty, but it would be irresponsible to suggest to someone the zaurus as something you could use for phone numbers or schedules, no matter how many other cool features it has.
Note: Right now, the linux community is in hardcore denial about usability problems in general, and any attempt to deny the truth of this post only further proves the truth of it.
Or you could just buy a pack of dobermans to guard your server farm and place fire hydrant stickers on your cpu's.
The reason why the mouse requires so much effort on Windows is that Microsoft (and by extension, most windows programmers) make UI's that take a lot of power away from the mouse.
The example most relevant to your post is the pull-down "File" menu. When they copied apple (or tried to), microsoft changed the location of the pull-down menu bar from the top of the screen (like on a mac) to the window of each respective application. With Apple's way, you can't possibly vertically overshoot the menu bar; with Microsoft's way, not only is it possible to overshoot the menubar horizontally, but you have to watch out for overshooting the menubar vertically as well. Putting it simply, a menu at the top of a screen has faster mouse access times than a menu on a window. I know it sounds counterintuitive, but it is a result of something called Fitts' Law, which states that the time to access a target is a function of the target's distance and it's size. For more information on Fitts' Law, check out this article on usability guru Bruce Tognizzini's website.
I think it's a neat idea, but you have the problem of people's haptic abilities (i.e. sense of touch) worsening as they get older. A touch-driven interface might really suck for some elderly person already trying to get a grip on computers in general.
Not that interfaces that use sight or sound will be invulnerable to aging-related isses, but it is something to keep in mind.
Possession is 11/10 of the law.
If the user is typing something really important in an IE text field, and then all of a sudden the text field loses focus and they hit the backspace button (thinking they're doing a backspace in a text field) guess what happens? They go back to a previous page and in many cases the text field they have been typing in get's completely blanked out when this happens. We UI designers would typically call this an "unexpected action". The user expected hitting backspace by itself would do a backspace in the text, and instead it brought them to a previous page. And wiped out all the valuable work they had done in the process.
Other examples of microsoft incompetance include window-in-window MDI, multi-row tabs, and their latest shennanigan, the adaptive menus that constantly change position on a user (which screws up the users motor muscle memory for where the menu selections are). All these "features" have been harshly criticized by many in the HCI community.
For further reading, check out the Interface Hall of Shame, of which Microsoft is the most frequent inductee.
To see Microsoft usability get slammed by one of the most prominent members of the UI design community, check out AskTog.com
Microsoft is so successful in the UI biz despite their poor usability for precisely the same reason they are so successful in the server biz despite their poor security: they've got a monopoly, a proprietary file format, and the ears, hearts and minds of every pointy-haired boss and every clueless IT manager in America.
- One camp of idiots in the linux community can't seem to understand how the basic file/folder/desktop paradigm works. They says things like "don't call what people put files into a 'folder'. That's too much like windows. Call it 'directory'."
That any HCI professional would say "folder keeps consistancy with the paradigm" is irrelevant to idiots such as these.
- Another camp of idiots, the polar opposite of the first, wants to blindly carbon-copy microsoft. They can't seem to understand that many designs that Microsoft came up with were dead wrong from a GUI design perspective and that Microsoft has been constantly criticized by the HCI community for implementing them in the first place. Multi-row tabs, window-in-window MDI, billions of tiny, cryptic, unlabeled toolbar buttons that are too small to have fast access time with a mouse are just some of the many skeletons in microsofts UI design closet. That an HCI professional would say "adaptive menus like the kind in office 2000 can be easily proven to be a stupid idea" is irrelevant to idiots such as these.
Clowns to the left of me, jokers to the right..."Warning: through sound and motion you might accidently paralyse nerves, shatter bones, set fires, suffocate an enemy or burst his organs."
Keep in mind that the mac development has had a long tradition of making usable, user friendly software. Many of the programmers in the mac development community are meticulous about designing their interfaces, and they understand that the will catch hell from their userbase if they put out a crappy, inconsistant UI that was developed for the software as an afterthought. Finally, the mac until recently has not had a command line (excepting some stuff in MPW), which has forced mac developers to think in a totally graphic way. They are coding for a desktop, but it is not a "Unix Desktop", it is a "Mac Desktop that Just Happens To Use Unix".
Linux programmers, on the other hand, come from a development community that has had a 30 year long tradition of calling end users stupid and telling them to go read the fine manual (which usually is anything but fine). There's an attitude that if an end user cannot perform a task that is confusing or ambiguous, it is because they "don't want to learn". Much of the linux development community thinks that HCI is a BS field of study and that usability design principles are sheer nonsense. Unlike in the mac community where deriding unusable software is widely accepted, in the linux community criticizing software usability is scorned by developers. It's referring to as "bitching" or "whining". Users who complain about badly designed interfaces are either told to shut up an code a better one or to "quit whining about what you get for free" or to "Stop spreading FUD about linux being hard-to-use". And finally, linux developers are severely tainted by their command-line heritage. Many interfaces they cook up have command line ways of thinking that just don't work on the desktop. Desktop linux will always crawl miles behind OS X if it's developers can't come out of their shell (pun intended).
Nothing prevents linux from taking over the desktop. Other than its entire developer community.
The first step in solving a problem is to admit that you have one