kernel 2.6.8 pwc patches and counterpatches
Quoted 9 lines · Prakash K. Cheemplavam
Quoted 4 lines
Intersting, that in the legal case, people are having a bad feeling, >>but in illegal case of doing reverse-engineering, drivers make it into the kernel...
It's not a matter of legal/illegal, like Linus pointed out; it's a >matter of respecting Nemosoft's work and not stabbing him in the back.
OK and if the authors of, say, SMP support say "back it out", Linux ends up without SMP support. Cool.
GPL is GPL. Period.
I think it's enough courtesy to say "Thank you".
Signature · 4 lines
Thomas Winischhofer Vienna/Austria thomas AT winischhofer DOT net *** http://www.winischhofer.net twini AT xfree86 DOT org
Quoted 2 lines · On Fri, 27 Aug 2004 16:07:04 +0200 Thomas Winischhofer…
OK and if the authors of, say, SMP support say "back it out", Linux ends up without SMP support. Cool.
If you could get each and every author of SMP support to agree, sure. But good luck getting that is quite a large group of folks and since Linus is one of those authors.
And last I checked, no binary-only module hooks were added to the SMP support recently, so nothing for them to get upset about. :-)
Quoted 7 lines · David S. Miller
Quoted 1 line · On Fri, 27 Aug 2004 16:07:04 +0200 Thomas Winischhofer…
OK and if the authors of, say, SMP support say "back it out", Linux ends up without SMP support. Cool.
Quoted 7 lines
If you could get each and every author of SMP support to agree, sure. But good luck getting that is quite a large group of folks and since Linus is one of those authors.
And last I checked, no binary-only module hooks were added to the SMP support recently, so nothing for them to get upset about. :-)
For crying out loud - that was an example. Forget the hook and forget the binary part of the pwc driver.
s/SMP support/RTL8139 driver/g
Get the point?
A license is a license. We can't just obey a single driver's author's daily mood. (And for those who don't know: IAAL. Fact.)
I understand that Nemosoft got mad... what GKH did was a little like calling you an a**hole with a smile on your face. No offense towards Greg, he does a brillant job. Nemosoft will calm down (if his pride permits and we don't make him look like he's losing his face - so shut up).
Don't behave like children, please.
You have no idea how often I god mad over a maintainer's opinion. No names now. Did I ever revoke the sisfb or my XFree86/X.org driver license?
Short version: I just think Greg and Linus are too nice in this very case. Keeping the open source part of pwc in the kernel is - based on the fact that the driver has been committed under the GPL - a matter that shouldn't be subject to discussions at all, of that kind and especially that wording. Hook or no hook. Once again, forget the hook and forget the binary part.
If we can agree on the driver without the hook, fine.
(Let the flamewar begin: The binary pwcx is 0x3000 bytes of asm code. No FPU stuff, just data tables. The 3 algos in question are pretty primitive. Ridiculous. If there is anybody who really wants it, we'll have an open souce C version within a week. People now thinking that I ignore the SCO case please write to me privately. I am ready to explain the difference between patent and copyright law and the world's general legal opinion on reverse-engineering. And even it's the 1013481791873th time. But don't be surprised if you get a form letter.)
Suggestion to Nemosoft (I hope your mail address still works): What about the following model:
1) We have an open source driver (without decompression) in the mainline kernel (which you maintain).
2) You provide a binary-only module which entirely replaces the mainline module? This way you don't need the hook. Just put everything in a single module that, by a patch and a binary, replaces the entire mainline version.
Extra maintainance effort: Zero.
Besides: If your NDA really expired a year ago, what's the point? (Please send me the NDA privately. I can give you a reliable legal opinion.)
Kids, please. How old are you? Please keep our common goal in mind.
Cheers.
Thomas
Signature · 4 lines
Thomas Winischhofer Vienna/Austria thomas AT winischhofer DOT net http://www.winischhofer.net/ twini AT xfree86 DOT org
Quoted 3 lines
A license is a license. We can't just obey a single driver's author's daily mood. (And for those who don't know: IAAL. Fact.)
Since you are a lawyer you know what "moral rights" are. And since you know what those are, I assume you did the same research about moral rights in the Netherlands (the country the PWC work was created) as I did and noticed that this issue isn't as black and white as you claim in your simplification I quoted above...
Quoted 8 lines · Arjan van de Ven
Quoted 1 line
A license is a license. We can't just obey a single driver's author's daily mood. (And for those who don't know: IAAL. Fact.)
Since you are a lawyer you know what "moral rights" are. And since you know what those are, I assume you did the same research about moral rights in the Netherlands (the country the PWC work was created) as I did and noticed that this issue isn't as black and white as you claim in your simplification I quoted above...
See, I know very well what "moral rights" are. But I am having a hard time seeing your point here.
Stating that things aren't just black or white gains us nothing in this discussion. This is like saying "you might be right, but you might also be wrong".
But the fact that this is a response to my previous statement, I take it you don't share my opinion.
Well, here is how I see things:
The GPL is the legal AND "moral" "glue" of the Linux kernel. It is the basis of our cooperation, and it defines our relation to the user.
A license is a contract. A contract is a *mutual* agreement on rights and obligations.
I am pretty sure, even the Netherlands rate "pacta sunt servanda" (as a "moral" law since at least stone age as well as a written law since the Roman empire) above obeying a child's immediate overreaction to a more or less painless slap on his rear. Even on a morality scale since "law" in any regard in the end is the result from a common "moral" of a society.
I confess, I did not study dutch law. Neither did I study nigerian, indonesian, or mexcian law in the very regard. But I know at least that dutch law is based on a mixture of old german and roman law (if Hugo Grotius wasn't entirely wrong). And both of these, I did study, and assume that the principles still apply.
If things are handled differently in the Netherlands, please educate me. Honestly, I am willing to learn.
Again, no personal offence intended. Any emphasis for the sake of clearity.
Thomas
Signature · 4 lines
Thomas Winischhofer Vienna/Austria thomas AT winischhofer DOT net http://www.winischhofer.net/ twini AT xfree86 DOT org
(Reply to Arjan's mail further below)
To avoid misunderstandings:
Quoted 2 lines · Thomas Winischhofer
See, I know very well what "moral rights" are. But I am having a hard
time seeing your point here.
"Moral rights" = copyright law term here.
[...]
Quoted 1 line
The GPL is the legal AND "moral" "glue" of the Linux kernel. It is the basis of our cooperation, and it defines our relation to the user.
"Moral" meaning "morality" here, not to be mixed up with the copyright law term of "moral rights".
[...]
Quoted 6 lines
I am pretty sure, even the Netherlands rate "pacta sunt servanda" (as
a "moral" law since at least stone age as well as a written law
since the Roman empire) above obeying a child's immediate overreaction to a more or less painless slap on his rear. Even on a morality scale since "law" in any regard in the end is the result from a common "moral" of a society.
Again, read "morality" or "morals".
Quoted 4 lines · Arjan van de Ven
Ok then you know that
1) there are moral rights you can't sign away no matter what contracts etc there are
Yes.
Quoted 1 line
2) the amount of moral rights changes per country
Yes.
Quoted 3 lines
3) berne convention roughly says "respect the copyright law from country f author" which is why dutch moral rights are relevant since the author lives in .nl
Yes. This is actually an important point for this discussion: Other countries' copyright laws DO NOT MATTER here. (Hence no need to broaden this discussion beyond dutch copyright law.)
Quoted 8 lines
Quoted 1 line
A license is a contract. A contract is a *mutual* agreement on rights and obligations.
I'm not sure the "license is a contract" thing is entirely valid in
europe.
License and contract are separte items; a contract can give you a license, but it's not clear that a license also is a contract. (because in europe, a contract without BOTH parties agreeing to eachother is not a real contract, and EULA's and such kinda are deep
into the gray area of that). But it's besides the point mostly.
Granted. The details of contract law, with respect to declaration of intention and declaration of acceptance, are somewhat a grey zone.
But if am not very mistaken, *all* juridical systems in Europe see a license as a contract. For the legally uneducated, this seems somewhat odd for the case of the GPL (since there is no real return) but "looks" entirely ok for license agreements where the license is granted in return for money.
But even donations are contracts and not unilateral transactions. The donor cannot claim the donation back (without very limited reasons that are not relevant here). Nothing else applies to a license without return.
The GPL itself, in clause 5, requires "acceptance". That is also a clear indication of its contractual nature.
And now for the moral rights (now: copyright law term):
Quoted 4 lines
Ok so in .nl the moral rights an author has are
1) The right to be visible as author in/with/at publications of the work
Not an issue here. (If it were, it would have been previously as well.)
Quoted 2 lines
2a) The right to object to publication when it is done such that it
appears that someone else is the author
Not an issue here. (If it were, it would have been previously as well.)
Quoted 2 lines
2b) The right to object to publication of the work under a different
name of the work
Not an issue here. The name is "pwc" or "pwcx". I am not aware of any intentions of changing this.
Quoted 3 lines
3) The right to object to any other modification in/to the work, unless the modification is of such nature that it is unreasonable to object to
See below
Quoted 2 lines
4) The right to object to mutilation/damage of the work when it hurts the good name or dignity of the author
Not an issue here.
These rights are the usual moral rights (now: copyright term) of most copyright laws I have come across yet.
Quoted 1 line
3) is the key one here but even 2) is somewhat. If the author wants to be rude and makes claim based on these,
... and this claim is validated by a court ...
Quoted 2 lines
nobody can distribute the
work under the GPL anymore since those requirements would then be a direct conflict with that license.
Granted. However, I don't see either 2) or 3) as a problem here.
2): As said, as long as the name isn't changed, there is no basis for a claim. And for 2a): If this has not been valid for the last 4 years, why should it be in case we keep the driver in the kernel now? The GPL states that copyright notices (and this is nothing else than a pointer to the author) are to be left intact. If the GPL is obeyed (and that is assumed, of course), this moral right isn't touched either.
For 3): The limit for any claim at court is chicanery, even if the claim by the letter of the law is valid. This is also expressed by "[un]reasonable" in the wording, and is a general principle of western juridical systems. "Chicanery" is commonly defined by "raising a claim where the intention to persecute or damage the defendant outweighs any other interests of the plaintiff."
The intention of this clause (3) is to avoid changes that compromise the copyright holder's creativity. But keep in mind that the "idea" or any algorithm is not covered by copyright law.
We are only speaking about the open source part of the driver, not the binary part. The binary part is outside the scope of the license of the open source part.
The open source part and the binary part is not "one" work. It is two. The one and only "work" is the open source part.
Of all changes I can think of (apart from removing the hook, see below), I find none whose nature would make it reasonable to object it beyond pure chicanery.
As for the hook, one could argue that removing the hook is such a change: It could compromise the work (respectively the author's creativity) by damaging its interoperability capabilities. I tend to assume that is a valid point, based entirely on the moral rights.
I will continue here later today. Time's up for now.
Quoted 3 lines
Now, it's not black/white. There is a clause in the same law that says that an author can chose to abandon rights 2a), 2b) and 3) "insofar as modifications to
the work are concerned" and one can argue that releasing under the GPL by the original author is such abandoning. But that is not quite clear.
So, in my opinion, the moral rights Nr 2) aren't even damaged by keeping the driver in the kernel. Basically, it is not even required to abandon these rights.
As regards 3), I will continue later as said.
Thomas
Signature · 4 lines
Thomas Winischhofer Vienna/Austria thomas AT winischhofer DOT net http://www.winischhofer.net/ twini AT xfree86 DOT org
Quoted 5 lines · Arjan van de Ven
Since you are a lawyer you know what "moral rights" are. And since you know what those are, I assume you did the same research about moral rights in the Netherlands (the country the PWC work was created) as I did and noticed that this issue isn't as black and white as you claim in your simplification I quoted above...
The GPL is very careful to explicitly grant rights in the text. There is a reason for that. Pwc should stay in the kernel.
Quoted 11 lines · Alan Cox
Quoted 5 lines · Arjan van de Ven
Since you are a lawyer you know what "moral rights" are. And since you know what those are, I assume you did the same research about moral rights in the Netherlands (the country the PWC work was created) as I did and noticed that this issue isn't as black and white as you claim in your simplification I quoted above...
The GPL is very careful to explicitly grant rights in the text. There is a reason for that. Pwc should stay in the kernel.
As much as I share your opinion, unfortunately, it's not that simple. However, I will get back to that subject as soon as I have time.
Thomas
Signature · 4 lines
Thomas Winischhofer Vienna/Austria thomas AT winischhofer DOT net http://www.winischhofer.net/ twini AT xfree86 DOT org