Slashdot Mirror
Doubts Raised About Legal Soundness of GPL2
svonkie writes "Two prominent IP lawyers have warned that the all-pervasive General Public License version 2 (GPLv2) is legally unsound. They claim GPLv3 and AGPLv3 are much better suited for the realities of modern open source software. 'If you go back in time to when GPLv2 was written, I don't think people were aware of just how ubiquitous this license would become and how closely scrutinized it would be,' said Mark Radcliffe, partner at the firm DLA Piper and general counsel for the Open Source Initiative (OSI). 'At that time, open source was not something as broadly used as it is now.' Radcliffe was joined by Karen Copenhaver, partner at Choate Hall & Stewart and counsel for the Linux Foundation, for a GPL web conference hosted by the license-sniffing firm Black Duck software"
The article essentially says that the terminology used needs more rigorous definition, and needs to match more closely with the existing legal terminology. For example, their use of "derivative work" might have legal connotations that don't completely follow from the terms of the licence. It's not like they've determined there's some fundimental legal principle which brings the whole thing crashing down, as you see in EULAs for example.
No kidding!!! What do you say at this point?
In other news, Darl McBride was seen dancing a little jig at the corner of 42nd and Broadway in New York City.
A source close to the situation informed Slashdot that he was in fact accepting small change to offset his legal fees for the next phase of his litigation against Linux users.
So, I actually count myself among the few that like Richard Stallman. I've met him, and he's a nice guy. But does anyone recall the furor over GPLv3 when it first came out, & some of the new provisions? This caused a lot of projects to stick with v2.
I'm wondering if this isn't just FUD to try to get people to switch to v3. Which is icky, but it did occur to me.
~dijjnn
They claim GPLv3 and AGPLv3 are much better suited for the realities of modern open source software.
...
"At that time, open source was not something as broadly used as it is now."
Maybe the way it was written is why FOSS is where it's at? Might not be such a bad idea to keep it around?
No comprende? Let me type that a little slower for you...
and 12 Microsoft employees have been assigned to post on /. using their wit and intelligence. If they can find it.
If you just take a GPL project, add a new UI skin and sell it in binary form, judge will make you release the source notwithstanding the license ambiguities. If what you are doing is not clearly a derivative work (like code inspired by reading a textbook) or you have a reasonable fair use case (like using the client part of client/server stack which is complex and not documented except for existence of the code itself), it's a good thing that the license will not be enforced.
Hindsight is 20/20.
This argument can be made for most of what's written into law. Where's the news here??
The plan was to create a proprietary-destroying license that spreads by itself and becomes more powerful the more it's used, but you couldn't do it all at once, because it would be too shocking a change to introduce. Hence, the GPL v2 was created as an interim step to soften people up, although RMS's goal was always v3. AGPLv3 was just plugging the gap he forgot. The plan hasn't been proceeding as fast as they hoped, so they seek to speed it up by spreading fear about v2.
"First they ignore you, then they laugh at you, then they fight you, then you win." - chiseled over the bearded guy's bed.
How much of this is about nudging Linus... pushing him, really... into applying GPL 3 to the Linux kernel?
Life is hard, and the world is cruel
These lawyers have some kind of agenda and should be kept out of the press. The GPL has been applied in court (successfully) many times.
I assume the agenda is to promote the truly awful GPLv3 and spread FUD about the not-so-awful GPLv2 so people will feel forced to 'upgrade.'
TFA claims that the term "derivative work" as used in GPLv2 requires further definition in the GPL itself because courts haven't clarified it, but this is wrong. The authors of the GPLv2 (i.e. RMS) clearly intended it to cover as much as possible: any and all works following under the statutory definition. It's true that software copyright cases are rare so the lgegal system hasn't developed the idea completely -- but that's not the GPL's duty.
But you don't like the patent provision, why not strike the patent language and call it GPLv2.1
Why can't people take responsibilities for themselves, do what is right in terms of copyright, and maybe then we wouldn't have lawyers nit picking us to death. Or as the old joke goes: "What is a thousand lawyers at the bottom of the sea?" Answer: "A good start."
Kosh: "Understanding is a 3 edged sword, your side, their side, the Truth."
How much of this is about nudging Linus... pushing him, really... into applying GPL 3 to the Linux kernel?
That can't happen without a rewrite. Too much of Linux is composed of patches written by unreachable authors and whose copyrights haven't been assigned to Mr. Torvalds or the Linux Foundation.
And even then, Mr. Torvalds has stated that he prefers the spirit of GPLv2 to that of GPLv3. I'm pretty sure that the spirit of GPLv2 can be expressed in the GPLv3 framework by adding a set of exceptions, much like the Classpath license and the LGPLv3 are sets of exceptions to GPLv3.
The license was already proven in court numerous times in different countries. It can definately hold up. I don't care that there are two big IP lawyers. Especialy when you keep in mind the fact that IP has less chance of holding up. IP laywers don't like copyleft for a reason.
Nothing got disproven with that useless article. That, together with you post, will not make an impact at how FLOSS advocates look at the license and it sure as well will not stop them slow them down or even irritate.
Resistance is futile, proprietary pussy.
Here be signatures
This is a real issue. For instance, I wrote a physics textbook, which is open-source, and I wrote a bunch of ruby and latex code that helps to produce the pdf from the latex sources, automatically handling some things relating to placement of figures on the page that are awkward to do with plain latex. My book, including the ruby and latex code, is under CC-BY-SA. I got an email from a guy at MSU who was writing a textbook, and had already started using my code to handle the illustrations. He wanted to check whether it was okay under the license, since he didn't intend to release his own book under a CC license. Well, my answer ended up being that I really didn't know whether it was okay or not. It wasn't clear to me whether his work counted as a derived work. On the one hand, you could say that what he was using was simply some software I wrote, so his book isn't a derived work based on my software any more than a book written in MS Word is a derived work based on Word. On the other hand, there's really no perfect separation between the software and our books. When you write a book in latex, the latex code *is* a piece of software. My code generates various boilerplate in its output, some of which is text that is visible to the reader, so it's under my copyright and license. Of course I could have just told him that it wasn't an issue, and I wouldn't sue him, but I had intentionally chosen the strong copyleft because that's what I wanted. I suspect that a lawyer would tell him his work was actually not a derived work, but I also suspect that he (and his eventual publisher) wouldn't even want to get into that issue.
Although the issue is real, it seems goofy to me to suggest GPLv3 as the fix for the problem. First off, there are huge philosophical differences between v2 and v3. Also, there is so much GPL v2 code out there that you can't necessarily just relicense under GPL v3 without causing yourself hassles with license incompatibilities. I also don't quite understand how they think they can bypass the fact that various countries have various inconsistent and ambiguous definitions of a derived work. The only thing that forces anyone to accept the GPL license attached to a work is that copyright law doesn't allow them to do certain things without a license from the author. Those things include (1) copying and redistributing the work, and (2) creating and distributing derived works from it.
Find free books.
What, so you can legally copy source code because your lazy ass isn't inteligent to code for itself? Luser...
Here be signatures
I wonder who pays these gentlemen. And, again, who pays those who pay them...
"Flyin' in just a sweet place,
Never been known to fail..."
Two Slahdot users are dissing your cowardance is public and got viewed and laughed at by aprox. 1 million readers due to you lack of social influence.
Here be signatures
The point of the GPL was that it was very simple and broad-sweeping. Naturally this does make it vulnerable to attack in the sense that the legal system might feel threatened by the massive impact of such a game-changing license; copyleft practically redefines IP law in a way that those in the legal institution (eg lawyers, judges, lawmakers, and the business interests that pay for them) don't acknowledge, understand, or otherwise feel comfortable with because they don't feel in control. All it takes is a mere technicality to disqualify the GPL from functioning at all, and the Free Software community is justifiably anxious about that, but the GPL has been successfully upheld in court time and time again, so I wouldn't worry.
At this point there should be no doubt of the legal soundness of any version of the GPL, but it all boils down to a matter of principle. If a society believes in Free Software, then the GPL's legal application is perfectly simple and valid. To those hostile to freedom in the society, then the application of the GPL becomes something artificially difficult/problematic.
I wonder who pays these gentlemen.
If you had read the summary you'd see they work for the OSI and the Linux Foundation. Hardly organizations that are anti-GPL, anti-FOSS or anti-Linux.
What, so you can legally copy source code because your lazy ass isn't inteligent to code for itself?
Most people find that mindlessly reimplementing everything isn't a good idea not being lazy. Secondly, how does the GPL not let you legally copy source code?
Attention yankee blockheads!
From the article:
And...
I don't understand how 'propagate' and 'convey' are any better than 'distribute'. It would seem to me that 'distribute' is the better term.
Now this, I understand. If you borrow the phrase 'derivative work', which is defined by copyright laws, and then the definition of 'derivative work' is changed in the copyright laws, does that mean the definition in the GPLv2 has also changed? Or is the definition of the GPLv2's 'derivative work' the same as that defined by copyright laws at the time the GPLv2 was drafted? Or is it some third definition not in any way tied to the copyright law definition? If it's the second or third meaning, then you're probably better off using a different phrase altogether.
I could also see a slight advantage to tying the GPL's definition of 'derivative work' and other phrases to whatever the current copyright law says they are: As copyright laws get more restrictive, the force of the GPL gets stronger. If copyright laws ever relax, then so too does the GPL. As I see it, as long as copyright laws are sane, the need for the GPL lessens.
We always knew Comcast was corrupt, here's the proof: http://tech.slashdot.org/comments.pl?sid=1909890&cid=34545432
I wonder who pays these gentlemen. And, again, who pays those who pay them...
Blackduck is founded and stocked by Microsoft employees. Though it would be damning enough in this context to point out that it is an active Microsoft partner.
SCO was a pre-existing company re-purposed several times, turned pump-n-dump, turned sock puppet. Blackduck was founded from the beginning for the activities it is engaged in.
Beta is broken and the link to classic doesn't work. Stop wasting our time or there won't be anybody left here.
So what? I HAS been proven to hold up, so the article still remains useless...
Here be signatures
GPL2 is not about to become invalid. But consider all of the changes we have gone through since GPL2 came out. Back then, the most complicated input device that people were likely to have in their homes was a touch-tone phone. Music came from phonograph records and cassette tape. The "@" sign was a little-used oddity on the typewriter keyboard for most people. Home computers were more the exception than the rule, and their CPUs used 16-bit addresses.
With the advent of consumer digital media we got a ton of law, both legislative and case law. Garbage legislation like ECPA and then DMCA, and a great increase in software patenting. All of that law essentially blind-sided the GPL, which had to cope with it but was not written with knowledge of it.
So, a license upgrade to deal with all of this is like installing a security patch on your operating system. It's just a sensible thing to do.
Bruce
Bruce Perens.
I was reffering to people who work at companies that write proprietary software and their lazy ass developpers. They cannot copy GPL code because they don't want to use the GPL. People get hostile when their job is jeopardy because of a free of charge and rivaling GLP project.
Here be signatures
It looks like you didn't RTFA.
What they're arguing, is that there might be corner cases (and these have not been in court) where the license isn't going to apply, or do what it is intended to do.
Copyright law (which copyright law? Well, let's say USA's...) plus all the loads of case law that have come from it, defines derivative work a certain way. Assuming you are a well-trained lawyer and have a shitload of money for research and an IQ and Judge-Psi scores of 300 each, you can look at, say, Project Beta and declare whether or not it is a derivative work of Project Alpha.
Let's say Project Alpha is GPL2 licensed.
If Project Beta is not a derivative work, then you don't need to bother to read GPL2 and worry about it's wording, so there's no problem here (for the Project Beta people -- let's get back to this in a little while).
If Project Beta is a derived work of GPL2-licensed Project Alpha, and you don't want to infringe copyright, then you need to follow the terms of GPL2. Now this is where it gets interesting: under the GPL2's definition of derived works -- not the same as copyright's -- Project Beta still might go either way. It might be a derived work under copyright law, but not a derived work under GPL2's defintions, which means that GPL2's terms that talk about derivative works, might not apply. So you might have something that is legally a derivative work, but GPL2 might not give you any distribution rights. Or it might give you distribution (GPL2 defininition) rights, but then legally you might find that you still don't have distribution (copyright definition) rights.
What has gone wrong, is that GPL2 is using copyright-related terms. The license doesn't change, but copyright changes all the time. This can make a big fucking mess and create ambiguities that no one can guess how courts will resolve.
Let's get back to the case where Project Beta is not (according to copyright law) a derived work of Project Alpha. But let's say that it is a derived work according to the definitions in the GPL, if only those definitions actually applied (which the don't). Project Alpha released their code under GPL2 because it does what they want. But now the GPL2 has let them down, because the people who built on their work, aren't having to comply. Maybe that's ok. From a public policy viewpoint, it is ok. But nevertheless, Project Alpha just got "tricked" by RMS into believing they were going to get what they want. (No, I don't really blame RMS.)
You went with BSD because you were willing to make a no-strings-attached gift of your software to everyone. Had you not been willing to do that, you would have found a license that did acceptably what you had wanted from GPL2.
For many of us, sharing-with-rules is more attractive than gift. This is especially true for business, because sharing-with-rules admits the potential for dual licensing. This is one of the few ways to carry out Open Source business that actually works.
GPL vs. BSD is essentially a matter of business (or non-business) strategy. You pick the rules that work for you, and then you pick a license.
Bruce Perens.
I was reffering to people who work at companies that write proprietary software and their lazy ass developpers.
Ahh yes because it's only proprietary software companies that have lazy developers. LOL.
They cannot copy GPL code because they don't want to use the GPL. People get hostile when their job is jeopardy because of a free of charge and rivaling GLP project.
So then all the companies that do copy GPL code aren't lazy, right?
Really? Why is that?
I don't see why they would. It's not like it prevents them from making any money and considering the legalese that the GPL is written in, it looks like a good thing for the IP lawyers - if I had a business that used the GPL in one way or another, I would have legal council. And this article is a prime example of why one would need legal council.
The law isn't like computer code where what's written down can only be interpreted one way. It's not as easy as black and white.
It's NOT me! It's the meds! I'm on 1000mg of Fukitol.
without the device to run the software, how do you "use" software?
You're just bitter because you can't take someone else's code and use it without paying.
It's responsible for FOSS.
And note:
BSD is lame: compared to GPL it's a backwater. Great OS comes from it, but it is an also-ran. Because the license is "give it away!". What company would do that?
Perl is released under GPL too at v5. And the AL 2 is GPL3 compatible, so no difference: you can merge code between them and the rights and responsibilities are equal, so not very different from GPL3.
Poor examples.
I was reffering to people who work at companies that write proprietary software and their lazy ass developpers.
Darn you, Apple! Why are you too lazy to write 100% of your operating system! Real coders write a new operating system every day, and those OSes have no need to ever be compatible in any way.
Kids these days. Next you'll be talking about "functions" and "object oriented" and other such hippy nonsense. Why, I've gotten by for years by copying and pasting all of that code where ever it goes. Now get off my lawn!
SSC
These are lawyers talking about a possible flaw in a legal document.
It's like a coder talking about a bug in a C program. You can say to yourself "Oh, that coder's just causing trouble. There's no real bug" - but that doesn't change the reality that the compiled program will do just what the C instructions tell it to do - nothing more, nothing less. (Well, except that a legal document is interpreted more subjectively - if the document is not written very precisely then different readers may interpret it differently... This is why we live in a world of "legalese" - it sounds arcane and needlessly complicated to non-lawyers but that's just because non-lawyers haven't learned the terminology and practice of precisely phrasing a legal document and guarding against the various rules which may be in effect "by default"...)
We have to be realistic about this - there have been some GPL-related court cases in which the GPL was upheld, but there are also issues surrounding the license, as well as how those are likely to play out in actual court cases, which may be unresolved or unfavorable to people who value software freedom. If one hopes to fix the problem, then one has to be realistic about where and what the problem is. We have a certain set of goals when we license something under the GPL - if we want to actually achieve those goals, we have to do our best to make sure the GPL is legally sound.
The problem of what constitutes a derivative work is one that I've heard before... The problem is that what you or I might consider a GPL violation could in fact be a very small piece of GPL code pulled into a very large non-GPL project. It's difficult to call that a "derived work" of the GPL'ed project. It's using GPLed code but the project as a whole may very well not be "derived" in a real sense from that GPLed work. If the project is big and the bit of borrowed GPL code is small, courts may not take the offense seriously. I don't know if this is something that can be solved with a better license, or if the kind of protection the GPL demands is beyond the scope of copyright... Anyway, it seems like a problem. Even if we want to tell people "you can't reuse parts of this code in other projects unless those projects are GPLed, too", we may not be able to rely upon that demand being fully effective...
Finally, it's worth emphasizing that law is not a static thing. It's a set of agreements between people subject to interpretation and alteration by people. Saying "it works and it doesn't need to change" may not be realistic. If people are working to undermine the GPL, then other people must work to reinforce and improve the license, if it is to be viable in the future. Basically, if the GPL matters to you then you need to fight for it.
Bow-ties are cool.
If you borrow the phrase 'derivative work', which is defined by copyright laws, and then the definition of 'derivative work' is changed in the copyright laws, does that mean the definition in the GPLv2 has also changed?
It doesn't matter how you word it, when the legal definition of the words you use change, what they mean in the license will change. You have to use words that have legal meaning to write your license, if the legal meaning of those words changes, the meaning of your license will change.
The truth is that all men having power ought to be mistrusted. James Madison
You really beleive it all to be this simple? Have you ever been in a room with either lawters from firms of this size, or with the technology marketing teams that fund them?
They are "interesting" rooms.
"DLA Piper is one of the largest law firms in the world and it is the only firm with more than 3,500 lawyers in North America and Europe. DLA Piper is a legal services organization whose members and affiliates are separate and distinct legal entities."
http://en.wikipedia.org/wiki/DLA_Piper
IBM and Novell both are certainly money this firm takes. Probably Microsoft, too. All of these players have "proxies" that they push markets with - like cold war intelligence and battle operations. I have heard the planners at one proprietary vendor - supposedly Open Source advocate - talk about funding CentOS - to damage RedHat. They'd do so through participation in a third vendor consortium/forum. This stuff is common.
"Speaking the Truth in times of universal deceit is a revolutionary act." -- George Orwell
I don't understand how 'propagate' and 'convey' are any better than 'distribute'.
In Anglo-Saxon law, the former terms have legal definitions and a set of case law.
what, you think all work is 100# scratch?
I'd like to remind you of that Kernel thing.
Oh right, why don't you make your own? loser. /barring actual logic and reality to your asinine comment.
This anonymous coward has some kind of agenda and should be kept out of the press. The GPL3 has been applied in court (successfully) many times.
I assume the agenda is to slur the truly wonderful GPLv3 and spread FUD about the so-awful GPLv3 so people will feel able to steal.
This is why it is recommended by the license itself to leave a clause in the license grant that reads "or any later version" which allows license upgrades to be performed by the originators of the GPL and now people without any copyright claim to the software to take a license out under the newest version. It has a repair function and it's built into the grant.
Oh I don't know....GPL V2 has been in and out of the court all over in Europe.
Most settle without a hitch.
Some, need to have it explained in front of a judge.
This guy doesn't know what he is talking about, and if the GPLv2 was unsound legally, it would have long been over turned.
GPLv2 even scares Microsoft's lawyers.
The guy just wants attention.
Send him a cookie.
-Hackus
Got Geometrodynamics? Awe, too hard to figure out? Too bad.
"They're a vendor"
"No they're not"
Thinks: I am NOT WRONG!!!! I'm NEVER wrong!!!!. "they are, they're a meta-vendor!!!"
A leap of weasel wording that would have Darl McBride blanch in horror...
OK, YOU'RE a meta-vendor. Therefore your point of view cannot be considered accurate and is therefore obviously biased.
Really? Why is that?
Because they are paid to dislike it.
Except that this is a story about a "web conference hosted by the license-sniffing firm Black Duck software". Blackduck is hardly going to allow any criticism of its partner, Microsoft, nor allow its major thorn, the GPL, to go unmolested. Go re-read plaintiff's exhibit 3096 about stacking conference panels. Even without a sock puppet organizing the conference, M$ has a prolific history now of interfering with and shutting down conferences on competing (that's everything by the way) technologies.
"So you want to love those conferences to death. I've killed at least two Mac conferences. First there was the Mac App Developers Conference. I was on the Board of Directors of the Mac App Developers Association long ago, and after I left I worked to try to turn it into a cross- platform developers conference, and I did. I managed to make their last conference was very cross-platformn, both Windows and Macintosh, which of course turned off their Macintosh audience; half of the conference was irrelevant to them. They didn't care about Windows. They were a bunch of Mac guys. Which diluted the value of the conference. And they didn't know how to advertise the Windows guys when the Windows guys showed up. So they lost money that year and the group folded. Oh, well. One less channel of communication that Apple canuse to reach its developers." Plaintiff's Exhibit 2456, Comes v MicrosoftWhen you're dealing with Microsoft, you're dealing with cockroaches. Get over it.
Beta is broken and the link to classic doesn't work. Stop wasting our time or there won't be anybody left here.
I wonder who pays these gentlemen.
If you had read the summary you'd see they work for the OSI and the Linux Foundation. Hardly organizations that are anti-GPL, anti-FOSS or anti-Linux.
OSI was founded by people who were unhappy with the Free Software Foundation and the GPL. OSI Founder Eric S. Raymond recently said that the GPL is no longer needed.
It's like a coder talking about a bug in a C program. You can say to yourself "Oh, that coder's just causing trouble. There's no real bug" - but that doesn't change the reality that the compiled program will do just what the C instructions tell it to do - nothing more, nothing less.
Programs don't run by coders arguing about them. Legal documents mean whatever a the last lawyer convincingly pretended they mean as long as that lawyer didn't get get caught by another lawyer.
Contrary to the popular belief, there indeed is no God.
The license was already proven in court numerous times in different countries. It can definately hold up.
But did any of these cases involve the "derivative work" aspects of the license? That seems to be the issue here, not the GPL as a whole.
The one issue I've always disagreed with the GPL folks on, for instance, is code that links to GPL libraries. I know the LGPL was created to allow code to link to libraries without being GPL itself, but what if I distribute, say, a dynamic linking executable which links to GPL code, but I don't distribute the GPL libraries themselves, instead directing users to download them elsewhere. Does my code have to be licensed under the GPL? I believe the intention of the GPL is that it would, but does that have a legal basis?
If I can be modded down for being a troll, can I be modded up for being an orc, or a balrog?
better story: OSI attempts earth-shaking announcement about GPL to draw attention away from fact that their status as nonprofit in California is suspended. Perhaps it was due to failure to file required tax documents (for California and U.S. IRS) for many years, that issue was discussed on Bruce Peren's now-defunct site Technocrat
http://kepler.sos.ca.gov/corpdata/ShowAllList?QueryCorpNumber=C2224685
I am not a lawyer, and this is not legal advice.
Let's say Project Alpha is GPL2 licensed.
If Project Beta is not a derivative work, then you don't need to bother to read GPL2 and worry about it's wording, so there's no problem here (for the Project Beta people -- let's get back to this in a little while).
If Project Beta is a derived work of GPL2-licensed Project Alpha, and you don't want to infringe copyright, then you need to follow the terms of GPL2. Now this is where it gets interesting: under the GPL2's definition of derived works -- not the same as copyright's -- Project Beta still might go either way. It might be a derived work under copyright law, but not a derived work under GPL2's defintions, which means that GPL2's terms that talk about derivative works, might not apply. So you might have something that is legally a derivative work, but GPL2 might not give you any distribution rights. Or it might give you distribution (GPL2 defininition) rights, but then legally you might find that you still don't have distribution (copyright definition) rights.
This is exactly backwards.
The GPL2 does not offer a separate definition of "derivative work". It instead defers to copyright law itself, saying :
The objection of the article is that by defering to the legal definition of "derivative work", you are creating a different set of responsibilities based on what that term means in various jurisdictions.
That's true, but as you point out the local legal meaning of copyright law is _always_ going to affect whether copyright applies to a particular work or not.
What GPL V3 does is avoid using the term "derivative work", using distinct terms from copyright law.
While that does offer more clarity as to what those terms mean within the license, it falls into the trap that you outlined above: with GPLV3-style licenses, you have 2 different sets of standards to determine (first, is this a "derivative work", needed for GPLV3 to apply at all; second, if it is, is it whatever the GPLV3-specific terminology intends to cover).
The ultimate impact is that you _still_ have to know what a derivative work is locally, and on top of that you also have to figure out the license-specific language.
What has gone wrong, is that GPL2 is using copyright-related terms
I'd say that what was right was that the GPLv2 used legal terms rather than trying to create yet another parallel structure. That way, with GPLv2 there's only one set of definitions to pin down, and those definitions are ones that have been pretty well adjudicated in the past so you have a fair amount of history to look at to figure out how they apply.
This is, incidentally, one of the reasons that I don't put language in my code allowing distribution under GPLv2 "or any later version".
It's not that I think that the FSF is going to go crazy and put out a GPLv3 saying "the FSF has sole rights to this code" or anything so sinister--it's that they have different motivations from me, and I don't want to put others in control of the license for my code (a decision I'm happy with, considering that the first "later version" since I made that decision seems to be a significant step in the wrong direction to me).
rage, rage against the dying of the light
The one issue I've always disagreed with the GPL folks on, for instance, is code that links to GPL libraries. I know the LGPL was created to allow code to link to libraries without being GPL itself, but what if I distribute, say, a dynamic linking executable which links to GPL code, but I don't distribute the GPL libraries themselves, instead directing users to download them elsewhere. Does my code have to be licensed under the GPL? I believe the intention of the GPL is that it would, but does that have a legal basis?
I think it's something along the lines of "probably not, but the litigation costs to verify this probably wouldn't be worth it".
A lot of the supposed ambiguity can be eliminated if you first look at the code as not being computer programs, but just being another written work, then evaluate any potential "derivative works" from that perspective, and let existing copyright law and case law guide the evaluation. Once you've done that, then look at the license-specific terms of the GPL regarding what it says constitutes distribution, as well as what it says triggers the requirement for distribution, and I think you'll find it's not nearly as convoluted as these two jokers are making it out to be.
This is very petty lawyer-ing and typical misunderstandings from software EULA lawyers chasing their own tails for so many years.
GPL covers SOURCE CODE, and thru "derivative works" covers binary "performances". The whole reason we even have EULA's (End User LICENSE Agreements) is that there was one case 30 years ago where somebody argued that typing source code from a book to RAM and from the RAM to CPU was "infringement" and duplication of the work. So because of the internal machine copies needed, you have to be granted a special LICENSE to USE any kind of software (source code or binary). EULA writers have used US law's reliance on "contracts" to throw the "kitchen sink" in EULAs and call them "contracts" rather than license for use.
yes, the terms they point out have been more precisely defined since 1991. Judges respect stability and don't fall for dizzying logic like this. Judges will realize terms change and favor the UNMODIFIED document nearly every time as a matter of good faith. GPL v2 has been in heavy use unmodified for 18 years, that's incredible stability in an industry where other EULA writers reserve the right to edit/change/modify their EULAs online, without notice, and you pre-agree to the new terms you haven't even seen yet. The GPL is a legal rock, if the best they can do is mince words there's no threat at all.
It might be a derived work under copyright law, but not a derived work under GPL2's defintions, which means that GPL2's terms that talk about derivative works, might not apply. So you might have something that is legally a derivative work, but GPL2 might not give you any distribution rights. Or it might give you distribution (GPL2 defininition) rights, but then legally you might find that you still don't have distribution (copyright definition) rights.
You are arguing that if you base your work on some commercial engine, and pay a fee, you might get a lawsuit later because you paid even if you didn't have to pay, perhaps because you qualified for a free license. I do wonder if that makes onto my top 10 of "most-stupid-arguments-ever" ;)
Religion is regarded by the common people as true, by the wise as false, and by rulers as useful.
I think you'll find that content of the license would be quite relevant should the original author want to take legal action against somebody who he believes is violating the GPL. If the terms of the GPL applied automagically, there'd be no reason to write the GPL in the first place.
The GPL is an agreement between two parties and the fact that it invokes copyright law or wasn't negotiated doesn't change that.
"The intent of the author of the GPL is irrelevant, but not for the reason you say. The intent of the author is irrelevant because the grant of rights is from the copyright holder to the recipient. The intent of the author in releasing his work under the GPL could be relevant."
The intent of the author of the original code is also irrelevant for the reasons I mentioned before.
This is not quite like mathematics. A legal document may hold for a long time until someone finds a loophole. At the stage that they do find a loophole then very quickly more and more people will start taking advantage of it. Think of tax breaks, where you save tax by having a company or something. To begin with there is no problem until someone realises that you can cheat by registering a company even if you aren't really working like a company. Soon everybody is doing this and the law is no longer working the way it was intended. This is more like crypto. When you start hearing of weaknesses (as with MD5) and someone has already developed a stronger algorithm (here I won't guess which is the right one; RIPEMD? SHA256?) then that is the time to start changing and developing better alternatives (SHA-3). In the case of the GPL-v2 the weaknesses are already showing; hosted/cloud/web applications provide mechanisms for providing people software access without triggering the protections of the GPL-v3. The weak patent protections included are being worked around by Microsoft. "attacks always get better, they never get worse". The time has come to walk, not run away from the GPL-v2. Each of those problems is already addressed in the GPL-v3.
=~ s,(.*),<sarcasm>$1</sarcasm>,g if any_point_you_wish();
Yes, because the OSI and the Linux Foundation are going to hire lawyers to be their counsel that are actually enemies of their cause. Yeah, that makes a ton of sense.
"You can dual-license something all you want."
Dual licenses don't work for open source ... without an assignment of rights.
Specifically, if I license something under the "GPL or the Artistic License", and someone takes it under the terms of the GPL, makes modifications to it, and donates those changes back to me, those changes are a derivative of a GPL licensed work, and therefore must be under the GPL. Only if in their donation back to me there is an assignment of rights to me, am I free to relicense the changes out under the Artistic License.
This is what bit SGI, when they thought they were gong to get a boatload of developers jumping on their attempt to step out in front of the open source parade with a baton and lead the parade; almost none of the changes that were made by outside developers were usable to them, due to them lacking an assignment of rights.
-- Terry
Trust me on this one. There are wheels within wheels.
"Speaking the Truth in times of universal deceit is a revolutionary act." -- George Orwell
OSI was founded by people who were unhappy with the Free Software Foundation and the GPL. OSI Founder Eric S. Raymond recently said that the GPL is no longer needed [osnews.com].
There's a very good reason few people listen to that fruitcake anymore.
What has gone wrong, is that GPL2 is using copyright-related terms.
Woah, this is hilarious. For years GPL advocates claimed that the strength of the licence was that it was based on "Copyleft", that is the normal enforcement of copyright laws. Now, all of a sudden, it dawns on people that this somehow a huge problem! I never thought I'd see the day that free software zealots were FUDding Copyleft.
Oh, there is no "GPL2 definition of derived works", except in the minds of people who wanted to extend the licence beyond what it plainly says.
Business. Numbers. Money. People. Computer World.
It's like a coder talking about a bug in a C program. You can say to yourself "Oh, that coder's just causing trouble. There's no real bug" - but that doesn't change the reality that the compiled program will do just what the C instructions tell it to do - nothing more, nothing less.
Programs don't run by coders arguing about them.
Well, it's not a perfect analogy. Note there were no cars involved, either...
Legal documents mean whatever a the last lawyer convincingly pretended they mean as long as that lawyer didn't get get caught by another lawyer.
I'm not sure what you mean here...
Each and every time a legal document is read by a new judge, the interpretation may be different. Every time a hostile lawyer reads the document, he will be looking for new ways to exploit flaws in the document.
To go back to the programming language analogy: it's as if lawyers are programmers who code in an ambiguous language, which runs on hardware which tries to find ways to subvert the intent of the program without contradicting the literal interpretation of its instructions - or find excuses to reject those instructions altogether.
My point with that sentence was that a legal document, once written, may contain flaws. These flaws, if they exist, are there whether or not we choose to acknowledge them. The whole point of the programmer analogy there was just to illustrate the point that flaws are possible, and that it's not productive to ignore them... Sooner or later, the flaw will be exploited.
Bow-ties are cool.
Read the article, the web conference is hosted by the "license-sniffing", Microsoft partner Black Duck software.
stacked panels are an ongoing tactic of M$.
Anyway, the GPL has already been proven more than a few times in court on both sides of the Atlantic.
First U.S. GPL lawsuit heads for quick settlement
A GPL compliance case against Iliad
Beta is broken and the link to classic doesn't work. Stop wasting our time or there won't be anybody left here.
Here is the problem. All the IT people here are looking like this as a XOR problem. Either it is good or it is not. The law really does not work that way. What has likely protected GPL from real challenges more than what letter of the law says, is simply most everyone that used it was more or less judgment proof. It was only recent years that big money started getting in to the picture. Very few lawyers that are worth a dam would bring a law suit against the one man band in his basement coding in his spare time, and few of those one man bands would have the resources to take on say the likes of IBM or whatever over not following the licenses.
Now however we are getting in to problems. The money is there.
Still, on a real practical level the purpose and goal of the GPL has been met. Basically, to either force or encourage people to share their work. Yea, there might be a few real court challenges, but will everyone just one day come along and claim that all GPL built can somehow be claimed to be closed? The very nature of the way it came about, insures that gene is not going back in the bottle for 99.9% of all works covered by it.
Living in Chile
Sorry, I don't trust you. I see no reason why either one of those organizations would willingly hire someone on as their general counsel when the person is working against their own cause.
One thing that's often confused me is the exact relationship between the GPL and interpreted languages. For example, if I write a perl script which calls perl functionality which is part of the base interpreter, my script need not be distributed under the terms of the GPL. This is akin to using a GPL word processor or other software, where the output of a program is not subject to the GPL.
If, on the other hand, my script calls a perl function which is itself written in perl (licensed under the GPL), the FSF argues that this constitutes a derivative work akin to dynamic linking. Thus, my script (if distributed) must be distributed under a GPL-compatible license.
I can see it both ways. On the one hand, calling a function written in the same interpreted language is very much like calling a function in a library from a compiled binary. On the other, it's strange to think that there's a distinction based on whether the function being called is written as part of the interpreter (in, for example, C) versus the interpreted language itself. In addition, there seems to be disagreement about whether the GPL really binds like the FSF claims. Lots of interpreted code gets released as the GPL when it seems likely that the LGPL is what the authors really intend; that is, they do not want to restrict scripts and functions which call the code.
A good example of this is R This statistical language has fairly small interpreter and a large set of both included and downloadable packages, themselves written in R (and licensed under the GPL). Clearly most of the primary authors do not intend for all R scripts using the most basic of functionality to be released in a GPL compatible way; for one, they make the header files necessary for writing C-based libraries for use in R LGPL to explicitly allow such libraries to be non-free. In addition, they are fine with a large number of downloadable packages which restrict commercial use (obviously not allowed under the GPL). Their interpretation of the GPL seems at odds with the FSF. Even if you want to release all your code in a GPL-compatible way, it may be (IANAL) that you cannot call both code restricted from commercial use and GPL-licensed code (basically unavoidable) in the same project.
"The universe seems neither benign nor hostile, merely indifferent." --Carl Sagan
Posting anonymously for legal reasons...
The company I work for released a product recently that is not much more than a fancy GUI that calls (via OS calls) various bits of software under the GPL and then manipulates the output a bit further. After careful explanation to the legal department about how it works, they looked over the GPL and decided that our program is NOT a derivative work. Our major competition argued against that saying that even though we never included any GPL code in our code, and instead just call the programs "externally", it IS still a derivative work by function, and therefore we must open our code.
This case is still in the courts... (note: not in the US)
I think most developers would not consider it a derivative work, however would be hesitant to do this just because it feels a bit "against the spirit of the thing". However from a legal standpoint, both sides can (and are being) argued pretty strongly. What the developers think is entirely irrelevant here.
...but what if I distribute, say, a dynamic linking executable which links to GPL code, but I don't distribute the GPL libraries themselves ...
This argument has been around for years, and you're probably right. However, it seems to entirely academic -- nobody's ever done it (except to make a point), and there aren't that many GPL DLLs out there in the first place.
The only time I can think this "problem" has ever affected anyone in the real world was with MySQL database drivers.
Business. Numbers. Money. People. Computer World.
for example.
Installation (in the US, this is explicitly allowed without license, EULA tries to say you only have that right when you agree: you do not. In the UK until Jan 2009 it was strict civil tort so the damages that could be awarded by installing a program would be Nil).
Criticising.
Moving to a new computer (see installation above).
Decompiling.
Reverse engineering.
Benchmarks.
Arbitration clauses.
Need more?
Blame your copyright law.
It defines for source code what a derived work means.
This is why you have to agree to an EULA to install Microsoft DirectX (shared libraries) or the SDK to make a game linking to Direct X (shared libraries).
If linking shared libraries were not derived works, MS's EULAs for Direct X et al would be worthless.
I think you'll find them fighting you tooth and nail over that one, but go ahead, if you think you're hard enough.
Except the derived works "loophole" in TFA wasn't just discovered, people have been arguing about this for decades.
This seems to be about changing the messaging point in the FOSS community - the new "party line" is apparently to stop handwaving around GPL2 issues and instead accept them and push people to GPL3.
Business. Numbers. Money. People. Computer World.
That argument makes no sense. Properly written software should be easily portable to *any* device unless it's a driver, performance considerations notwithstanding.
Check out my sci-fi/humor trilogy at PatriotsBooks.
Except the derived works "loophole" in TFA wasn't just discovered, people have been arguing about this for decades.
In fact people have even been suing about these issues, so there's already case law. However the very fact that people argue about something is bad. If something can be made clearer then that's better.
This seems to be about changing the messaging point in the FOSS community - the new "party line" is apparently to stop handwaving around GPL2 issues and instead accept them and push people to GPL3.
Correct. That's why it's called the GPLv3 rather than having a completely new name. Just as you aren't recommended to install Linux 1.2 any more but rather 2.6.x. If you want you can still use the old version if you want to for some special reason. What's wrong with that? You make it sound like some insidious secret plot rather than a clearly documented upgrade strategy.
=~ s,(.*),<sarcasm>$1</sarcasm>,g if any_point_you_wish();
OSI was founded by people who were unhappy with the Free Software Foundation and the GPL. OSI Founder Eric S. Raymond recently said that the GPL is no longer needed [osnews.com].
There's a very good reason few people listen to that fruitcake anymore.
Just one?
Well played, sir.
Sorry if I wasn't clear, but the 'insidious secret plot' was to handwave around GPLv2 and derived works (see anything GNU has ever said about "linking").
If OSI's lawyers just discovered this issue, which people argued about on Usenet back in the 80s, they're not doing a very good job.
Business. Numbers. Money. People. Computer World.
proprietary pussy.
That's the one case where proprietary is preferable to free/open source.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
In all the cases I've seen, the alleged violator was doing something that would violate copyright if not done under license. No one has ever seriously questioned, as far as I am aware, that the license would not hold up in those kinds of cases.
However, the problem with the GPLv2, as FSF reads it, is that they think it covers some things that are not prohibited by copyright. No one has ever went to court over those, and if they did, the FSF would lose.
The FSF said 'FFS'.
If most programs went to GPL 3, you would see a wholesale abandoning of OSS and a return to Microsoft and closed solutions.
Simple reason: Companies have trade secrets. All it takes is one GPL 3 product in a cluster, and suddenly every single manufacturing thing they have, has to be disclosed to any customers who ask for it by law.
GPL 2 isn't perfect, but Linux distro makers know that Microsoft would eat their lunch the second that any production OS was GPL 3.
Sorry if I wasn't clear, but the 'insidious secret plot' was to handwave around GPLv2 and derived works (see anything GNU has ever said about "linking").
Sort of fair enough. Though I think it was more down to insularity (these things are more clear in US law and they have turned out to be right during any actual court cases) and the understanding that since the GPL was a US agreement it would be interpreted in the context of US definitions. Remember lots of what was argued about on Usenet wasn't always stuff that mattered in real life (at least judging from alt.flame :-) )
If OSI's lawyers just discovered this issue, which people argued about on Usenet back in the 80s, they're not doing a very good job.
More likely they now have a better answer than they had before, so it's a good time for them to bring up an issue they were less comfortable with previously.
=~ s,(.*),<sarcasm>$1</sarcasm>,g if any_point_you_wish();
GPL covers SOURCE CODE, and thru "derivative works" covers binary "performances". The whole reason we even have EULA's (End User LICENSE Agreements) is that there was one case 30 years ago where somebody argued that typing source code from a book to RAM and from the RAM to CPU was "infringement" and duplication of the work. So because of the internal machine copies needed, you have to be granted a special LICENSE to USE any kind of software (source code or binary). EULA writers have used US law's reliance on "contracts" to throw the "kitchen sink" in EULAs and call them "contracts" rather than license for use.
Except that under Copyright Law, you don't need any special license to USE software. See Section 117, which was amended in 1980:
Source: http://www.copyright.gov/title17/92chap1.html#117
I believe that the Copyright Act preempts your statements. :)
It was a really good paper.
My point with that sentence was that a legal document, once written, may contain flaws. These flaws, if they exist, are there whether or not we choose to acknowledge them. The whole point of the programmer analogy there was just to illustrate the point that flaws are possible, and that it's not productive to ignore them... Sooner or later, the flaw will be exploited.
And if it does not contain flaws, flaws will be claimed to exist anyway.
When there is a conflict, intended purpose and understanding of the document is supposed to trump language, and the original purpose of legal language is to make it easier to determine the intent and support a reasonable interpretation of such intent by all parties involved, not to plug all loopholes that may be imagined by creative interpretation.
Contrary to the popular belief, there indeed is no God.
The GPL has been held to be valid a number of times in a court of law in different countries. This is true. However, the lawyers didn't say it was invalid, so it is also irrelevant.
They said that it was unclear what is meant by derived work, and therefore it was unclear what could be licensed with a different license when combined with GPLv2 software, and what had to have a GPLv2 license. It all hinges, according to them, on if one takes an expansive view or a narrow view reading of the independent work clause. This is something that's very much up in the air right now, with many people playing fast and loose with the rules. You have a continuum of behavior here. Everything from "I wrote this file, therefore I don't have to license it under the GPL, even though it is linked into the kernel" to "I GPL'd the shims to my proprietary driver, but not the driver itself." The authors point out it is unclear how much of this behavior is safe and how much isn't. The ambiguity and shifting attitudes about what is and is not a derived work creates risk and uncertainty when using this license.
They claim GPLv3 doesn't suffer from these weaknesses.
Nowhere to do they claim the GPLv2 is not legally valid. Just that ambiguity exists,
Bruce is quite right. License incompatibility can be a real problem; by sticking to a small set of certain widely-used licenses, you avoid the problem. If you're interested in license compatibility issues, you might look at The Free-Libre / Open Source Software (FLOSS) License Slide and Make Your Open Source Software GPL-Compatible. Or Else
- David A. Wheeler (see my Secure Programming HOWTO)
Good article
Under a definition like "calls open source code", anything that runs under Linux would then be a derivative work. Because Linux operating system functions (like allocating memory for the executable, obtaining a socket, getting output to the display, reading keystrokes, creating threads, etc.) have to be called in order for anything to function. Even scripting. A Python script must run under Python, Python runs under Linux, etc.
Personally, I think the OS community is reaping the "benefits" of having ever tried to use the legal system to "protect" what they do. The legal system is a complete failure at serving the people. It may have (at one time, long, long ago) been designed to do so, but that is no longer its function. Now, it serves only to make money and consolidate power for itself and those who patronize it (by which I mean inject large sums of money into it.)
I've fallen off your lawn, and I can't get up.
Hey Bruce,
sorry for hijacking this thread, but could you give your opinion on the Vernor vs AutoDesk case, and the "threat" this is supposed to have on Free Software, as PJ on groklaw and more recently Nimmer have been writing. Do you believe that software being sold not licensed poses a significant threat to FOSS?
Cheers
Cyco
Having written a multitasking 'OS' for a microcontroller ...
I can say openly that real men aren't so stupid as to start from scratch in this day and age. You need some sort of framework to support your own framework until its far enough along that you no longer depend on it.
Yes, you CAN start from scratch, but WHY WOULD YOU?
Persistent Volume manager for Kubernetes - https://github.com/dwimsey/openshift-pvmanager
"To go back to the programming language analogy: it's as if lawyers are programmers who code in an ambiguous language, which runs on hardware which tries to find ways to subvert the intent of the program without contradicting the literal interpretation of its instructions - or find excuses to reject those instructions altogether."
So, it's pretty much exactly like writing Javascript and CSS?
You are not a brain: http://books.google.com/books?id=2oV61CeDx-YC
A more limited interpretation is that ownership of a copy embodies a limited right to use, period. Which doesn't bother us.
Bruce Perens.
There's a very good reason few people listen to that fruitcake anymore.
He's an armed fruitcake though, he can force you to listen.
May contain traces of nut.
Made from the freshest electrons.
Even if as a result of that ownership model, the AGPL's required redistribution clauses become toothless?
Thanks for your reply
Cheers
C
I don't think it will happen. If it does, we'll have to modify the license language.
Bruce Perens.
"To go back to the programming language analogy: it's as if lawyers are programmers who code in an ambiguous language, which runs on hardware which tries to find ways to subvert the intent of the program without contradicting the literal interpretation of its instructions - or find excuses to reject those instructions altogether."
So, it's pretty much exactly like writing Javascript and CSS?
Well, almost. Internet Exploder will subvert the intent of the program even if that means contradicting the literal interpretation of the instructions. :)
Bow-ties are cool.
My point with that sentence was that a legal document, once written, may contain flaws. These flaws, if they exist, are there whether or not we choose to acknowledge them. The whole point of the programmer analogy there was just to illustrate the point that flaws are possible, and that it's not productive to ignore them... Sooner or later, the flaw will be exploited.
And if it does not contain flaws, flaws will be claimed to exist anyway.
When there is a conflict, intended purpose and understanding of the document is supposed to trump language, and the original purpose of legal language is to make it easier to determine the intent and support a reasonable interpretation of such intent by all parties involved, not to plug all loopholes that may be imagined by creative interpretation.
Well, except that the document in question (the GPL) is a license. Recipients of the license can't be expected to follow the intent of the license if it somehow contradicts the document's wording...
Bow-ties are cool.
Ok. Clearly you did not attend the webinar (I did). The IP lawyers did not try to disprove anything. They were discussing the intricacies of the GPL v2. They discussed how it used copyright type language (and how the GPL v3 did not); how copyrights differed from country to country despite the Bern convention; what may or may not be considered a derivative work and so on. They did not advocate for or against anything. You refer to it as a "useless article." What article are you talking about? They presented a webinar with some PowerPoint slides.
You'd do better at pretending to be stupid if it was really an act.
I'll explain your mistakes in full, but I'll do them to someone who isn't such an ass about it. You can die ignorant for all I care.
I do not write, or interpret, the Copyright law. The Parliament, and the courts, respectively, do. The Parliament or whatever you have in your country.
Copyright law disallows certain acts *unless I give permission*.
If an act is not regulated by Copyright law, then the question of my consent does not arise. If an act by its nature is regulated by Copyright law, but the object of the act is a work that is not so related to any work in which I hold rights, as to to require my permission according Copyright law, then the question does not arise. I case of doubt, any court should first determine if my permission is required, then ask if I have given such permission. Only then my answer is the GPL.
The GPL must then be interpreted in light of the question to which it is answering.
The GPL v. 2 contains this passage:
This may seem like a contradiction: Does the license permit running the Program? But the whole document is clearly written in non-legalese language, with a long preamble, and many passages that clearly tries to explain to non-lawyers how copyright law works. What here is expressed is that the law, as the author of the license knew it at the time, did not require permission for the running of the program. This is of course unfortunate, as there are different laws in different countries. However, the GPL also says
If any act requires permission from a copyright holder, and the text of the GPL forgets to give that permission, this is not a disaster. The copyright holders can always issue additional permissions later. In any case, permission is not given in the GPL in such cases.
I am not denying the virtues of clear unambiguous licenses. But courts routinely interpret contracts and agreements written by laypersons. Courts are supposed to estimate the intentions of the parties, and whether they have made reasonable efforts to communicate their intentions to the other party, and made reasonable efforts to understand the other party's intentions. No party shall be bound by conditions that he reasonably remained unaware of when he entered into the contract.
Underlying this is the understanding that an agreement is an attempt by two or more parties to establish a cooperation to mutual benefit. While standards may vary in different regions as to the extent of the obligation of the parties to actually read the small print -- that is, receive the communications from the other party about his intentions, or conversely the extent of the obligation to communicate your conditions with clarity, in no region is the entering into an agreement seen as a challenge to the other party to "game me if you can".
In the case of the GPL, the intention of the person who chooses the GPL for his work transpires with clarity from the whole document. The long preamble helps setting the context. It transpires that the copyright holder wish to initiate a commons, to which others may contribute, or to contribute to an existing one. In no statement of the GPL is it expressed any consent, supposing such consent is required, to contribute in any way to anything that is not itself in such a commons.
There is no substitute for common sense. Especially, no body of rules will do.