This is a copyright issue just like any other. Thus the normal court system should be able to make a decision one way or the other. If the reality is that literally no one is capable of adjudicating this issue, then there is literally no teeth to the GPL since who could make a decision on it. I believe that's not really the case though.
And of course lawyers are always going to argue, that's literally their job after all. However, having some actual solid ground here would be helpful since "exceptions" abound, and it's basically "if we like you or we're forced to do because you have market power (nvidia), then it's totally cool". Land on the wrong side of that and you're getting bashed (again note the difference between Canonical distributing zfs vs nvidia's graphics binary).
>If the reality is that literally no one is capable of adjudicating this issue, then there is literally no teeth to the GPL since who could make a decision on it.
Not true at all. Most claims of GPL violation are very straightforward.
True, but we were not discussing general GPL violations.
But specifically the murkiness of the binary kernel module 'exceptions' (or if they are not actually derivative works in the first place, and thus not really 'exceptions', but rights). That's what 'this issue' refers to above.
But you claim that if "no one is capable of adjudicating this issue" then there are "literally no teeth to the GPL".
That is nonsense. Binary modules that may-or-may-not derive from open source programs are an edge case. The GPL can do many useful and powerful things without touching that edge case.
The other poster stated a partial argument that "who the hell are we gonna find that A) is actually qualified to judge this sort of thing". I then stated that IF literally no one was able to adjudicate it, then there would no teeth to the GPL (since anyone who wanted non-compliance would just binary kernel module everything). In other words, for the sake of argument, I was granting the other poster their position, and stating the outcome of that line of logic. Then I immediately disagreed with the base assumption saying "I believe that's not really the case though.".
Continuing the other poster's logical line (again, even though I disagree that a court could not settle the issue), if at some point a court came down and said, "Regardless of the text of the GPL itself, it is completely legal to link binary blobs against the linux kernel, and that linking does not create a derivative work", then the result of that line of reasoning would end in a weakening of the GPL because it would allow companies to 'lock away' the source code in direct opposition to the spirit of the GPL. Also note, that if the courts said something like "We cannot find any person in the entire world who is qualified to judge this issue", then it would have the same effect because abusers of the GPL would know that enforcement was not a possibility (thus the GPL losing it's teeth). I think it's hard to argue that if (in the counterfactual world where there was no one qualified to judge the violation) it would be possible to enforce copyright on the GPL.
I kept that "if" in my wording. I disagree with the line of logic. No matter what you think of the premise.
> (since anyone who wanted non-compliance would just binary kernel module everything)
Even if binary modules made you immune to the GPL, there are a lot of things the GPL would still do. It would not be toothless. On top of that, the issue here is binary modules that questionably derive from open source code. If you make a binary module that unambiguously derives from open source code, that would still be an easy question for the courts. That's why this is an edge case that remains an edge case, even in an alternate reality where the courts literally cannot decide on zfs.ko or nvidia.ko
I'm not sure what you mean by that. 'rhetorical device' is a very broad category. Can you explain if you think that's good or bad, and why?
dsp1234 made an argument that X, though not true, would lead to Y bad thing. I agree that X is not true, but object strongly to the idea that it would lead to Y.
Are you pointing out that the X->Y argument is only for effect, because X is false? I am aware of that, and I disagree with the effect, because I don't think X->Y is at all plausible. So I made a comment.
I agree that X is not true, but object strongly to the idea that it would lead to Y.
If the courts threw up their hands and said "It is literally impossible to adjudicate this issue of the binary kernel module", then who is left to enforce the copyright law? If a court said that, and an individual/company wanted to ignore the GPL, then they could put all of their code into binary modules. If a company's proprietary code is in a binary format, there is no one available to enforce licensing rules, and the company knows those two facts, then what reason does that company have to fear from the GPL?
Please provide an answer to both questions independently, so they are both addressed and no assumptions are left invisible. My answers are 'no one' and 'none'. I'd like to hear what you think the answers to those questions are, then I'd also be curious to hear what you think the logical outcome of the binary kernal module license issue being unenforceable would be.
> If the courts threw up their hands and said "It is literally impossible to adjudicate this issue of the binary kernel module", then who is left to enforce the copyright law?
The court would still be able to deal with the vast majority of copyright law, even if they decided that a specific license term was too vague to rule either way.
> If a court said that, and an individual/company wanted to ignore the GPL, then they could put all of their code into binary modules.
That's not true for two reasons. First, there are GPL violations that have nothing to do with binary modules, like a company embedding busybox and not distributing source code. Second, we are talking about binary modules where the question of "do they derive from the kernel" doesn't have a very clear answer. In most cases if a company tried to hide their code in a binary module, it would still blatantly derive. It would take a lot more engineering effort to get into the same gray zone as nvidio.ko and zfs.ko. Also note that right now, today, you can put in significant engineering effort to create a binary module that isn't affected by the GPL. All you have to do is make a non-GPL program that exposes the same API. But nobody goes through all this effort to skirt the GPL. So I doubt in this alternate world people would spend a similar amount of effort, at least not very often.
> If a company's proprietary code is in a binary format, there is no one available to enforce licensing rules, and the company knows those two facts, then what reason does that company have to fear from the GPL?
The courts would still in the general case enforce the rules, so they would fear legal action.
>I'd also be curious to hear what you think the logical outcome of the binary kernal module license issue being unenforceable would be.
The situation at hand is not one of all binary modules being GPL-proof. But even if they were, even if the GPL could do nothing at all about binary modules, it would still have uses. Anytime someone used a GPL module, you would be able to get source code. It would make the GPL act much like the LGPL.
And of course lawyers are always going to argue, that's literally their job after all. However, having some actual solid ground here would be helpful since "exceptions" abound, and it's basically "if we like you or we're forced to do because you have market power (nvidia), then it's totally cool". Land on the wrong side of that and you're getting bashed (again note the difference between Canonical distributing zfs vs nvidia's graphics binary).