Re:Hostile takeover of SCO
on
SCO Roundup
·
· Score: 2, Informative
How much would a hostile takeover of SCO cost?
It would be impossible. As has been repeatedly detailed here, they are 95% privately held. The shares you would need to buy to take them over simply aren't available on the market.
yeah, but why should this happen? i know why it does happen, but why should it?
Well, its pretty simple. gcc had some standards-compliance and performance issues. In order to address them, they made some changes which broke binary compatibility with older versions. Every linux distribution switches to the newer version, because, well, its better. Most third-party binary packages, like Mozilla, also switch to match the evolving standard Linux platform. This is overwhelmingly positive, except for some barely-maintained proprietary software whose developers can't manage to recompile it every few years to keep up with the times.
It would be a major impediment to the development of the platform if it was required to keep the kind of binary compatibility that would be necessary to keep ancient proprietary plugins working. What if the last release of RealPlayer was compiled with gcc 2.7, or required libc5, or used a.out libraries? Would you expect Mozilla to base their releases on those ancient technologies? I hope not. If you really want to use realplayer, you can compile Mozilla with gcc 2.95 yourself, but for the overwhelming majority of users, keeping up with modern standards makes the most sense (even if proprietary plugins are most important, it is much more likely that Real will eventually rebuild their plugin with a modern compiler, than Java or Flash will be recompiled with an older one).
Download a new version of a web browser, break all your old plugins because of a compiler incompatibility.
Actually, while it may break RealPlayer (which AFAIK hasn't been updated in ages), this is actually absolutely necessary in order to use the latest Flash and JRE plugins, which being targeted to the latest version of Red Hat, are compiled with that gcc 3.2.
This is just moving in the direction that every distribution has been going in for some time. Basically all Linux Mozilla binaries in regular use, aside from those provided by mozilla.org, have been compiled with this for quite a while, because it is the standard compiler on every distribution. It is very sensible for mozilla to make this switch, since every distribution is using gcc 3.2 as its compiler anyway, and it is what proprietary plugin makers should be targeting.
As I understand it, you probably can't. They are not selling you a license to Linux; they have acknowledged that they don't have the right to do so. What they are selling is a license to UnixWare. A useless license, since it does not include the software. But it includes is a promise that they won't sue you. That is what you get for your $699. If it turns out they have no valid infringement claims against Linux, well, you weren't paying for Linux in the first place so there's little you can do.
Their EBITDA was $168K for all last year? On sales of 65+ million? That's close enough to zero to arouse my suspicions about their accounting.
Not really, when you think about it. They probably had a good idea how much money they were losing off of Unix sales, so they could arrange to receive just as much money as they needed from Microsoft and Sun to stay profitable through the lawsuit.
I wonder if the new Amiga OS4 that is about to be released could eventually run on one of these?
The G5 is nothing other than Apple's name for the 970, so the processor should be compatible. But doesn't the new Amiga OS need some kind of special encrypted ROM to work? Also, this is being marketed as a server, so there is no telling whether it will actually support such frivolancies as graphics adapters or keyboards.
OS X/Darwin is entirely optimized for ppc and is developed entirely towards that end. ppc Linux is a port from something else (albeit a good one)
NextStep was originally written for the 68k, thence ported to SPARC, x86, and PA-RISC. So PPC was the fifth architecture the basic underlying system has been ported to. So if you don't like ports, you had better throw away your Mac and switch to Windows now (oh wait sorry...the NT kernel was actually developed for the i860 first).
And remember, Linux will be a native, fully supported OS for these machines alongside AIX -- the firmware will be designed to boot Linux, and all the hardware will be fully supported.
Yes, you can. The GPL gives you certain rights beyond those granted under copyright law, provided you comply with its terms. If you don't like them, you have to negotiate different license terms with the copyright holder, just like with any other copyrighted work.
The BitKeeper license attempts to take away the well-established right, under most countries' laws, to reverse-engineer software. As many others have pointed out, in most countries these terms simply can't be legally binding, and "breaking the license" by exercising your legal rights is probably not going to be against the law.
I almost wish someone would try to sue linksys (now cisco) for the wrt54g surce code issue so that the GPL could be shot down in court.
Yeah that would be great wouldn't it? Then you would lose the right to use every single piece of GPL'd software, unless you've pre-emptively signed copyright licensing agreements with the authors ahead of time. No, there's no magic law that says if an open-source license is shot down, the code goes into public domain. If the GPL goes away, the copyright is still there, you simply have no longer been granted any rights to use the work by the copyright holder. The GPL may not be legally enforceable, in fact, but the reason it has never gone to trial is that in every instance a company has violated the GPL, they have realized that it is more beneficial to abide by the license and continue to be allowed to use the code, as opposed to having the license voided in court and losing those rights.
Afghanistan is not a "former Soviet Country", really, although it was under a Soviet-backed government for a time. It never had the state-supported education and health-care infrastructure that developed in the Soviet Union, Eastern Europe, and even North Korea. The American-backed guerilla campaign against the government began almost as soon as the communists took power (and six months before the Soviet invasion), and any efforts they might have made to reorganize the society were made difficult by the almost continual state of war after 1979. So it is really the last 25 years that they have been "falling behind", and things were not amazingly great before then.
I just ran grep -ri " sco " on the kernel source tree (well, 2.4 anyway). There are no SCO copyright notices. All the references are either third-party drivers ported from SCO versions, the SCO binary compatibility support (which no one has implied contains any copied code), and some references to something called "SCO" in the bluetooth driver which seems to be totally unrelated to our friends from Utah. I did the same thing with AT&T and also found nothing.
If what SCO has been saying is true (granted, they have been changing their story so much you can't really believe anything they say), the copyrights are probably from IBM or Sequent, or possibly SGI, and they copied the code straight across because they thought they owned it. Hopefully a sane judge will not disagree with that.
As for BSD, yeah, there is a lot of stuff like this:
./include/linux/if.h: * Authors: Original taken from Berkeley UNIX 4.3, ( c) UCB 1982-1988
Rights are the biggies--life, liberty, pursuit of property---not this laundry list crap. Calling it a 'right' is just a cynical ploy to make an entitlement impossible to remove or de-fund at some future date.
But isn't "making it impossible to remove or de-fund at some future date" exactly the purpose of setting out laws granting your "biggies"? Those things are hardly absolutes, and one doesn't have to look far to find places where those "rights" don't exist. Killing people, taking away their freedom, and stealing their property are all things which are very much physically possible and in the absence of laws people have not shown any qualms about doing them. We say that they are rights because we believe it is more desirable to live in a society where these rights exist than one in which they do not, despite the obvious high cost of hiring police and courts to deal with murder and theft, not to mention the possible threat to government power posed by freedom of speech, and the difficulties dealing with pesky dissidents without being able to summarily kill or imprison them.
Likewise, the Estonians have decided that a society in which all people can access the Internet is more desirable than one in which some cannot, and that writing this entitlement into the consitution was the best way to move towards this goal, and ensure that this access is permanent. There are no objective grounds for declaring this right to be any less of a right, or any less "natural", than any other entitlement established by a state.
As of 2.5.72 I think, the Matrox framebuffer at least compiles again. There is too much screen corruption on my 8mb G200 for it to be really usable yet...but its an improvement over the previous total breakage.
If mathematicans aren't really interested in helping understand the world, why should society fund them?
These are two separate things. Many people are attracted to the natural sciences, and even engineering disciplines, not because of a desire to improve the world, but because they find pleasure and abstract beauty in those fields. Yet undeniably work in those areas can lead to benefits for "society", and therefore people doing research in those areas are funded, even if their personal reasons for doing the work have nothing to do with those benefits. Likewise with mathematics, many ideas thought of as purely abstract and disconnected from practical application have turned out, later on, to be useful tools in understanding various real-world phenomena.
It is totally unscientific and ultimately counter-productive to close off areas of inquiry because at the time they are undertaken no one can know exactly what the consequences will be. And ultimately the motivations of the people involved are irrelevant; we know based on history that there could turn out to be uses for it in the future, even if neither "we" (the society making the decision to support the research), nor those doing the research, can see any at this time, and this potentiality alone should justify providing support.
The mozilla.org tarballs don't include the development headers. The RPM packages do. If you want to compile Epiphany, Galeon, Skipstone, or one of the other fine gecko-based browsers, you have to have these installed. The alternative, if you don't want to use the packages, is to either compile mozilla from scratch also (fine if you have 12 hours or so free), or pick the relevant parts out of the mozilla source tree and manually install them somewhere yourself. Of course, you can install any of the above browsers from binary packages...but these will have a dependency on the RPMs for Mozilla, so you will have to have them installed then too unless you want to break your distribution's package system.
Not to mention the fact that the Mozilla binary releases usually have a lowest-common-denominator feature set enabled, and are unlikely to have support for Xft, or GTK 2.x, etc.
Well, I assume XINE can play MP3's (it has to support the format anyway for all of those video encodings that use MP3 for audio), and as you can see from the plugins page, XMMS can be made to play a wide variety of non-audio media formats. So that description is technically correct.
Here is the problem. You are basing your conclusions on "logic", or as it shall be known now, Old Logic. Old Logic is no longer in effect: we are now operating under New Logic, also known as SCO Logic. Here is how it works, taking the simple example first.
Sequent has licensed Unix from AT&T. Sequent develops some nifty stuff like NUMA, RCU, etc. for large parallel computers. They then incorporate this into Unix. The result is now a derived work of Unix, and AT&T/successors have "sole relicensing rights" to that derived work. But here's where the New Logic comes in: Not only that version of Unix, but also the original implementations of those technologies themselves, are also derived works, and the Unix license grants them to AT&T, or their successor in holding the license, in this case SCO. You'd think that Sequent's copyrights (now held by IBM) would let them do whatever they want, but this is erroneous: The Unix license grants these rights to SCO, and IBM can distribute implementations of these technologies only according to the terms of the System V license. This is what SCO has claimed all along.
Now lets apply this to Linux. IBM has incorporated implementations of these ideas, which can only be licensed through SCO even though IBM owns them, into Linux, under the GPL. Under Old Logic, even if we accept the above, IBM has simply broken the GPL, and the kernel without these parts still belongs to the original contributors, and can be licensed under the GPL. But as always, New Logic changes everything: Since Linux, with RCU and NUMA support, is now a derived work of System V Unix (even if it has been created illegally), SCO now has "sole right of relicense" to not only this derived work, but as with Sequent, all previous implementations, even those without the licensed code (just as it applies to implementations of RCU and NUMA before they were incorporated into Unix). And under the New Logic, Linus' copyright matters no more than Sequent's.
SCO has no right to "bless" non-GPL redistribution
The most sickening thing is that it sounds like SCO really doesn't want the (hypothetical) code removed. They want to convince a court that somehow IBM's AIX contract gives them ownership of anything that happens to have Unix code in it, and this overrides anyone else's copyrights, even if they wrote it, Linux as well s Dynix and AIX.
If they really cared about their Linux customers, they would flag the (alleged) offending code immediately and, if not replace it themselves, at least alert the kernel developers, so that a legal fix could be released. The records are clear enough that it seems they ought to be able to win damages from whoever was responsible for adding it, even if the offending code had been removed from current kernels. If there is no real offending code, though, what they are doing makes perfect sense: the more extreme and ludicrous their claims are, the higher their stock price goes, and the higher their (still slim) chances of being bought out or getting a nice settlement, rather than being embarassed in court.
There is something to this. However, if it were the case that some code included in the Linux kernel weren't distributable under the GPL, it would be illegal to distribute it at all. Why? Because all of the rest of the code, going back to the original bits written by Linux in 1991, is under the GPL. Combining that code with other code creates a derivative work of that code. Under the GPL's famous viral provisions, any derivative work of GPL'd software has to be distributable under the terms of the GPL (ie, you can add code which is GPL'd, or you can add code under another license that does not conflict with the GPL).
So if there is violating code, SCO has no right to "bless" redistribution unless they agree to non-GPL licensing terms with EVERY SINGLE KERNEL DEVELOPER who contributed non-offending source. If they are implying that distribution of their linux kernel is legal, therefore, it implies that they are offering whatever portions they claim ownership of under the terms of the GPL. If not, they are infringing upon the copyrights of hundreds of kernel developers.
Actually I wasn't trying to imply that anyone is going to switch from Mac to Linux (aside maybe from some Mac users dual-booting), I agree that main growth will be from Windows users. If anyone is using a Mac right now, it is because they believe fervently that it is a superior platform. And unless new Macintosh development ceases I think those people will continue to believe that, and Macintosh usage will stay at its current level.
What you have to understand is that after Jobs came in, Apple permanently gave up the idea of unseating Wintel dominance, or even gaining market share. Everything they do now is focused on extracting maximum profits on sales to their current fanatical user base, and keeping them locked into the platform for as long as possible. Even through the last few years of "success" their sales have been relatively stagnant and the overall market share has dropped. They are making money, more or less, though, so no one is too concerned. But there is nothing of the sense of manifest destiny and that empowers the Linux/open source community, nor the ability to run on practically any computer hardware, so in the long run I don't see how they could hope to stay on top of Linux (as long as the usability of desktop Linux distributions improves and continues to approach parity with Windows and Mac OS).
Slashdot Mirror
How much would a hostile takeover of SCO cost?
It would be impossible. As has been repeatedly detailed here, they are 95% privately held. The shares you would need to buy to take them over simply aren't available on the market.
yeah, but why should this happen? i know why it does happen, but why should it?
Well, its pretty simple. gcc had some standards-compliance and performance issues. In order to address them, they made some changes which broke binary compatibility with older versions. Every linux distribution switches to the newer version, because, well, its better. Most third-party binary packages, like Mozilla, also switch to match the evolving standard Linux platform. This is overwhelmingly positive, except for some barely-maintained proprietary software whose developers can't manage to recompile it every few years to keep up with the times.
It would be a major impediment to the development of the platform if it was required to keep the kind of binary compatibility that would be necessary to keep ancient proprietary plugins working. What if the last release of RealPlayer was compiled with gcc 2.7, or required libc5, or used a.out libraries? Would you expect Mozilla to base their releases on those ancient technologies? I hope not. If you really want to use realplayer, you can compile Mozilla with gcc 2.95 yourself, but for the overwhelming majority of users, keeping up with modern standards makes the most sense (even if proprietary plugins are most important, it is much more likely that Real will eventually rebuild their plugin with a modern compiler, than Java or Flash will be recompiled with an older one).
Download a new version of a web browser, break all your old plugins because of a compiler incompatibility.
Actually, while it may break RealPlayer (which AFAIK hasn't been updated in ages), this is actually absolutely necessary in order to use the latest Flash and JRE plugins, which being targeted to the latest version of Red Hat, are compiled with that gcc 3.2.
This is just moving in the direction that every distribution has been going in for some time. Basically all Linux Mozilla binaries in regular use, aside from those provided by mozilla.org, have been compiled with this for quite a while, because it is the standard compiler on every distribution. It is very sensible for mozilla to make this switch, since every distribution is using gcc 3.2 as its compiler anyway, and it is what proprietary plugin makers should be targeting.
As I understand it, you probably can't. They are not selling you a license to Linux; they have acknowledged that they don't have the right to do so. What they are selling is a license to UnixWare. A useless license, since it does not include the software. But it includes is a promise that they won't sue you. That is what you get for your $699. If it turns out they have no valid infringement claims against Linux, well, you weren't paying for Linux in the first place so there's little you can do.
Whats more, Microsoft's Windows UNIX layer is probably as close to modern Unix standards as anything SCO sells.
Their EBITDA was $168K for all last year? On sales of 65+ million? That's close enough to zero to arouse my suspicions about their accounting.
Not really, when you think about it. They probably had a good idea how much money they were losing off of Unix sales, so they could arrange to receive just as much money as they needed from Microsoft and Sun to stay profitable through the lawsuit.
I wonder if the new Amiga OS4 that is about to be released could eventually run on one of these?
The G5 is nothing other than Apple's name for the 970, so the processor should be compatible. But doesn't the new Amiga OS need some kind of special encrypted ROM to work? Also, this is being marketed as a server, so there is no telling whether it will actually support such frivolancies as graphics adapters or keyboards.
OS X/Darwin is entirely optimized for ppc and is developed entirely towards that end. ppc Linux is a port from something else (albeit a good one)
NextStep was originally written for the 68k, thence ported to SPARC, x86, and PA-RISC. So PPC was the fifth architecture the basic underlying system has been ported to. So if you don't like ports, you had better throw away your Mac and switch to Windows now (oh wait sorry...the NT kernel was actually developed for the i860 first).
And remember, Linux will be a native, fully supported OS for these machines alongside AIX -- the firmware will be designed to boot Linux, and all the hardware will be fully supported.
You can't have it both ways, chief.
Yes, you can. The GPL gives you certain rights beyond those granted under copyright law, provided you comply with its terms. If you don't like them, you have to negotiate different license terms with the copyright holder, just like with any other copyrighted work.
The BitKeeper license attempts to take away the well-established right, under most countries' laws, to reverse-engineer software. As many others have pointed out, in most countries these terms simply can't be legally binding, and "breaking the license" by exercising your legal rights is probably not going to be against the law.
I almost wish someone would try to sue linksys (now cisco) for the wrt54g surce code issue so that the GPL could be shot down in court.
Yeah that would be great wouldn't it? Then you would lose the right to use every single piece of GPL'd software, unless you've pre-emptively signed copyright licensing agreements with the authors ahead of time. No, there's no magic law that says if an open-source license is shot down, the code goes into public domain. If the GPL goes away, the copyright is still there, you simply have no longer been granted any rights to use the work by the copyright holder. The GPL may not be legally enforceable, in fact, but the reason it has never gone to trial is that in every instance a company has violated the GPL, they have realized that it is more beneficial to abide by the license and continue to be allowed to use the code, as opposed to having the license voided in court and losing those rights.
Afghanistan is not a "former Soviet Country", really, although it was under a Soviet-backed government for a time. It never had the state-supported education and health-care infrastructure that developed in the Soviet Union, Eastern Europe, and even North Korea. The American-backed guerilla campaign against the government began almost as soon as the communists took power (and six months before the Soviet invasion), and any efforts they might have made to reorganize the society were made difficult by the almost continual state of war after 1979. So it is really the last 25 years that they have been "falling behind", and things were not amazingly great before then.
Back to playing sed pong for me...
Watch your case. "SCO" != "sco"
From the grep manpage:
-i, --ignore-case
Ignore case distinctions in both the PATTERN and
the input files.
If what SCO has been saying is true (granted, they have been changing their story so much you can't really believe anything they say), the copyrights are probably from IBM or Sequent, or possibly SGI, and they copied the code straight across because they thought they owned it. Hopefully a sane judge will not disagree with that.
As for BSD, yeah, there is a lot of stuff like this:If thats it, they are on even shakier ground.
Rights are the biggies--life, liberty, pursuit of property---not this laundry list crap. Calling it a 'right' is just a cynical ploy to make an entitlement impossible to remove or de-fund at some future date.
But isn't "making it impossible to remove or de-fund at some future date" exactly the purpose of setting out laws granting your "biggies"? Those things are hardly absolutes, and one doesn't have to look far to find places where those "rights" don't exist. Killing people, taking away their freedom, and stealing their property are all things which are very much physically possible and in the absence of laws people have not shown any qualms about doing them. We say that they are rights because we believe it is more desirable to live in a society where these rights exist than one in which they do not, despite the obvious high cost of hiring police and courts to deal with murder and theft, not to mention the possible threat to government power posed by freedom of speech, and the difficulties dealing with pesky dissidents without being able to summarily kill or imprison them.
Likewise, the Estonians have decided that a society in which all people can access the Internet is more desirable than one in which some cannot, and that writing this entitlement into the consitution was the best way to move towards this goal, and ensure that this access is permanent. There are no objective grounds for declaring this right to be any less of a right, or any less "natural", than any other entitlement established by a state.
As of 2.5.72 I think, the Matrox framebuffer at least compiles again. There is too much screen corruption on my 8mb G200 for it to be really usable yet...but its an improvement over the previous total breakage.
It brings you antialiased fonts (at least in RH9). It is built for Gnome 2.x.
You can compile against GTK 1.2.x and still get antialiased fonts if you also enable Xft support.
Building against 2.x lets you use Galeon 1.3 and Epiphany though, which is why I do it (and am eagerly awaiting the source release).
If mathematicans aren't really interested in helping understand the world, why should society fund them?
These are two separate things. Many people are attracted to the natural sciences, and even engineering disciplines, not because of a desire to improve the world, but because they find pleasure and abstract beauty in those fields. Yet undeniably work in those areas can lead to benefits for "society", and therefore people doing research in those areas are funded, even if their personal reasons for doing the work have nothing to do with those benefits. Likewise with mathematics, many ideas thought of as purely abstract and disconnected from practical application have turned out, later on, to be useful tools in understanding various real-world phenomena.
It is totally unscientific and ultimately counter-productive to close off areas of inquiry because at the time they are undertaken no one can know exactly what the consequences will be. And ultimately the motivations of the people involved are irrelevant; we know based on history that there could turn out to be uses for it in the future, even if neither "we" (the society making the decision to support the research), nor those doing the research, can see any at this time, and this potentiality alone should justify providing support.
The mozilla.org tarballs don't include the development headers. The RPM packages do. If you want to compile Epiphany, Galeon, Skipstone, or one of the other fine gecko-based browsers, you have to have these installed. The alternative, if you don't want to use the packages, is to either compile mozilla from scratch also (fine if you have 12 hours or so free), or pick the relevant parts out of the mozilla source tree and manually install them somewhere yourself. Of course, you can install any of the above browsers from binary packages...but these will have a dependency on the RPMs for Mozilla, so you will have to have them installed then too unless you want to break your distribution's package system.
Not to mention the fact that the Mozilla binary releases usually have a lowest-common-denominator feature set enabled, and are unlikely to have support for Xft, or GTK 2.x, etc.
Well, I assume XINE can play MP3's (it has to support the format anyway for all of those video encodings that use MP3 for audio), and as you can see from the plugins page, XMMS can be made to play a wide variety of non-audio media formats. So that description is technically correct.
But they probably did mean the opposite, yes.
Here is the problem. You are basing your conclusions on "logic", or as it shall be known now, Old Logic. Old Logic is no longer in effect: we are now operating under New Logic, also known as SCO Logic. Here is how it works, taking the simple example first.
Sequent has licensed Unix from AT&T. Sequent develops some nifty stuff like NUMA, RCU, etc. for large parallel computers. They then incorporate this into Unix. The result is now a derived work of Unix, and AT&T/successors have "sole relicensing rights" to that derived work. But here's where the New Logic comes in: Not only that version of Unix, but also the original implementations of those technologies themselves, are also derived works, and the Unix license grants them to AT&T, or their successor in holding the license, in this case SCO. You'd think that Sequent's copyrights (now held by IBM) would let them do whatever they want, but this is erroneous: The Unix license grants these rights to SCO, and IBM can distribute implementations of these technologies only according to the terms of the System V license. This is what SCO has claimed all along.
Now lets apply this to Linux. IBM has incorporated implementations of these ideas, which can only be licensed through SCO even though IBM owns them, into Linux, under the GPL. Under Old Logic, even if we accept the above, IBM has simply broken the GPL, and the kernel without these parts still belongs to the original contributors, and can be licensed under the GPL. But as always, New Logic changes everything: Since Linux, with RCU and NUMA support, is now a derived work of System V Unix (even if it has been created illegally), SCO now has "sole right of relicense" to not only this derived work, but as with Sequent, all previous implementations, even those without the licensed code (just as it applies to implementations of RCU and NUMA before they were incorporated into Unix). And under the New Logic, Linus' copyright matters no more than Sequent's.
errg.
SCO has no right to "bless" redistribution
Should be:
SCO has no right to "bless" non-GPL redistribution
The most sickening thing is that it sounds like SCO really doesn't want the (hypothetical) code removed. They want to convince a court that somehow IBM's AIX contract gives them ownership of anything that happens to have Unix code in it, and this overrides anyone else's copyrights, even if they wrote it, Linux as well s Dynix and AIX.
If they really cared about their Linux customers, they would flag the (alleged) offending code immediately and, if not replace it themselves, at least alert the kernel developers, so that a legal fix could be released. The records are clear enough that it seems they ought to be able to win damages from whoever was responsible for adding it, even if the offending code had been removed from current kernels. If there is no real offending code, though, what they are doing makes perfect sense: the more extreme and ludicrous their claims are, the higher their stock price goes, and the higher their (still slim) chances of being bought out or getting a nice settlement, rather than being embarassed in court.
There is something to this. However, if it were the case that some code included in the Linux kernel weren't distributable under the GPL, it would be illegal to distribute it at all. Why? Because all of the rest of the code, going back to the original bits written by Linux in 1991, is under the GPL. Combining that code with other code creates a derivative work of that code. Under the GPL's famous viral provisions, any derivative work of GPL'd software has to be distributable under the terms of the GPL (ie, you can add code which is GPL'd, or you can add code under another license that does not conflict with the GPL).
So if there is violating code, SCO has no right to "bless" redistribution unless they agree to non-GPL licensing terms with EVERY SINGLE KERNEL DEVELOPER who contributed non-offending source. If they are implying that distribution of their linux kernel is legal, therefore, it implies that they are offering whatever portions they claim ownership of under the terms of the GPL. If not, they are infringing upon the copyrights of hundreds of kernel developers.
Actually I wasn't trying to imply that anyone is going to switch from Mac to Linux (aside maybe from some Mac users dual-booting), I agree that main growth will be from Windows users. If anyone is using a Mac right now, it is because they believe fervently that it is a superior platform. And unless new Macintosh development ceases I think those people will continue to believe that, and Macintosh usage will stay at its current level.
What you have to understand is that after Jobs came in, Apple permanently gave up the idea of unseating Wintel dominance, or even gaining market share. Everything they do now is focused on extracting maximum profits on sales to their current fanatical user base, and keeping them locked into the platform for as long as possible. Even through the last few years of "success" their sales have been relatively stagnant and the overall market share has dropped. They are making money, more or less, though, so no one is too concerned. But there is nothing of the sense of manifest destiny and that empowers the Linux/open source community, nor the ability to run on practically any computer hardware, so in the long run I don't see how they could hope to stay on top of Linux (as long as the usability of desktop Linux distributions improves and continues to approach parity with Windows and Mac OS).