Apps/libs running on the OS isn't linked in any way to the kernel. The ELF format is mere data to be processed, and data processed by GPL'ed programs isn't subject to the GPL.
This is where I think the logic fails. The program flow surely is transferred between the program and kernel. For example, when I open a file, I make a call which causes kernel file system (and possibly lot of other) code to be run. What you're talking about sounds like you were thinking of some byte code, or otherwise interpreted language. At least in the case of machine language, I still don't understand why this doesn't qualify as linking and libraries do.
Dynamic linking constitutes derived work because you are enhancing your program by using library functions(libc is an exception). There is nothing different about static, shared or dynamic in this sense.
Yes, the linked program is derivative, I agree on that. But this linking is done at run-time, so this makes me wonder whether only the resulting memory footprint is derivative. And licenses can only control distribution, not how you use the work. It's like potato chips and dipping sauce. Dipping sauce would be pretty much useless without chips, yet it clearly is not a derivative work. I can also think about examples in the IP field, like a data sheet and a user's manual or instruction reference for a microprocessor. Neither is very useful alone, yet they are two separate works (neither is necessarily a derivative of the other) and their copyrights can at least in theory be owned by two different legal entitites.
Correct me if I'm wrong, but I'm under the impression that linking GPL code with Motif libraries is not infringing, nor is running the program, only distributing the resulting code/binary is. That is, even if the original toolkit in your example was not LessTif but, say, GTK and I somehow managed to replace GTK with Motif, I believe I wouldn't be infringing unless I actually distributed the program.
But how about this scenario:
I write an operating system, possibly under GPL. Be it that it only includes a boot loader, memory allocator and disk access + file system routines, so I'm not talking about any end user OS. I take other people's GPL code and run it as an application (not in the kernel space). Now, I proceed to write e.g. a massive and proprietary game as just another application.
Now, I believe and hope that this would be infringing. I can almost hear the authors of that GPL'd code yell at me. Yet I think it's quite similar to what's being done with Linux.
There might not be any other way to get the system to change. As long as the people who are exploiting the system are making lots of money, they'll fight any attempt to change the system to a form where they can't - and because they're making lots of money, they have the resources to make such a fight really painful for their opponents. If you want to win, you need to cut off their revenue stream, and unauthorized copying is the most straightforward way of showing how ephemeral their "right to profit" really is.
I think this is exactly where entirely legal means, especially philosophies like open source / free software step in. Really, if something is now threatening big companies like Microsoft, which one is it - piracy or open source? I think the answer should be quite obvious. And one of the greatest things is that this way the exploiters will have really tough time trying to figure out who their opponent is.
I'd go actually so far to say that whoever pirates e.g. MS products is actually supporting the exploiters, because this way they make Windows a profitable platform for MS and other companies.
I doubt the intent is to "hurt" the lawmakers (although I have a sudden mental vision of the RIAA filing suit against a Congresscritter for using Gnutella...nah, it'd never happen:-), just make it painfully obvious to the industry that society does not view their current business model as a legitimate good or service. If they're smart, they'll find an alternative.
Or as an alternative, these big companies go to the lawmakers crying out loud "piracy" and get fancy new laws which give them more and more power over how their products are used - which as has been seen give more and more power to fight those that really try to change the model (for example DeCSS - remember, it was created to make it possible to have an open source DVD player).
Oh well, one can still hope that one day the lawmakers themselves have to realize how draconian these laws have become and have to change them. This day is however further pushed forward by those who give big companies the very chance to justify the need for these laws by endorsing piracy.
By the way, aren't these exceptions too a bit tricky thing?
I mean, if (and only if) it would be held that without these exceptions distributing proprietary programs for Linux would be infringing, there's a little problem. It's no longer that the distributors of proprietary software are infringing the Linux kernel copyright, but that making such exceptions requires the acceptance from every contributor.
That is to say, if I wrote a GPL'd program originally to run on, say, FreeBSD, and someone ported it to Linux, this would be infringing my copyright. Let me explain why.
This way, the Linux kernel was sort of linked to my program (or vice versa, it really doesn't matter). Now, also the proprietary program, e.g. Netscape, is in similar stance. The net effect is that my program has been linked with proprietary Netscape code, albeit through an abstraction layer, and entirely without my consent, thus infringing my copyright.
Again, before flaming, please read again the word 'if' in the beginning of this post. I'm not trying to attack GPL. It's a truly great license, and the one which I have chosen for some of my programs out there.
I know. And what you say of what I may or may not do, I very well know that's what FSF holds to be legally true. I was only wondering how it can be infringing to distribute a non-GPL'd program which at runtime only is linked with GPL'd code. Note that I'm not attacking GPL - I certainly hope that it holds in court one day.
Your attempt to differenciate "distrubution time" and "runtime" shoes me that you don't have alot of experience as a programmer. There is not difference.
I might have more experience as a programmer than you think, and I still maintain there's a HUGE difference between runtime and distribution time.
To me, running my program or uploading it via FTP are two very distinct things.
You make a program that is proprietary. You make in your program a mecanism that allows plugins through dynamic linking. Someone writes a GPL plugin. Who's the infringer? The user, you or the plugin writer.
Well, this is basically what I tried to ask, but you put it better than me. And again that the plugin writer is the infringer is the legal opinion of FSF. But my point was, how is he infringing? Now he doesn't do the linking, which only happens after distribution, where the license no longer can have effect.
I think about this in a way analogously to having references in a book you write. I think this is analogous because
1) In both cases, the copyright is owned by whoever wrote the thing
2) Making 'full use' of the thing requires the user to already have the linked-to material (shared libraries or other books)
3) No material under copyright from other authors' books or software ever is distributed.
Now, please tell me I'm wrong (and if you do, point out why). After all, I want to have the freedom to write GPL programs myself without the fear that they will end up in a proprietary product which "extends" the functionality of my original program, only now as a GPL'd library.
While I might agree with the long-term goal of changing the copyright system, I don't think piracy is the way to fight the system. Yes, I know about civil disobedience, and I agree that it's necessary sometimes. However, pirates are not harming the lawmakers, but those authors who don't have other system to build upon. That's the difference between civil disobedience as a means of politics and immoral criminality.
No silly analogies are needed: Writing a peice of software that decrypts books simply should not be a crime!
Being a programmer myself, I wholeheartedly agree. I also agree that this program should be legal, and DeCSS should be legal. Note however, I was not talking about writing a program to decrypt, I was talking about illegal copying (and so was the parent of my post). And I don't agree that copying other people's works, distributed for profit, should be legal (except for fair and noninfringing uses).
You want to apply corporeal rules to the ethereal? Fine, then do it ALL THE way. If I can break into my own car, I should be able to break into my own copy of an ebook.
Had you read my post, you would perhaps have noticed this is what I said. Oh well, sometimes that's too much asked.
Another thing in GPL which makes me wonder is that whenever a shared library (or DLL) is distributed under the GPL, and a proprietary program (allegedly illegally - I certainly hope so) is linked to it in runtime, who is actually the infringer? I.e. is it legal to distribute a proprietary program which isn't linked to code under GPL at distribution time, but only in runtime, and isn't the infringer then the user?
FSF clearly thinks this is not legal, and I hope they're right about that. Otherwise this would foil the entire GPL.
Well, the FSF has a history of forgiving companies
and groups for initial mistakes with the GPL so long as they choose to become compliant. The latest one
was KDE/Qt debacle.
Umm, I don't think the FSF had any say in this. As far as I know, they didn't own any components included in KDE nor any in Qt. Qt copyright is owned by Trolltech and KDE copyright by those who wrote KDE. And the mere choice of GPL as the license does not give FSF the right to pursue violations unless it's _their_ software that is violated.
And you keep "missing" the point that copying something is NOT stealing - it's COPYING. It's illegal, yes, but it's illegal COPYING.
If this illegal copying causes less people to buy the book, it suddenly will no longer be feasible to write books, and people will be fully deprived of books. It's taking something that is not yours while depriving others of something they would otherwise have. In other words, it's stealing.
With a similar logic, should people be allowed to chop a few trees from every forest because? Surely there will be new trees growing which couldn't have in a cramped forest. So it's not theft.
While I'm opposed to illegal copying and wood theft, I still maintain that fair and non-infringing uses (which I believe includes access, not only the mere right to view as in eBooks) are far too important to squash with a stupid piece of legislation.
No, you are missing the point. No matter what, these "breaking into my house" stories will never be relavent to an IP discussion because stealing my stero and stealing a copy of an electronic book are two entirely different acts.
I believe you're missing the point, too. Let's see.
1. Reading a book you have legally bought is not stealing.
2. Taking a snippet from a book you have legally bought is not stealing, if done for fair use purposes.
3. Transferring a book you have legally bought to another computer so you don't need to keep that old 286 lying around is not stealing.
4.Removing a measure that "protects" the book from being fed to your Text-to-Speech system is not stealing (whether you are blind or not).
Still, stealing is stealing. Copying this book to your friend would be wrong. Accessing it, no way.
Quit splitting hairs. You know what I mean. You can, of course, break into your own house. Though using dynamite might get you some questions from the ATF...
Don't you find breaking into your own house and accessing a book you own at least a bit analogous?
Sorry, but USA have a long history regarding this type of logic. For example, in the US of A (and not, for example, in the EU), not only it's illegal to sell drugs,
Huh, are you serious? I believe it's still illegal in most EU countries to possess illegal drugs for any purpose. Yes, I live in the EU, and at least here in Finland it's illegal. And, really, I'm happy this way, and I terribly fail to understand your logic.
A bit by bit comparison doesn't really make sense. Neither does frequency response analysis. These encoders were designed specifically to alter these things.
The only thing which matters is, "how good does it sound to the listeners". And that's indeed the most scientific way of testing it I can think of. Blind tests, comparing with real CD quality.
I can't though, so it doesn't matter. I'm not a musician by any means, nor can I detect the difference between 160 and 192 mp3 compression. So I'll continue using my inferior, yet cross platform, non-license restricted, used-everwhere, mp3 format.
Funny. Licensing issues are the single most important reason for not using MP3, at least for me. Looking for a non-license restricted format? Try Vorbis.
Umm, you sound more than a little confused about this technology.
The entire issue of how to encode sound is left, perhaps unsurprisingly, to the encoder. If you find a decoder (player) which gives crappy output on a good mp3/ogg file, well then that player is crappy. Yes, I know that such players exist.
Reverse engineering also has nothing to do with this, and reverse engineering an algorithm doesn't really make it "non-native" or crappy in any other way.
Now, I'm not much of a MS fan, however, it must be pointed out that their office software does work very well, and works well with itself. So, while you can argue that the OS is brain-dead, the software, (minus the paperclip) is of decent quality.
I'm not a MS fan either, I use Linux exclusively (at home, anyway).
But one thing I always wonder is people saying that the Windows OS/kernel is crap and office software is their strength. I would say it's vice versa - I can use Windows (although I prefer Linux), but Word I wouldn't touch with a 5 foot long pole. Installing LaTeX (or some other tool, suitable for the task at hand) is usually the first thing to do when I have to use Windows.
And the Windows kernel's not really that lousy, although a little buggy (but certainly better than the office tools, I'd say).
If they had just approached the author of KIllustrator and asked, "Would you please change your name?" I'll bet he would have quietly changed it and this would have never been an issue.
Aside: I learned ASM on a vax, and on man did it kick ass over PC ASM... memory to memory copies (or was it register to register?), no 640k limit to mess with, no weird offsets because Bill thought it was a good idea....
Well, most of that really isn't an issue anymore: The 1M (instead of 640k) limit can be turned off for protected mode and you can use linear addressing, no need for segmentation (which I guess you meant by saying "weird offsets"). Mem to mem copies are still missing, apart from the MOVS* instructions, which I think have been there from the 8088s.
Agreed, x86 is a real mess for a processor. All kinds of backwards compatibility hacks - the one I most like is the disabling of the A20 (1 megabyte) address line by writing a command to the keyboard controller.
There's one thing I miss a lot on x86 processors, though. I think it would be great to have at least a dozen more general purpose registers (though even 1 or 2 would be an improvement).
Well, not in any country. Certainly not in Finland. And not in a lot of other countries. I don't know about US or Australia, though.
Even further, if you refused to give the key, how on earth would they know if you've forgotten it or simply don't want to give it? You certainly can't be jailed or fined for forgetting your own passphrase, can you?
Slashdot Mirror
This is where I think the logic fails. The program flow surely is transferred between the program and kernel. For example, when I open a file, I make a call which causes kernel file system (and possibly lot of other) code to be run. What you're talking about sounds like you were thinking of some byte code, or otherwise interpreted language. At least in the case of machine language, I still don't understand why this doesn't qualify as linking and libraries do.
Dynamic linking constitutes derived work because you are enhancing your program by using library functions(libc is an exception). There is nothing different about static, shared or dynamic in this sense.
Yes, the linked program is derivative, I agree on that. But this linking is done at run-time, so this makes me wonder whether only the resulting memory footprint is derivative. And licenses can only control distribution, not how you use the work. It's like potato chips and dipping sauce. Dipping sauce would be pretty much useless without chips, yet it clearly is not a derivative work. I can also think about examples in the IP field, like a data sheet and a user's manual or instruction reference for a microprocessor. Neither is very useful alone, yet they are two separate works (neither is necessarily a derivative of the other) and their copyrights can at least in theory be owned by two different legal entitites.
Correct me if I'm wrong, but I'm under the impression that linking GPL code with Motif libraries is not infringing, nor is running the program, only distributing the resulting code/binary is. That is, even if the original toolkit in your example was not LessTif but, say, GTK and I somehow managed to replace GTK with Motif, I believe I wouldn't be infringing unless I actually distributed the program.
But how about this scenario:
I write an operating system, possibly under GPL. Be it that it only includes a boot loader, memory allocator and disk access + file system routines, so I'm not talking about any end user OS. I take other people's GPL code and run it as an application (not in the kernel space). Now, I proceed to write e.g. a massive and proprietary game as just another application.
Now, I believe and hope that this would be infringing. I can almost hear the authors of that GPL'd code yell at me. Yet I think it's quite similar to what's being done with Linux.
I think this is exactly where entirely legal means, especially philosophies like open source / free software step in. Really, if something is now threatening big companies like Microsoft, which one is it - piracy or open source? I think the answer should be quite obvious. And one of the greatest things is that this way the exploiters will have really tough time trying to figure out who their opponent is.
I'd go actually so far to say that whoever pirates e.g. MS products is actually supporting the exploiters, because this way they make Windows a profitable platform for MS and other companies.
I doubt the intent is to "hurt" the lawmakers (although I have a sudden mental vision of the RIAA filing suit against a Congresscritter for using Gnutella...nah, it'd never happen :-), just make it painfully obvious to the industry that society does not view their current business model as a legitimate good or service. If they're smart, they'll find an alternative.
Or as an alternative, these big companies go to the lawmakers crying out loud "piracy" and get fancy new laws which give them more and more power over how their products are used - which as has been seen give more and more power to fight those that really try to change the model (for example DeCSS - remember, it was created to make it possible to have an open source DVD player).
Oh well, one can still hope that one day the lawmakers themselves have to realize how draconian these laws have become and have to change them. This day is however further pushed forward by those who give big companies the very chance to justify the need for these laws by endorsing piracy.
I mean, if (and only if) it would be held that without these exceptions distributing proprietary programs for Linux would be infringing, there's a little problem. It's no longer that the distributors of proprietary software are infringing the Linux kernel copyright, but that making such exceptions requires the acceptance from every contributor.
That is to say, if I wrote a GPL'd program originally to run on, say, FreeBSD, and someone ported it to Linux, this would be infringing my copyright. Let me explain why.
This way, the Linux kernel was sort of linked to my program (or vice versa, it really doesn't matter). Now, also the proprietary program, e.g. Netscape, is in similar stance. The net effect is that my program has been linked with proprietary Netscape code, albeit through an abstraction layer, and entirely without my consent, thus infringing my copyright.
Again, before flaming, please read again the word 'if' in the beginning of this post. I'm not trying to attack GPL. It's a truly great license, and the one which I have chosen for some of my programs out there.
<snip>
I know. And what you say of what I may or may not do, I very well know that's what FSF holds to be legally true. I was only wondering how it can be infringing to distribute a non-GPL'd program which at runtime only is linked with GPL'd code. Note that I'm not attacking GPL - I certainly hope that it holds in court one day.
Your attempt to differenciate "distrubution time" and "runtime" shoes me that you don't have alot of experience as a programmer. There is not difference.
I might have more experience as a programmer than you think, and I still maintain there's a HUGE difference between runtime and distribution time.
To me, running my program or uploading it via FTP are two very distinct things.
You make a program that is proprietary. You make in your program a mecanism that allows plugins through dynamic linking. Someone writes a GPL plugin. Who's the infringer? The user, you or the plugin writer.
Well, this is basically what I tried to ask, but you put it better than me. And again that the plugin writer is the infringer is the legal opinion of FSF. But my point was, how is he infringing? Now he doesn't do the linking, which only happens after distribution, where the license no longer can have effect.
I think about this in a way analogously to having references in a book you write. I think this is analogous because
1) In both cases, the copyright is owned by whoever wrote the thing
2) Making 'full use' of the thing requires the user to already have the linked-to material (shared libraries or other books)
3) No material under copyright from other authors' books or software ever is distributed.
Now, please tell me I'm wrong (and if you do, point out why). After all, I want to have the freedom to write GPL programs myself without the fear that they will end up in a proprietary product which "extends" the functionality of my original program, only now as a GPL'd library.
While I might agree with the long-term goal of changing the copyright system, I don't think piracy is the way to fight the system. Yes, I know about civil disobedience, and I agree that it's necessary sometimes. However, pirates are not harming the lawmakers, but those authors who don't have other system to build upon. That's the difference between civil disobedience as a means of politics and immoral criminality.
Being a programmer myself, I wholeheartedly agree. I also agree that this program should be legal, and DeCSS should be legal. Note however, I was not talking about writing a program to decrypt, I was talking about illegal copying (and so was the parent of my post). And I don't agree that copying other people's works, distributed for profit, should be legal (except for fair and noninfringing uses).
Had you read my post, you would perhaps have noticed this is what I said. Oh well, sometimes that's too much asked.
FSF clearly thinks this is not legal, and I hope they're right about that. Otherwise this would foil the entire GPL.
Umm, I don't think the FSF had any say in this. As far as I know, they didn't own any components included in KDE nor any in Qt. Qt copyright is owned by Trolltech and KDE copyright by those who wrote KDE. And the mere choice of GPL as the license does not give FSF the right to pursue violations unless it's _their_ software that is violated.
If this illegal copying causes less people to buy the book, it suddenly will no longer be feasible to write books, and people will be fully deprived of books. It's taking something that is not yours while depriving others of something they would otherwise have. In other words, it's stealing.
With a similar logic, should people be allowed to chop a few trees from every forest because? Surely there will be new trees growing which couldn't have in a cramped forest. So it's not theft.
While I'm opposed to illegal copying and wood theft, I still maintain that fair and non-infringing uses (which I believe includes access, not only the mere right to view as in eBooks) are far too important to squash with a stupid piece of legislation.
I believe you're missing the point, too. Let's see.
1. Reading a book you have legally bought is not stealing.
2. Taking a snippet from a book you have legally bought is not stealing, if done for fair use purposes.
3. Transferring a book you have legally bought to another computer so you don't need to keep that old 286 lying around is not stealing.
4.Removing a measure that "protects" the book from being fed to your Text-to-Speech system is not stealing (whether you are blind or not).
Still, stealing is stealing. Copying this book to your friend would be wrong. Accessing it, no way.
Don't you find breaking into your own house and accessing a book you own at least a bit analogous?
Huh, are you serious? I believe it's still illegal in most EU countries to possess illegal drugs for any purpose. Yes, I live in the EU, and at least here in Finland it's illegal. And, really, I'm happy this way, and I terribly fail to understand your logic.
Huh. What money from shareware?
I'd still say, if they want to help, let them help. As long as it's open source, they can't do much evil.
A bit by bit comparison doesn't really make sense. Neither does frequency response analysis. These encoders were designed specifically to alter these things.
The only thing which matters is, "how good does it sound to the listeners". And that's indeed the most scientific way of testing it I can think of. Blind tests, comparing with real CD quality.
Funny. Licensing issues are the single most important reason for not using MP3, at least for me. Looking for a non-license restricted format? Try Vorbis.
The entire issue of how to encode sound is left, perhaps unsurprisingly, to the encoder. If you find a decoder (player) which gives crappy output on a good mp3/ogg file, well then that player is crappy. Yes, I know that such players exist.
Reverse engineering also has nothing to do with this, and reverse engineering an algorithm doesn't really make it "non-native" or crappy in any other way.
I'm not a MS fan either, I use Linux exclusively (at home, anyway).
But one thing I always wonder is people saying that the Windows OS/kernel is crap and office software is their strength. I would say it's vice versa - I can use Windows (although I prefer Linux), but Word I wouldn't touch with a 5 foot long pole. Installing LaTeX (or some other tool, suitable for the task at hand) is usually the first thing to do when I have to use Windows. And the Windows kernel's not really that lousy, although a little buggy (but certainly better than the office tools, I'd say).
Forget not bp.
Would you change your name if Adobe asked you to?
Well, most of that really isn't an issue anymore: The 1M (instead of 640k) limit can be turned off for protected mode and you can use linear addressing, no need for segmentation (which I guess you meant by saying "weird offsets"). Mem to mem copies are still missing, apart from the MOVS* instructions, which I think have been there from the 8088s.
Agreed, x86 is a real mess for a processor. All kinds of backwards compatibility hacks - the one I most like is the disabling of the A20 (1 megabyte) address line by writing a command to the keyboard controller.
There's one thing I miss a lot on x86 processors, though. I think it would be great to have at least a dozen more general purpose registers (though even 1 or 2 would be an improvement).
Even further, if you refused to give the key, how on earth would they know if you've forgotten it or simply don't want to give it? You certainly can't be jailed or fined for forgetting your own passphrase, can you?
Umh. Somehow misread your comment in a terrible way. Sorry :-(
In general, x.y.z is a development branch if y is odd, otherwise a stable branch.