|
|||
FUD
Quote:
I do agree that the idea that one should be seeking legal counsel in order to apply a license is outlandish and unworkable for hobbyist programmers. The GPL has always invoked those general feelings for me - those feelings of uneasiness over how I should interpret the license. It's sooooo much easier to pick the BSD license after you compare it to the verboseness of the GPL. I never thought the LGPL was very workable. Clients never have the right stuff, and they don't know how to get it, even with help. I think many large app producers who go that route really do not abide by the license, but they just say that they do. Stallman knows this, and it aggravates him. Admittedly, driver producers can benefit from the LGPL, as they can make calls into the kernel and (for the userland counterparts) - the C library, and those things don't change as often. Maybe that's an area where the GPL v3 was seen to be rubbing the wrong way. I suppose I'd have to consult a lawyer to know for sure. What it amounts to is that if you're going to use GPL'ed code in some way inside or connected to your own code, then release all of your own code in harmony with the GPL. This makes it a lot easier to deal with, even if there are "loopholes" you could otherwise use, and your conscience will be clear. If you use BSD or MIT code, and use BSD or MIT licenses, then you can keep (all or part of) your code secret, or release all of it. So, we really are talking about two separate camps. Pick a camp. That's what FreeBSD and Apple have done. Separate camps aside, IMO there is much FUD going about town, around compiler issues and other things. That FUD turns this whole thing into a religious war. Really, I think the two camps can get along. No reason for war. Last edited by censored; 22nd May 2014 at 03:29 PM. |
|
|||
On related note
On a related note, we should consider the attack against the idea of the API as being outside of the scope of copyright. Google just lost against Oracle on that basis. If APIs are now suddenly, in spite of thirty year old precedents, copyrightable, then the GPL automatically degenerates to the LGPL by the opinion I've seen expressed in an article I've recently read. Not that it matters. Again, I'm not a lawyer, so disregard all.
I'm surprised there hasn't been more trade press about the API brouhaha. It's very serious stuff, very negative for all of us, no matter the camp we're in... Edit: Attempts to relocate the "GPL automatically degenerates to the LGPL" article failed, and so the statement is hereby retracted. It was just my imagination, apparently. Relative to my previous post regarding GPLv3/v2: The kernel exception in the Linux kernel lets us use the kernel more liberally than if the exception wasn't there. But, the idea of applying exceptions to a context otherwise ruled by GPL licensing is probably more complicated than it seems, and IMO brings more FUD into the picture. It's Just easier to pick a license that fits the purpose, use it harmoniously, and forget the loopholes. I suppose that's what Apple and FreeBSD have done with the complete ejection of GNU. No more loopholes. Each to their own. Last edited by censored; 23rd May 2014 at 10:12 PM. |
|
|||
If you're interested in that case, Bradley Kuhn and Karen Sandler discussed it on their recent podcast episode here: http://faif.us/cast/2014/may/13/0x44/
This being a complicated case (and if you'll excuse me saying so your post showing great confusion or indication of being influenced by severe misinformation, e.g. the copyrightable arguments by Alsop were largely rejected but fair use considerations are remanded back to the lower court, so "loss" hasn't happened yet), if you have the time I'd recommend also listening to their earlier podcast 0x35 and especially reading over Judge Alsop's decision followed by the recent circuit court decision. The pdfs of those two decisions are linked from episode 0x44. Bradley Kuhn also has a blog entry on the recent decision and there is discussion with him and others linked from there on identa.ca: http://www.ebb.org/bkuhn/blog/2014/0...le-google.html From what I hear most technical press coverage of the recent decision has been complete crap. I'd not read any more of that and try to forget what you've already read. |
|
|||
Quote:
Edit: Now I'm back from the reading of the text associated with your links. In it I see that the "loss" refers only to Google's loss of the original, favorable district court decision, because of Oracle's appeal. That's what I meant in my post (bad wording), as I did absorb that info from the news about the case. Now the case gets kicked back to the courtroom for a determination of whether or not, in the specific referenced case, the alleged infringing copy could be claimed a "fair use" copying of the code, and by direct association - (only in some people's opinions) - a green light for the use of API declarations as "fair use". That favorable result has yet to be had. It would not be a green light, and barely a yellow one. If the new courtroom result is that the allegedly infringing usage was not "fair use" - then there is a precident for APIs being generally copyrightable, with fair use disallowed in particular cases. Subsequently, all allegedly infringing code would rely on courtroom actions, decided on a case by case basis. That is the problem with the utilization of the "fair use" defense. "Fair use" is vague to the point where it almost always has to be hashed out in a courtroom. Thus, it does not protect the general immunity (relative to copyright) of APIs. It does make more money for the legal profession, because nearly every case has to go to trial. Talk about FUD! If one considers that "fair use" is often limited to a small number of lines, depending upon the form of expression, that does not bode well for Google. There are some 7,000 lines of declarative code (the API) and fair use has in some cases been limited to a hundred. It depends upon the percentage - so with such a large API the courtroom result could allow for a larger number of lines of code protected under fair use. Sometimes, the result in such cases has been a substantial percentage - say ten percent or so. Unfortunately, due to all the vagueness associated with the doctrine, that figure goes all over the house. Relying on fair use to protect our ability to re-implement APIs will be nothing short of a nightmare. The only way it isn't a nightmare is if "fair use" for purposes of software interoperability is precisely defined (and that is light years away from current applications of the "fair use" doctrine in courtrooms). The ideas expressed on the sites you linked, about the court and specifically the jury not understanding the idea of "APIs" well enough to separate them from straight forward copying of code in an ordinary sense, does not help necessarily. Even if that is an error, will any subsequent actions see it as one? A precedent can be sited, error or not, and subsequent juries (and courts) could compound the issue with the same errors and lack of understanding. The whole thing's a complete nightmare. Again - I'm not a lawyer, so disregard everything I write. Last edited by censored; 23rd May 2014 at 06:16 PM. |
|
|||
Quote:
Quote:
Quote:
http://www.cafc.uscourts.gov/images/...5-7-2014.1.PDF In the first quote, Oracle seems to be saying that the merger doctrine (the basis for what programmers usually think of as "API freedom" - something needed for interoperability, and not subject to copyright) does not necessarily exempt the names of elements in API declarations, and/or the declarative line as a whole, from copyright. Statements in the court's opinion refine this to the names of methods and classes. If this holds, then there is no longer any API freedom in the manner that programmers have accepted it for thirty years. The second quote expresses this idea (that really kills the effective sharing of APIs) - put in the words of the appellate court that sent the case back to trial for "fair use". It means the court interprets the idea of what the API does as exempt from copyright, but not the names of the specific methods and classes. Without the ability to use the specific names and classes of the API, there is no interoperability. To that - the court says that interoperability (a doctrine that exempts some copyright) is a fair use issue. Nightmare! What constitutes "fair use" is vague and almost always requires court time to decipher. A good law or doctrine minimizes the amount of time citizens have to spend in courtrooms. Could "fair use" then be more than exemption based upon a "small amount" that was copied and needed to have effective "public discourse" - and now extend to ideas such as "interoperability" of software? Perhaps, but in that case would usage be hamstrung by some vague limit as to how much you could create interoperability with? Like the fair use of news articles? This issue is huge. Oh yeah. I'm not a lawyer. Disregard my opinion. Last edited by censored; 23rd May 2014 at 06:41 PM. |
|
|||
Well, another thing they say in the show is to be careful about generalizing any result from this case to other cases. It doesn't necessarily set a strong precedent for later cases with slightly differing details (or different courts?) even if the fair use argument doesn't go google's way. I don't understand this, not knowing much about how law works, but this quote from Alsop, "development of the law reveals a trajectory in which enthusiasm for protection of "structure, sequence and organization" peaked in the 1980s," makes me think it wouldn't be an on/off sort of thing where Google's loss leads to Novell (if they even wanted to) suing everyone doing unix and wine and reactos have to retract their work. Precedent seems not to be such a solid predictable thing here. There's something about it being only one jurisdiction of many too and with the copyright act giving individual courts a lot of discretion to decide each case on its merits. Sort of like, "we legislators don't really get this software thing so you're going to have to figure it out over time, courts."
If they get a fair use argument, I'd think that should make people feel comfortable enough, especially given that API declaration copying hasn't even been considered worth trying to pursue in copyright cases before. As far as size, well nothing much is as big as Java, so if that's covered anything I actually like ought to be pretty safe. And then the circuit court seemed more to be interested on the interoperability thing whether prior java programs existed that might run on Android or not. Given what I skimmed I'd be surprised if arguments resembled fair use arguments as applied to quoting newspaper articles. The cases they considered were all heavily technological, except that accounting method one, but that seemed more for historical context. I certainly wouldn't thinks there's worry about a software case setting precedents for print. There's been lots of cases about print over the centuries. They don't need guidance from software cases to help there. Probably the judges would leave software cases on the shelf as much as possible given how much annoying technical detail, detail they clearly struggle to grasp, is involved. It would be funny to see the court weighing different APIs based on their size and how much "three dimensional structure with all the chutes and ladders" they have. So imagine Unix's API is not copyrightable cause it's a flat namespace (one dimensional) and relatively small, but java's is cause it's large and hierarchical. Or maybe they'd have a tighter standard and only Scheme R4RS would be fair game and everyone's screwed except Nils Holm (Scheme 9 from Empty Space). It seems stupid to me to have APIs copyrightable (at least if copyrights are going to be more than 10 years), but it might be kind of interesting (may you live in interesting times) if free software had to adapt to that. Like Karen says, they'd just have to use their own designs instead. Or maybe they could copy people less likely to sue. e.g. who has the copyright on smalltalk and common lisp? In a way it might be kind of cool to have Java and Unix banned for use by free software. Assuming people could find the motivation once we have nothing again, maybe we could get something a little different this time. But it's pretty early to worry about that. Let's see where this goes first and if anyone else decides to try to sue from a similar place. The most jarring idea in your first post was that APIs being copyrightable weakens the GPL. I can't even begin to see the logic there. If anything, as Bradley I think is getting at with his mixed feelings comment, it makes the GPL that much more enforceable. e.g. GNU could sue to have mg GPLed since it replicates the "structure, sequence, and organization" of emacs's command set despite its implementation being independent (not that they ever would do that). |
|
|||
Was just reading Judge Debevoise's rejection of USL injunction on Berkeley. Sounds to me like in the particular case of Unix, assuming APIs were somehow made copyrightable in general from precedent after some cases like Oracle v. Google, BSD would probably be okay. He was suggesting USL wouldn't have luck because they didn't put copyright notices on lots of code and that mattered back in the pre-Berne convention days:
http://sco.tuxrocks.com/Docs/USL/Doc-92.html |
|
|||
Quote:
Quote:
Quote:
|
|
|||
|
|
|||
It's very simple. Without Richard Stallman, we'd never have gained the huge amount of inertia as it's been acquired by the open source realm, fueled by the Free Software initiative and its iconic sparkplug promoters. Open source would never have risen to the level it has, and the popularity of Linux may have been much less, even to the point of it's not being a political force.
Without the likes of Google Android, closed-system cell phones would be the dominant or only force in that arena. The "lock down" of the entire hardware ecosphere would likely have already happened by now, including the entire Microsoft hardware and software ecosystem, which would likely have followed much more closely in the politic of Apple. This could easily have made the open source domain an elite but small niche group of people who could build their own hardware. While some big companies (like Samsung) embrace open source, it's not by a large political/economic margin that they do. Without all the hard work of people in the FSF community, such companies would not see any benefit to follow an open source model, and would likely submit to locked down NDA type software contracts for their goods. I prefer the BSD/MIT license. It's less legalese for me to worry about. But - just because I live in that camp doesn't mean I don't appreciate Richard Stallman. |
|
|||
It's a battle that has not been won. Make no mistake - the major powers in this arena are in favor of not only proprietary software, but also of proprietary hardware. A completely closed, neat little locked down system, wherein they control all aspects of the computing environment - end to end - is what most see as the holy grail. They're working on it. Dedicatedly.
Do you know what software is running your cell phone? Probably not. You may think you know what's running on your cell phone. Regardless of the operating system your cellphone/smartphone purports to use, it is running (in most cases I'm aware of) on top of a hypervisor which simultaneously runs a second RTOS operating system for phone specific processing. You probably don't have the source code for either of those items (hypervisor or RTOS). Those items, running at the lowest level, or at a higher level with priviledge, can do *anything*. Such an idea is not comforting for me. In the old days, your PC was very transparent. You could run all of the software, and be assured that you could look at all of that software (given it was open sourced). The trend is to close this off. I think it's for the obvious reasons... Mobile devices are replacing PCs. Mobile is convenient, it's true, but it's disconcerting. We lose the kind of control we had with the old machinery when we switch to mobile. That loss is in concert with the trend to close off the workings of the hardware we use to execute "PC functionality" in the new era. We suffer a loss of control at a critical level, and simultaneously a loss of who controls "our" systems. Perhaps it's the latter loss that is the most distressful... Last edited by censored; 3rd June 2014 at 10:41 PM. |
|
||||
Actually, I have a pretty serious issue with all of this.
RMS was working on GNU for a while, around the same time the BSD camp was writing what essentially was an open source replacement for AT&T's Unix. The BSD camp didn't call it "open source", but for all intents and purposes it was open source. (GNU was announced 1983, BSD's first release was in 1978). Linus came along and went "Gee, it'd be nice to have a free version of Unix I could use at home! Oh look, BSD!" Then Linus saw the law suit taking place between AT&T and BSD, and thought "well, no sense in bothering with BSD, I don't know if it'll still be here in six months". Then he noticed GNU and thought "well, if only I had a *kernel* to go with this GNU operating system." And Linux was born. Now, did "GNU/Linux" lift off because of GNU, or because of Linus? It's *really* hard to say...but looking at GNU's kernel choices, I'd say Linus had a great deal to do with it...because it's *really* difficult to run applications on top of a kernel that doesn't work (see question #13 here: http://www.redditblog.com/2010/07/rms-ama.html). This can be attributed to a law suit by a massive US corporation who cared nothing of freedom (user *or* software). I'm willing to bet, though, that if Linus had put his skills to use in BSD, the open source landscape wouldn't be incredibly different from what it is today...Stallman or not.
__________________
Linux/Network-Security Engineer by Profession. OpenBSD user by choice. Last edited by rocket357; 4th June 2014 at 07:57 PM. |
|
||||
Quote:
For example: - Large parts of the Android userland are not open source. - Pretty much all Android code is made by Google. There are no public discussions on anything, really. - Almost all Android drivers are closed-source (they're usually not provided by Google but by the phone vendors). - Most Android phones are "locked", you need all sorts of stupid tricks ("hacks") to do stuff with them (like, install a different OS/ROM, unlock more features, etc).
__________________
UNIX was not designed to stop you from doing stupid things, because that would also stop you from doing clever things. |
|
|||
Quote:
Whoops, okay, looks like I'm off here if you look at Cyanogenmod perhaps being a more significant effort. I'm unclear what their motivations are or were. Still, don't surveys of free or open source software developers tend to show a significant portion who list as motivation or partial motivation ideological reasons that sound a lot like what rms and friends promote? I don't see very much of the crusade feel or phrasing it as a moral question from any other camp. Maybe the world didn't need that to get some large amount of open source software but it can't have hurt. He's certainly influenced my thinking, not that I'm contributing much of anything. |
|
||||
Quote:
For me, it's clearly #2 and #3 : replace the FSF/GNU/GPL staff by any BSD staff, and see. I don't think the motivations would change a lot, nor the result (except the licenses).
__________________
ThinkPad W500 P8700 6GB HD3650 - faultry ThinkStation P700 2x2620v3 32GB 1050ti 3xSSD 1xHDD |
|
|||
Quote:
|
|
|||
Quote:
However; we need to hang onto what we've got. So .... relative to my previous posts in this thread, what WAS Oracle REALLY after? |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Stallman: How Much Surveillance Can Democracy Withstand? | J65nko | News | 0 | 15th October 2013 12:27 AM |
Opinion for new SOHO Machine - Processor and Windows compatibility | IronForge | OpenBSD Installation and Upgrading | 4 | 1st January 2013 12:44 AM |
Stallman: Did I say Jobs was evil? I meant really evil. | J65nko | News | 1 | 30th October 2011 08:18 PM |