Licence text and fabs
Licence text and fabs
Posted Oct 4, 2006 14:16 UTC (Wed) by mingo (guest, #31122)In reply to: Licence text and fabs by nim-nim
Parent article: Busy busy busybox
The GPL even in it's v2 form controls more than the code itself. For example you can not claim you can only provide source on gold plated media available at extorsionate price (that is unless you can prove you have to use gold-plated media yourself).
that is different. I'm not talking about common-sense definitions about how the "payment" for our creative works (which payment is mostly in the form of providing source code) happens.
The thing i'm talking about here is that the GPLv3 includes another work (the key) which is independent of our GPL-ed works, in the definition of Source Code, which source code we purport to license!
(and then later on the GPLv3 uses this "defining of other people's creative works into our license's Source Code" act as a tool to prevent the hardware from being redistributed, unless the newly-defined "source code" (which now includes another person's independent creative work!) is provided too.)
Posted Oct 4, 2006 20:51 UTC (Wed)
by nim-nim (subscriber, #34454)
[Link] (18 responses)
I can't imagine how anyone could construct the key as a "work", it costs nothing to create or duplicate.
Take the same hardware with and without DRM. It should be painfully obvious the only "value" "created" by adding DRM and making it apply on GPLed software is the possibility to stop the accesses the GPLv2 used to imply reneging on part of the "payment" for GPLed code use.
Someone has been breaking a bargain. Someone has been extending itself unilaterally without counterpart. Is this someone the FSF? I don't think so
Posted Oct 4, 2006 21:55 UTC (Wed)
by mingo (guest, #31122)
[Link] (17 responses)
I can't imagine how anyone could construct the key as a "work", it costs nothing to create or duplicate.
Let me give you an example of a 1024-bit key:
(The fact that it's easily copied - like any piece of digital information - has no bearing on the strength of copyright protection.)
Posted Oct 5, 2006 6:34 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link] (16 responses)
Copyright "works" are only protected if they are the produce of human creative expression (could not have been created by a stupid computer alone)
99,99999% of DRM keys obviously aren't.
Now you artificial unrepresentative example isn't, but IIRC in a similar case when Sega tried to lock its console by checking for the Sega word/logo at the beginning of ROMs, a judge ruled since the Sega word/logo use was 100% technical and the check could have been implemented by Sega the same way with a non trademarked/copyrighted/protected string/bitbucket, using Sega instead didn't win them any special additional protection.
Which again is academic because actual DRM keys do not look like yours.
Posted Oct 5, 2006 7:30 UTC (Thu)
by mingo (guest, #31122)
[Link] (3 responses)
Now you artificial unrepresentative example isn't, but IIRC in a similar case when Sega tried to lock its console by checking for the Sega word/logo at the beginning of ROMs, a judge ruled since the Sega word/logo use was 100% technical and the check could have been implemented by Sega the same way with a non trademarked/copyrighted/protected string/bitbucket, using Sega instead didn't win them any special additional protection.
well, what i wrote into the key wasnt a simple word/logo. And yes, it can be used as a totally valid cryptographic key - it has enough random content to be cryptographically strong, and it has enough expressive content to be copyrightable. And if it's kept secret by the hardware, it can very much be a trade secret as well. The Sega logo is hardly a trade secret. Do you have any case law that says that trade secret works that are /independent/ of the copyright holder's work are trumped by copyright law? In my reading it would be contrary to the letter and the spirit of both the law and the Constitution.
This was a relatively small key i constructed, but what if that key includes actual, functional code as well, which would have to be executed by the CPU for the hardware to work correctly. What if that key is your family inheritance that your ancestors signed their legal documents with for decades, and which includes expressive content that has sentimental value to me as well. Your assumption that just because it's used as a key it cannot have expressive properties was false when you first said it, and it is still false now.
And you are still missing the larger point as well. This portion of the GPLv3 is a seemingly small but dramatic departure from the spirit of the GPLv2, which explicitly said that it does not attempt to control other works (unless those works being based on GPL-ed code), be that source code, keys or anything else. GPLv2 Section 2 says:
this language is gone from the GPLv3, and the DRM language does a first step towards using distribution rights to control other works. Today it is only used to control certain keys. So far I have not seen anyone even try to limit the scope of this new step, it seems the party line of the FSF is that "everything that /we/ think or ever thought would be needed for the 4 freedoms is a fair game". Would you ever sign a contract with that kind of open-ended language in it?
And in what way will you change the bargain next time around? Until today it was in essence source code for source code, and it took 10 years to explain even this simple concept to developers. Tomorrow it's the hardware's keys too in some cases, the day after tomorrow it will be what? Will the GPLv4 say:
Will the GPLv4 perhaps include:
is that the plan?
Future contributors will be asking me that, and I have no answer for them, do you have any?
We are going to lose alot more contributors this way than due to this uncertainty of the deal than whatever freeloading might become possible via DRM. Freeloading was always possible - heck, a simple end-user of GPL-ed code who never contributes back and never distributes GPL-ed works is a "freeloader" almost by definition.
I'm not worried about freeloaders as long as it's economically unfeasible for them, both the GPLv2 and the GPLv3 has loopholes that enable freeloading and leaching. I'm alot more worried about having a fair and clear deal towards developers, because they wrote this 1 billion lines of GPL-ed code, not the FSF.
Or was this section of the GPLv2 just a false promise to lure developers into writing 1 billion lines of code?
Posted Oct 5, 2006 8:45 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link] (2 responses)
Of course, that's assuming no developper shares the FSF concerns, which is patently false. You can no more claim the developers support as the whole that the FSF can
Posted Oct 5, 2006 9:18 UTC (Thu)
by mingo (guest, #31122)
[Link] (1 responses)
And you are still missing the larger point as well. This portion of the GPLv3 is a seemingly small but dramatic departure from the spirit of the GPLv2, which explicitly said that it does not attempt to control other works (unless those works being based on GPL-ed code), be that source code, keys or anything else. GPLv2 Section 2 says:
this language is gone from the GPLv3, and the DRM language does a first step towards using distribution rights to control other works. Today it is only used to control certain keys. So far I have not seen anyone even try to limit the scope of this new step, it seems the party line of the FSF is that "everything that /we/ think or ever thought would be needed for the 4 freedoms is a fair game". Would you ever sign a contract with that kind of open-ended language in it?
And in what way will you change the bargain next time around? Until today it was in essence source code for source code, and it took 10 years to explain even this simple concept to developers. Tomorrow it's the hardware's keys too in some cases, the day after tomorrow it will be what? Will the GPLv4 say:
Will the GPLv4 perhaps include:
is that the plan?
Future contributors will be asking me that, and I have no answer for them, do you have any?
Posted Oct 5, 2006 11:19 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link]
To be honest, I didn't feel I needed this portion of your argument needed to be rebuted, as better people than me (Alan Cox for example) already stated they didn't feel it was the case.
I happen to agree with them.
Posted Oct 5, 2006 7:31 UTC (Thu)
by mingo (guest, #31122)
[Link] (3 responses)
And no-one has explained it to me yet how it would reduce the amount of DRM in hardware if the GPL adopted anti-DRM language. All the indicators are that such a step /increases/ the amount of DRM done, because a fair portion of current embedded Linux would switch to other OSs. (partly because they are forced by other content makers to use DRM, partly because they'd sense a fundamental uncertainty in the licensing foundations of Linux.)
Also, the FSF should be well aware of the fact that it's always the first layer of software that matters for DRM, most of the time. So it's the kernel that has to deal with hardware and DRM issues. It's the kernel's contribution rules that you are playing with here. In fact without the kernel moving to GPLv3 the whole DRM section is almost totally pointless. (unless you count Hurd, which would have to remove tons of Linux code to begin with, if it wants to switch to the GPLv3) The remaining free software projects are mostly just curious bystanders. Yeah, i oppose DRM just as much as you do, but do /you/ have to deal with the fallout of this happy activism?
Posted Oct 5, 2006 8:50 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link] (2 responses)
Read the transcripts of MP discussions when local EUCD/DMCA law are passed. One question they ask is "it will help Hollywood, but will blanket legal DRM protection/mandate hurt any other economic actor ?". Of course the major representatives try to convince them the answer is no.
The GPLv3 is a very clear way to show MPs blanket legal DRM protection/mandate hurts someone.
Posted Oct 5, 2006 9:37 UTC (Thu)
by mingo (guest, #31122)
[Link] (1 responses)
The GPLv3 is a very clear way to show MPs blanket legal DRM protection/mandate hurts someone.
How would that happen in practice? Lets take Tivo, an FSF-selected example (presumably because it's the best example of harmful DRM they could find), ok? How does the GPLv3 show that blanket legal DRM protection hurts someone, in the Tivo example? Lets assume the kernel goes GPLv3. Please outline the scenario of likely action that would/could happen in your opinion.
The only way i could theoretically imagine would be for a large non-DRM hardware base with GPLv3 kernel+userspace to be built up right now, which industry would complain loudly if a change in laws would made their product illegal to distribute under the GPLv3. Given that essentially every medial player and every mobile phone on the market includes some form of DRM currently, as per the wishes of the content industry (whose content consumers actually want to enjoy), how do you see the likelyhood of that happening in practice, /without/ a huge body of free content that consumers would be equally crazy about?
Please outline the scenario that you think would/could happen.
Posted Oct 5, 2006 10:51 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link]
Very easily.
Mr V. goes to his local MP to make DRM mandatory and protected by law on all media processing boxes. The MP asks if it will hurt someone. Mr V says "of course not, it will only catch evil pirates".
Mr F can then come saying "But this will ban use of FLOSSTV, which is licensed under the GPLvx which is uncompatible with the proposal" competition and free market is good, you can't do that. This is a strong argument.
Right now M F can only say "But this will harm FLOSSTV, which is licensed under the GPLvx which intent/spirit is uncompatible with the proposal" competition and free market is good, you can't do that. This is a weak argument, spirit and intent count for zip.
In case you think I'm making all this up, I'm only repeating from memory some of the arguments used in the French Parliament when the EUCD directive was translated in local law this year.
Some of the articles proposed by the majors were clearly uncompatible with existing Free Software license terms and could be amended. For others the uncompatibility was not so clear cut and unfortunately opposing them proved impossible.
The harm BTW was not only FLOSS-side. The law hurt numerous actors, but they all found out nothing but the most blatant problems (such as law/license uncompatibilities) had any weight against the majors' money.
Posted Oct 5, 2006 7:36 UTC (Thu)
by mingo (guest, #31122)
[Link] (7 responses)
Which again is academic because actual DRM keys do not look like yours.
[sarcasm] Oh, thank goodness no-one else will think of this and they will not produce DRM keys that are copyrightable. Phew, that was close, i was almost getting worried! [/sarcasm]
Posted Oct 5, 2006 8:42 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link] (6 responses)
But you know what? I don't care. They'll be laughted out of court since judges have a very low tolerance for this kind of legal trick.
The judge's test will be the following: are the copyrighted/trademark/whatever parts functionnaly necessary? If the answer is no and they've been knowingly inserted only to make copyright/trademark/whatever law conflict with another legal requirement, game over. You're trying to game the law and wasting the judge's times. Too bad for you if you loose some IP property in the process, no one forced you to expose it.
Even if someone managed to construct some functionnal reason for the use of these parts, copyright/trademark/whatever law wouldn't win automatically -- the judge would balance one requirement against another, and the more artificial convoluted one would lose.
Posted Oct 5, 2006 9:12 UTC (Thu)
by mingo (guest, #31122)
[Link] (1 responses)
Sensing some fundamental disagreement that causes us to argue circularly
i have read back some of your other replies in other threads and I think your core argument in favor of DRM is in one of your sentences:
Does that fairly summarize your points around DRM up in essence?
If yes then i contest your suggestion above. (i thought i did that before, but hey)
GPLv2+DRM still gives us the source code. BSD doesnt. Tivo gives us their source code for the DRM-ed kernel they utilize. So where does the GPLv2+DRM combination create the "BSD on GPL code" situation in the Tivo case? I'd really like to understand your logic here.
I'd agree with you that if GPLv2+DRM would be equivalent to the BSD then that would not be acceptable to me and i'd find little joy in hacking the kernel. (when i started developing Linux years ago i made a conscious choice against BSD, mostly due to their license)
Posted Oct 5, 2006 11:16 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link]
Very easily, you only need to make DRMed PCs the cheap norm and open PCs the expensive exception.
for every-market-without-GPLv3-software ; do
// Making exceptions costs money and the majors don't like hairies
// Need some GPLv2 devs with access to DRMed hardware
// This only takes a few years with modern lifecycles
// This takes a tad longuer, as hairies will hoard previous gen hardware
// Without hobbyists there are less floss devs on the job market
note (floss-in-market-nonexistent);
This is the usual BSD pattern of hire key community people, stop contributing downstream, redeploy once the community is anemic
Posted Oct 5, 2006 10:15 UTC (Thu)
by mingo (guest, #31122)
[Link] (3 responses)
The judge's test will be the following: are the copyrighted/trademark/whatever parts functionnaly necessary? If the answer is no and they've been knowingly inserted only to make copyright/trademark/whatever law conflict with another legal requirement, game over.
But this is not "another legal requirement". This is the GPLv3's attempt to extend its scope to works created independently of that work. How do you know that it's "defendant laughed out of court" instead of "GPLv3 plaintiff's copyrights being judged unenforceable, due to misuse of copyright"?
Posted Oct 5, 2006 11:33 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link] (2 responses)
As the GPLv2 enforcements efforts have shown, a judge will tend to consider someone availing himself of some provisions of a licence (free code access) has no legs to stand on contesting the obligations part later (as long as the obligations are clearly expressed).
Nobody forces anyone to use copyrighted/trade secrets/trademarks as DRM keys.
Posted Oct 5, 2006 14:25 UTC (Thu)
by mingo (guest, #31122)
[Link] (1 responses)
As the GPLv2 enforcements efforts have shown, a judge will tend to consider someone availing himself of some provisions of a licence (free code access) has no legs to stand on contesting the obligations part later (as long as the obligations are clearly expressed).
Oh, the GPLv2 enforcements so far were mostly about clear-cut violations like "not giving out the source code", or trying to hide the origin of the code and copyright ownership by modifying binaries.
I believe you are over-simplifying the picture. A more realistic case would be like this:
- hardware maker fully living up to all provisions of the GPL. Source code made available on their website, etc.
- separate content provider (provider of expensive GPS maps, movies, pay-per-view events, etc.) insists on copy protection of their content.
- hardware maker use of DRM to restrict the extraction of that content from the hardware, by only allowing trusted installs approved by the hardware maker.
GPLv3 copyright owner sues. The judge will weigh: on the one side, the legitimate interests of the content owner, and the trade secret of the hardware maker, and the contractual obligation of the hardware maker to not make the content copiable, against the interest of the GPL owner, who gets his modified source code. The hardware maker even contributed back to the copyright owner's project.
But besides the source code (which was the main object of trade in a previous version of the same license, which was "similar in spirit"), the GPL work owner now also insists on a new-found and pretty far-fetched "right to tweak the hardware".
The hardware does not look all that tweakable to the judge, it's a sealed pink plastic cube. The copyright holder insists on the defendant either ceasing distribution (which likely bankrups the defendant) or requires the giveout of the trade-secret crypto key (which likely bankrupts the defendant too, because he was contractually obligated to the content owner to protect stuff and to not give out the key).
Furthermore, the judge will look who the people with this "right to tweak the hardware" are. They describe themselves as certain "hackers", and they want to "hack on the hardware". To do good stuff, amongst them, as the defendant will point out, "to enable pay-per-view piracy".
On the other hand, the judge will weigh the "harm" that the denial of this request for keys causes the copyright owner, but that harm seems rather remote: the copyright owner has only a personal interest in modifying the hardware. Out of curiosity. No matter how difficult technically. And he could run the manufacturer's code on his home PC. No, of course he does not want to watch pirated content.
Do you think the decision is so clear-cut against the defendant, especially if the defendant insists on a jury trial?
Posted Oct 5, 2006 15:22 UTC (Thu)
by nim-nim (subscriber, #34454)
[Link]
1. hardware maker uses GPLvx code -> he's now bound by the license requirement
2. here comes content provider, asking something incompatible with the hardware maker licensing obligations :
No judge will defend someone promising to do foo to A and not-foo to B.
> The thing i'm talking about here is that the GPLv3 includes another workLicence text and fabs
> (the key) which is independent of our GPL-ed works, in the definition of
> Source Code, which source code we purport to license!
Licence text and fabs
mQGiBDlqJigRBADpPRhgY7lPy42Hy1gk6YKT81QOGkptbUuSkKZ4lSvGT5DEsmwG
q6/l+/WdeE9401Njs55bnztcetCzcadyBIxPJz+JpzbaOjwdxTZuLHstQ5+ddK54
fdDBw76W0N/DZTV/rypPvl0p9NtnmT+Q5tWF1j/s7bLxyNBBNxRpiBm+xwCg/zEy
2YVUb9hnsna1Lknp4sxGBN8D/RvG4wXHLzFIsEjY4qlY+W0hN3dNW0mBuqLEZ3gD
E1rZyI9/6 wHQeGIB
m5euYv1YJ This chunk of metal, like it or not, 1ASDFlk
WIAfA/4kS Is our love and hate, my bright spot. xAIQusQ
cfSBBwn68 Treat it with respect, dear Alice, AOX9D/v
VKb/Ia5Wb To protect its keys, we promise. y1IYXNz
byBTdGFtZ WomKAQL
AwIBAAoJE Copyright, 2006, Ingo Molnar. 8WD8g0T
mo9TCht0r 2AmDLM9
j2R35j0DF9sxrMLwz2JhAJ9eUCud0CxsVANJg3BsTvftvbgVIIhGBBARAgAGBQI7
Wm7tAAoJEMGC+zEYbLxm1OMAnjlzJ2MZJ2yoBzagnLesayo1qHhxAKC32HkRdxt2
bZSCQrDA7CftJLYBjLkEDQQ5aiYpEBAA+RigfloGYXpDkJXcBWyHhuxh7M1FHw7Y
4KN5xsncegus5D/jRpS2MEpT13wCFkiAtRXlKZmpnwd00//jocWWIE6YZbjYDe4Q
tu+Pp0NGCMbMHXdXJDhK4sTw6I4TZ5dOkhNh9tvrJQ4X/faY98h8ebByHTh1+/bB
This work, besides still being a pretty strong cryptographic key, is also a literary work, a creative original expression, and is copyrightable.
Bzzt. Thank you for making my pointLicence text and fabs
Licence text and fabs
Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.
You must ship 10 free samples of hardware and design specifications
to the FSF headquarters and offer free on-site training and free
tools for tweakers to be able to make use of their right to tweak.
You may only distribute copies of works that are GPL-compatible in their
totality. Aggregation with any non-GPL work is prohibited.
> I'm not worried about freeloaders as long as it's economically unfeasibleLicence text and fabs
> for them, both the GPLv2 and the GPLv3 has loopholes that enable freeloading
> and leaching. I'm alot more worried about having a fair and clear deal
> towards developers, because they wrote this 1 billion lines of GPL-ed code,
> not the FSF.
Is there any reason why you have not replied to my points below, i think they are important:
Licence text and fabs
Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.
You must ship 10 free samples of hardware and design specifications
to the FSF headquarters and offer free on-site training and free
tools for tweakers to be able to make use of their right to tweak.
You may only distribute copies of works that are GPL-compatible in their
totality. Aggregation with any non-GPL work is prohibited.
> And you are still missing the larger point as well. This portion of theLicence text and fabs
> GPLv3 is a seemingly small but dramatic departure from the spirit of the
> GPLv2,
Licence text and fabs
> And no-one has explained it to me yet how it would reduce the amount of DRMLicence text and fabs
> in hardware if the GPL adopted anti-DRM language
Licence text and fabs
> How would that happen in practice?Licence text and fabs
Licence text and fabs
Oh, some people will think of it.Licence text and fabs
Licence text and fabs
GPLv2+DRM happens to enable corporations to do the BSD on GPL code they
wanted all along.
> So where does the GPLv2+DRM combination create the "BSD on GPL code"Licence text and fabs
> situation in the Tivo case
{
// Get MS/Hollywood money and help
release (spec-mandating-DRM-on-next-hardware-gen);
apply (DRM-on-GPLv2-software);
// Or the FLOSS ecosystem will collapse brutaly, regulators notice
// and floss users not buy the new hardware generation
hire (some-gplv2-devs-to-do-life-support);
wait (DRM-hardware-is-the-norm);
// However without hardware access they're bound to disappear and find
// a cheaper hobby
wait (floss-hobbyists-have-disappeared);
// All the new software contributions come from competitors
// May as well make a consortium and work together
// Denying entry to new would-be competitors
replace (GPLv2-software,consortium-software);
}
Licence text and fabs
> How do you know that it's "defendant laughed out of court" instead of "GPLv3Licence text and fabs
> plaintiff's copyrights being judged unenforceable, due to misuse of
> copyright"?
Nobody forces anyone to use GPLv3 software.
Licence text and fabs
The answer is simple :Licence text and fabs
A. hardware maker can refuse
B. hardware maker can accept to breach his legal obligations, and be sued
C. hardware maker can accept, replacing the GPLed code with something compatible with the content provider demands. He won't be sued. If he's smart, he'll make the content provider pay for the change