Like most free software, it's mostly portable, and does not depend on any non-free APIs. The sound engine is also abstracted, so if need be JACK could be rewritten.
Other dependancies, like GTK2 would need to be at least recompiled (and run through the X server).
However, nobody has actually done this. It's far simpler to just install Linux.
For those not in the know, I should prolly point out that Wine does not emulate any specific Windows version, it does whatever the app needs in order for it to run.
Still, there are lots of APIs in NT4 that Wine still does not support (primarily through underuse though).
Now, I understand the "security through obscurity" theory that basically says Mac's have far fewer virii problems than PCs because not nearly as many people use Macs, but that's sort of a dead idea nowadays.
Why? You realise how many computers there are in the world? Those numbers sound mighty impressive, but it still adds up to, perhaps, 2-3% of users? I know that I've only ever actually met 2 Mac users in my entire life, including geek friends, normal friends, business people etc.
So let's be reasonable - any worm that works by randomly firing off emails and hoping enough sticks simply will not propogate if it only targets Mac users (or Linux users, or NetBSD users or whatever). The goal of the worm is to propogate, in order to do that it has to be able to hit as many people as possible, and it's really that simple.
Anyway, to write a worm that targets the Mac, you'd have to own one. The Mac community is notoriously close, I can't really imagine anybody writing a virus for their chosen platform. Note that this doesn't really apply to Linux, there are plenty of Linux/UNIX worms/trojans etc lying around, simply because the loot is so much more attractive - Linux boxen tend statistically to be beasts sitting on the end of a phat pipe, so greed would overcome any affection for the platform.
Just take it as one of the few advantages of relative obscurity.
Those "bug reports" aren't extensive at all. They contain an extremely brief description of the problem, followed by a cause if you're lucky (many of them don't have causes apparently) then some boilerplate text which bulks it out.
Obviously open source projects vary. You have to look at the best of breed apps, Linux, XFree, GNOME, KDE, Mozilla etc. These will generally run public bug trackers, which you can query at will to get as much information as you need or desire. It's very handy to have. Short KB "articles" aren't really a replacement.
Gather 100,000 close friends? I'd love to meet somebody who's that popular!
Anyway, it'd be far far far far far easier to simply make it run under Wine. DirectX 9 support for WineHQ has started in the last few days, and I'd guess TransGaming will be supporting it soon also.
Of course, if the game is really boring and expensive, that might mean that nobody cares enough to make it work. But convincing a few Wine developers is a hell of a lot easier than a mega-corp on a budget.
The fact is, that I'd play a Star Wars game because I want to be in the movies. I don't give a wamprats ass about building up my character, completing quests and so on. I want to be a Jedi, I want to fly spaceships, I want to travel to exotic planets and so on.
The game that defined the Star Wars experience for me was Jedi Knight Dark Forces. Despite being primarily a first person shooter, it had most of those ingredients. I could be a Jedi AND use the force powers. I could travel to exotic locations. It even had a plot with some (albiet weak) character development!
I last played it (completed it) years ago, but I still remember many of the missions, settings and atmosphere. It had a quality sound track, classic Star Wars style, with the tracks tailored to each level. When you were creeping about infiltrating an Imperial base and trying not to get spotted, it was quiet and spooky. When you were trying to escape a Rebel base under attack from legions of stormtroopers, it was fast and frentic.
Even though the gameplay was basically shooting things, the world was epic enough that I felt I was in the movies.
There was only two problems. It was single player, and it was focussed too much on blasting stuff (well, that was the genre).
What I want is not a MMORPG. They bore me. Real life is boring. A poor fake of real life on Tatooine even more so. I want you to sell me a game like Jedi Knight, with a plot, with vast levels and worlds with convincing characters, that I can play with a group of friends. Not a big group. A small one is fine. We can get together online and play it in sessions, like how we all go round to a friends flat on Sunday nights to watch TV.
I want to be able to fly in space, and I'd like a stronger focus on convincing worlds and missions, and less on shooting things (though it should still have fights), ie the balance should be more like in the movies. We should be able to work together or compete as major characters. It should be like going to the movies together, except that we're in the movies together.
It should have a clear beginning, middle and end. When it's over, then you can sell us sequels. Like the films.
That's all I want. It's not all that hard. Jedi Knight wasn't hugely far off, and that was years ago. Why does nobody do this?
Nothing wrong with it as far as I'm concerned, and I've written my fair share of GPLd code (and will write much more in the upcoming years).
That's why this whole "looking at GPLd code will contaminate you" stuff is wrong IMHO. Copying code is a clear violation, it can be proven if necessary (given the balance of probabilities). OTOH taking a useful idea or technique cannot be.
If taking ideas or algorithms is violating the license, then what about code style? Working on others open sourced code has massively improved my coding style. Have I "taken" that too?
Personally, if I was to write some free software and somebody ripped my ideas off for a proprietary product, I might be pretty pissed off. But then again, they could have stolen that idea even if the code wasn't open. So the only thing they could steal is actual code, which would violate the license, or algorithms/constructs - to be frank, I'm not a super hot-shot coder, and the code I write tends to be simple and to the point. Fancy tricks are not my style. So that's not a risk either.
To sum up then, open vs closed code really doesn't make a huge amount of difference here.
Already, there is *significant* variance amongst different linux distros, even ignoring forking those.
Oh, I would argue that it's not so bad. Distros are normally entirely source code compatible and mostly binary compatible if you know what you're doing as well.
So yes, you cannot test on every distro out there. But if iD put up a Linux binary of Enemy Territory and say "only tested on Red Hat", do you honestly think a Slackware user is going to give a damn? If he downloads it and it runs, he'll be happy. If you know what you're doing, it's not hard to make binaries that are fairly portable.
Coming soon! Tools and documents from autopackage HQ showing you how to do this, explaining common pitfalls and making it much easier to do without special setups
Unlike like 99.9% of our community, you don't want to just throw some source up and hope that an *end user* can 'make' it.
(I edited the above line as the original didn't seem to make much sense....)
Like I said, I think you're overexaggurating. The main problems caused by incompatabilities between distributions are down to software installers and package management. There are only 3 real package managers in widespread use now, RPM, DPKG and emerge.
It's most certainly possible to abstract the remaining differences between a well designed API.
Personally, I think there may be some funky logic behind using some of the principles behind JCP.
I think you want the LSB. That's for when you need industrial strength compatability. For 99% of all software (including proprietary software) however, that simply isn't necessary.
Now think about how much further along Linux would have been if that time, money, and effort had not been squandered on dead ends
Dead ends and wastage are a part of life. It's not possible to plan everything, tell everybody what to do. It's like saying "just imagine how advanced ours cars would be if everybody had simply bought Ford".
Now think about how much time, money, and effort was spent on Gnome or KDE. Now think about how much further along Gnome or KDE could have been if nobody had wasted their time on the other one.
That split was unfortunately inevitable, and it will hopefully serve as a lesson to anybody that would start a major product while ignoring the philosophies that started it all.
Now think about Gecko. Gecko, as a browser technology, is essentially dead. KHTML, thanks to Apple, rules the day.
According to my desktop neutral but Linux based website (see sig), over 45% of my hits come from Mozilla. Only 8% come from Konq (and in fact it's normally lower, more like 4-5% on most days, but the figures do vary). Interestingly, Internet Explorer makes up the rest, presumably from people visiting at work. See here for details
So, in the real world, Gecko is most certainly not dead. "Dead" in this context would have to be determined by development speed and user base. Clearly, I have to cater to my users, almost all of which are on Mozilla based browsers.
Not really.... for starters there's no guarantee that an app will have an entire CPU to itself. Secondly, no OS I know of, not even MacOS X, takes several milliseconds to process network packets.
Anyway, bear in mind that the OS controls IO on the system, so your game of quake has to synchronize with the other parts of the system on the other CPU at some point.
Finally for audio mixing, that's why buffers are used. What matters in pro audio work is latency, ie the amount of time it takes for audio to travel through the system. That can affect things like synchronization of multiple audio streams and so on. Again multiple processors don't necessarily lead to lower latency, if anything, they can sometimes increase it.
The only real thing that would take CPU time there is video transformation, which is hardly a common desktop activity.
The other two wouldn't push usage above two or three percent, and there's little to no benefit in dual CPUs when your current one is multitasking just fine.
Of course they are boosted. But Apple didn't compare themselves against Dells own benchmarks. They did their own and crippled the Dells while they were at it.
THAT is the point of this article. You don't have to examine Dells numbers, you can already be assured that they had all optimizations enabled (but they are published on the spec website so they aren't just making them up either)
2) I don't know about you, but it is normal for me to be doing several things at once on my computer. Listening to music, downloading email, munging video, plus about a hundred background tasks. The OS itself balances these separate tasks between the processors, so there is a very real and significant advantage to the dual processor even if the individual programs don't take advantage.
That's true. However, those hundreds of backgrounds tasks are normally asleep. As an example, open up ten different desktop apps, run top or whatever and note that CPU usage is only a few percent. Those apps are blocking in an event loop, and until they receive events the kernel won't allocate them any timeslices.
Because of the way pre-emptive multitasking works however, having a dual CPU machine generally simply gives you more cycles to burn. You could get the same effect by buying a chip that's twice as fast - in fact, performance would be better as you don't have the overhead of the communication between the two CPUs.
So, this is useful if you spend a lot of your time doing very processor intensive things, because adding extra CPUs is generally easier than finding chips double the current speed of what you're using (assuming you're already on the cutting edge).
But, for most desktop users, it wouldn't make any difference, because no matter how many apps they have open, only a few of them will actually be doing any processing at any given time.
Well the Dell benchmarks are on the SPEC site itself, and more importantly Dell don't use them to try and make the competition look bad (or not in a way quite as obvious as Apple are).
So, the problem this guy has is not that Apple are tweaking their machines to get the best results, but that they didn't do the same for the competition. They then used that to make their products look better than they really are, which is lying.
No dude, he said that the 2nd processor wasn't used by the nature of the SPEC test. So even when running that test on a dual CPU machine, the 2nd CPU would have always been idle.
This is what really gets my goat sometimes. Calling him a "known troll" and saying "his credibility" is zero does not address his points.
Are you going to deny that Apple cheated at the benchmarks by disabling various optimizations on the competition? Are you going to deny that most software uses integer math, as one "software coder" clearly did (hint: i write a lot of software, and integer math practically always dominates)?
The guy may, or may not be a troll. However, the sheer amount hate mail, and the level of it, was stunning. What kind of people write stuff like that? Very few of them even attempted to address the guys points, and those that did made a hash job of it (nobody uses int math? wtf?).
The fact is that anybody outside the Mac community, having read that essay, is going to come away with a bad impression of said community. Nobody deserves to get hate mail like that for pointing out the other side of the statistics.
You want easy to administer, but Red Hat isn't cool enough right?
Whatever. I use Red Hat for that exact same reason, I like Linux but don't want to spend ages pissing about with config files. And you know what? I nearly never do. P
I can say, its not ever going to happen. Every single person I've ever talked to about it who believed otherwise hasn't used OSX.
Hi. I've used OS X quite a bit, and I believe that Linux is going to totally blow it away in ease of use, consistancy, looks, compatability and speed.
I know this, because I'm deep into the developer community now. I can see a few years ahead, can divine roughly where things are heading. I can see some of the cool stuff that's being worked on right now.
And the fact is, that you can't beat exponential scalability in developer resources.
Yes, consistancy. In some respects Linux is already more consistant than MacOS. Almost all apps in the latest versions of Red Hat are consistantly themed, with OpenOffice being the only real exception. Contrast that with MacOS, in which applications have two entirely different themes, applied at random for no good reason. Hell, Apple have violated their own HIG twice now simply to make things look shiny.
What kind of message does that send to developers?
In contrast, developers all over Linux land are busy learning about and getting the usability religion. When a developer or team doesn't co-operate, fork! It's happened several times already.
Oh and FWIW there are indeed UI experts working on Linux (well, gnome at any rate). So, I think you're being overly pessimistic.
I agree, but WHY is that? WHAT do Apple's programmers and designers have the Linux/GNOME programmers do not?
Numbers. Plain and simple. Me and Paul were chatting the other day, and I did some maths.
Best case scenario, full time Linux desktop hackers are outnumbered by full time Apple desktop hackers approximately 70:1
This comes from totalling up all the people that I know hack on mostly desktop related stuff at Red Hat, SuSE, Ximian, CrossOver, Sun, TrollTech etc and the number of people that apparently work on MacOS.
Obviously, volunteer time makes that a bit ratio slightly less harsh. But a lot of volunteer time is spent writing and perfecting the applications which Linux needs, but are then ported to other platforms as well. So, really the number of people hacking on the desktop itself is really quite low. High for an open source project, but low compared to Windows or the Mac.
People tend to forget the sizes were talking about here. Linux has been broke since day 1. Apple and Microsoft have several billion dollars. That makes a big difference.
I personally think this is one reason that Apple hasn't released a port of Quicktime for Linux. They're busy wooing Linux users, and doing it very successfully.
I think it's more to do with the fact that Linux media players have been able to play QuickTime movies for a long time now, and there's no real market for compressing stuff into a proprietary format that requires extra plugins to play.
Soon, there will be nothing cool that comes out for the Linux Desktop that doesn't soon run on the Mac.
Ignoring iD games, virtually all popular open source apps have Win32 ports and Mac ports now anyway. Not a big deal. Linux is still growing rapidly.
Yes, Linux will eat everything for lunch, simply because of the way it scales. Nobody, not Microsoft, not Apple, can compete with the rest of the world combined, which is essentially what they're up against here.
It's only a matter of time in fact before somebody takes Linux and does an Apple on it - producing quality laptops etc that run a highly polished and integrated form of Linux out of the box. When you have access to all the source, there's almost no limit to the things you can do. Even if the first company to try this screws up, there will always be others.
Slashdot Mirror
Like most free software, it's mostly portable, and does not depend on any non-free APIs. The sound engine is also abstracted, so if need be JACK could be rewritten.
Other dependancies, like GTK2 would need to be at least recompiled (and run through the X server).
However, nobody has actually done this. It's far simpler to just install Linux.
For those not in the know, I should prolly point out that Wine does not emulate any specific Windows version, it does whatever the app needs in order for it to run.
Still, there are lots of APIs in NT4 that Wine still does not support (primarily through underuse though).
s/dumb/innocent/
Why? You realise how many computers there are in the world? Those numbers sound mighty impressive, but it still adds up to, perhaps, 2-3% of users? I know that I've only ever actually met 2 Mac users in my entire life, including geek friends, normal friends, business people etc.
So let's be reasonable - any worm that works by randomly firing off emails and hoping enough sticks simply will not propogate if it only targets Mac users (or Linux users, or NetBSD users or whatever). The goal of the worm is to propogate, in order to do that it has to be able to hit as many people as possible, and it's really that simple.
Anyway, to write a worm that targets the Mac, you'd have to own one. The Mac community is notoriously close, I can't really imagine anybody writing a virus for their chosen platform. Note that this doesn't really apply to Linux, there are plenty of Linux/UNIX worms/trojans etc lying around, simply because the loot is so much more attractive - Linux boxen tend statistically to be beasts sitting on the end of a phat pipe, so greed would overcome any affection for the platform.
Just take it as one of the few advantages of relative obscurity.
Obviously open source projects vary. You have to look at the best of breed apps, Linux, XFree, GNOME, KDE, Mozilla etc. These will generally run public bug trackers, which you can query at will to get as much information as you need or desire. It's very handy to have. Short KB "articles" aren't really a replacement.
A common problem is the use of NTLM authentication on work proxies, or the inability to install new software.
There is also the issue of people requiring IE for their intranets, and a lot of people simply cannot be bothered using two browsers at once.
Anyway, it'd be far far far far far easier to simply make it run under Wine. DirectX 9 support for WineHQ has started in the last few days, and I'd guess TransGaming will be supporting it soon also.
Of course, if the game is really boring and expensive, that might mean that nobody cares enough to make it work. But convincing a few Wine developers is a hell of a lot easier than a mega-corp on a budget.
That would be the deciding factor to me.
The fact is, that I'd play a Star Wars game because I want to be in the movies. I don't give a wamprats ass about building up my character, completing quests and so on. I want to be a Jedi, I want to fly spaceships, I want to travel to exotic planets and so on.
The game that defined the Star Wars experience for me was Jedi Knight Dark Forces. Despite being primarily a first person shooter, it had most of those ingredients. I could be a Jedi AND use the force powers. I could travel to exotic locations. It even had a plot with some (albiet weak) character development!
I last played it (completed it) years ago, but I still remember many of the missions, settings and atmosphere. It had a quality sound track, classic Star Wars style, with the tracks tailored to each level. When you were creeping about infiltrating an Imperial base and trying not to get spotted, it was quiet and spooky. When you were trying to escape a Rebel base under attack from legions of stormtroopers, it was fast and frentic.
Even though the gameplay was basically shooting things, the world was epic enough that I felt I was in the movies.
There was only two problems. It was single player, and it was focussed too much on blasting stuff (well, that was the genre).
What I want is not a MMORPG. They bore me. Real life is boring. A poor fake of real life on Tatooine even more so. I want you to sell me a game like Jedi Knight, with a plot, with vast levels and worlds with convincing characters, that I can play with a group of friends. Not a big group. A small one is fine. We can get together online and play it in sessions, like how we all go round to a friends flat on Sunday nights to watch TV.
I want to be able to fly in space, and I'd like a stronger focus on convincing worlds and missions, and less on shooting things (though it should still have fights), ie the balance should be more like in the movies. We should be able to work together or compete as major characters. It should be like going to the movies together, except that we're in the movies together.
It should have a clear beginning, middle and end. When it's over, then you can sell us sequels. Like the films.
That's all I want. It's not all that hard. Jedi Knight wasn't hugely far off, and that was years ago. Why does nobody do this?
That's why this whole "looking at GPLd code will contaminate you" stuff is wrong IMHO. Copying code is a clear violation, it can be proven if necessary (given the balance of probabilities). OTOH taking a useful idea or technique cannot be.
If taking ideas or algorithms is violating the license, then what about code style? Working on others open sourced code has massively improved my coding style. Have I "taken" that too?
Personally, if I was to write some free software and somebody ripped my ideas off for a proprietary product, I might be pretty pissed off. But then again, they could have stolen that idea even if the code wasn't open. So the only thing they could steal is actual code, which would violate the license, or algorithms/constructs - to be frank, I'm not a super hot-shot coder, and the code I write tends to be simple and to the point. Fancy tricks are not my style. So that's not a risk either.
To sum up then, open vs closed code really doesn't make a huge amount of difference here.
Oh, I would argue that it's not so bad. Distros are normally entirely source code compatible and mostly binary compatible if you know what you're doing as well.
So yes, you cannot test on every distro out there. But if iD put up a Linux binary of Enemy Territory and say "only tested on Red Hat", do you honestly think a Slackware user is going to give a damn? If he downloads it and it runs, he'll be happy. If you know what you're doing, it's not hard to make binaries that are fairly portable.
Coming soon! Tools and documents from autopackage HQ showing you how to do this, explaining common pitfalls and making it much easier to do without special setups
Unlike like 99.9% of our community, you don't want to just throw some source up and hope that an *end user* can 'make' it.
(I edited the above line as the original didn't seem to make much sense....)
Like I said, I think you're overexaggurating. The main problems caused by incompatabilities between distributions are down to software installers and package management. There are only 3 real package managers in widespread use now, RPM, DPKG and emerge.
It's most certainly possible to abstract the remaining differences between a well designed API.
Personally, I think there may be some funky logic behind using some of the principles behind JCP.
I think you want the LSB. That's for when you need industrial strength compatability. For 99% of all software (including proprietary software) however, that simply isn't necessary.
Dead ends and wastage are a part of life. It's not possible to plan everything, tell everybody what to do. It's like saying "just imagine how advanced ours cars would be if everybody had simply bought Ford".
Now think about how much time, money, and effort was spent on Gnome or KDE. Now think about how much further along Gnome or KDE could have been if nobody had wasted their time on the other one.
That split was unfortunately inevitable, and it will hopefully serve as a lesson to anybody that would start a major product while ignoring the philosophies that started it all.
Now think about Gecko. Gecko, as a browser technology, is essentially dead. KHTML, thanks to Apple, rules the day.
According to my desktop neutral but Linux based website (see sig), over 45% of my hits come from Mozilla. Only 8% come from Konq (and in fact it's normally lower, more like 4-5% on most days, but the figures do vary). Interestingly, Internet Explorer makes up the rest, presumably from people visiting at work. See here for details
So, in the real world, Gecko is most certainly not dead. "Dead" in this context would have to be determined by development speed and user base. Clearly, I have to cater to my users, almost all of which are on Mozilla based browsers.
Anyway, bear in mind that the OS controls IO on the system, so your game of quake has to synchronize with the other parts of the system on the other CPU at some point.
Finally for audio mixing, that's why buffers are used. What matters in pro audio work is latency, ie the amount of time it takes for audio to travel through the system. That can affect things like synchronization of multiple audio streams and so on. Again multiple processors don't necessarily lead to lower latency, if anything, they can sometimes increase it.
The other two wouldn't push usage above two or three percent, and there's little to no benefit in dual CPUs when your current one is multitasking just fine.
THAT is the point of this article. You don't have to examine Dells numbers, you can already be assured that they had all optimizations enabled (but they are published on the spec website so they aren't just making them up either)
You can run MS Office and Photoshop via CrossOver. Which does indeed use antialiasing.
That's true. However, those hundreds of backgrounds tasks are normally asleep. As an example, open up ten different desktop apps, run top or whatever and note that CPU usage is only a few percent. Those apps are blocking in an event loop, and until they receive events the kernel won't allocate them any timeslices.
Because of the way pre-emptive multitasking works however, having a dual CPU machine generally simply gives you more cycles to burn. You could get the same effect by buying a chip that's twice as fast - in fact, performance would be better as you don't have the overhead of the communication between the two CPUs.
So, this is useful if you spend a lot of your time doing very processor intensive things, because adding extra CPUs is generally easier than finding chips double the current speed of what you're using (assuming you're already on the cutting edge).
But, for most desktop users, it wouldn't make any difference, because no matter how many apps they have open, only a few of them will actually be doing any processing at any given time.
So, the problem this guy has is not that Apple are tweaking their machines to get the best results, but that they didn't do the same for the competition. They then used that to make their products look better than they really are, which is lying.
No dude, he said that the 2nd processor wasn't used by the nature of the SPEC test. So even when running that test on a dual CPU machine, the 2nd CPU would have always been idle.
I mean by that logic we should never criticise Microsoft because they gave the world Office, or something.
Are you going to deny that Apple cheated at the benchmarks by disabling various optimizations on the competition? Are you going to deny that most software uses integer math, as one "software coder" clearly did (hint: i write a lot of software, and integer math practically always dominates)?
The guy may, or may not be a troll. However, the sheer amount hate mail, and the level of it, was stunning. What kind of people write stuff like that? Very few of them even attempted to address the guys points, and those that did made a hash job of it (nobody uses int math? wtf?).
The fact is that anybody outside the Mac community, having read that essay, is going to come away with a bad impression of said community. Nobody deserves to get hate mail like that for pointing out the other side of the statistics.
You want easy to administer, but Red Hat isn't cool enough right?
Whatever. I use Red Hat for that exact same reason, I like Linux but don't want to spend ages pissing about with config files. And you know what? I nearly never do. P
Hi. I've used OS X quite a bit, and I believe that Linux is going to totally blow it away in ease of use, consistancy, looks, compatability and speed.
I know this, because I'm deep into the developer community now. I can see a few years ahead, can divine roughly where things are heading. I can see some of the cool stuff that's being worked on right now.
And the fact is, that you can't beat exponential scalability in developer resources.
Yes, consistancy. In some respects Linux is already more consistant than MacOS. Almost all apps in the latest versions of Red Hat are consistantly themed, with OpenOffice being the only real exception. Contrast that with MacOS, in which applications have two entirely different themes, applied at random for no good reason. Hell, Apple have violated their own HIG twice now simply to make things look shiny.
What kind of message does that send to developers?
In contrast, developers all over Linux land are busy learning about and getting the usability religion. When a developer or team doesn't co-operate, fork! It's happened several times already.
Oh and FWIW there are indeed UI experts working on Linux (well, gnome at any rate). So, I think you're being overly pessimistic.
Numbers. Plain and simple. Me and Paul were chatting the other day, and I did some maths.
Best case scenario, full time Linux desktop hackers are outnumbered by full time Apple desktop hackers approximately 70:1
This comes from totalling up all the people that I know hack on mostly desktop related stuff at Red Hat, SuSE, Ximian, CrossOver, Sun, TrollTech etc and the number of people that apparently work on MacOS.
Obviously, volunteer time makes that a bit ratio slightly less harsh. But a lot of volunteer time is spent writing and perfecting the applications which Linux needs, but are then ported to other platforms as well. So, really the number of people hacking on the desktop itself is really quite low. High for an open source project, but low compared to Windows or the Mac.
People tend to forget the sizes were talking about here. Linux has been broke since day 1. Apple and Microsoft have several billion dollars. That makes a big difference.
I think it's more to do with the fact that Linux media players have been able to play QuickTime movies for a long time now, and there's no real market for compressing stuff into a proprietary format that requires extra plugins to play.
Ignoring iD games, virtually all popular open source apps have Win32 ports and Mac ports now anyway. Not a big deal. Linux is still growing rapidly.
Yes, Linux will eat everything for lunch, simply because of the way it scales. Nobody, not Microsoft, not Apple, can compete with the rest of the world combined, which is essentially what they're up against here.
It's only a matter of time in fact before somebody takes Linux and does an Apple on it - producing quality laptops etc that run a highly polished and integrated form of Linux out of the box. When you have access to all the source, there's almost no limit to the things you can do. Even if the first company to try this screws up, there will always be others.