When Google first unveiled its WebM project, there were quite some concerns over its license. This license was incompatible with version 2 and 3 of the GPL, and was effectively a new license, causing unnecessary confusion. Google has now cleared everything up by switching to a regular BSD license.
The original license for the WebM project had a clause that if you brought patent action against Google, the patent license was terminated. This isn’t unusual in open source licenses (the second Apache license and version 3 of the GPL have something similar), but the WebM license had an additional problem.
“The twist was that ours terminated ‘any’ rights and not just rights to the patents, which made our license GPLv3 and GPLv2 incompatible,” explains Google’s Chris DiBona, “Also, in doing this, we effectively created a potentially new open source copyright license, something we are loath to do.”
They’ve resolved the issue by decoupling patents from copyright, meaning the copyright part is now a pure BSD license. They’ve used “patent language borrowed from both the Apache and GPLv3 patent clauses” for their own patent clause. They’re no longer creating a new license, and the patent clause can stand on its own. Additionally, they have updated the patent grant language to make it clear that it includes the right to modify the code and give it away to others.
“Thanks for your patience as we worked through this, and we hope you like, enjoy (and most importantly) use WebM and join with us in creating more freedom online,” DiBiona adds, “We had a lot of help on these changes, so thanks to our friends in open source and free software who traded many emails, often at odd hours, with us.”
A good solution from Google.
ok, it’s official, now I really do care about codec and where it’s heading.
Google, you did a very good thing here. on one hand you might just have done this so that no one can complain it’s not open enough (the _____BSD folks), on the other hand though you could have just done this to ensure it will now make it into everything ever to be made giving you an even larger presense in our daily lives that you control most of our data and what not…
But, i am not going to take that route and instead I will give you something I never have before, a sincere “Thank you.” (this is a one time thing, savor it).
(That’ll do pig, that’ll do…)
Edited 2010-06-04 22:52 UTC
None of the BSD folks opposed the license, AFAIK, it was the GPL folks.
No, it was the Open Source Initiative (OSI) folks who opposed it. The Free Software Foundation (FSF) publicly supported WebM.
Microsoft, Apple and Google along with the rest of the corps all have there moments. Credit must be given to each where due. Today, it’s Google’s turn and for this, they do deserve credit. Viva La youtube/html5/WebM!
Edited 2010-06-05 01:42 UTC
Except it’s hard to see how a video codec will intrude into your life …
They can afford it and wanted a solution, now.
A big “thank you” to Google for doing this.
Now that the license is all sorted, the focus can
move to further improving the already-very-good software.
It is hard not to love google.
Yes, when you have no brain.
I suggest the WTFPL license
http://en.wikipedia.org/wiki/WTFPL
it seems better…
Google should also pay everyone who want to use WebM.
WTFPL wouldn’t do what they want, though – they need to defend VP8 against patent trolls, and WTFPL fails miserably in the face of that.
Pay them for what,exactly?
Moron.
I think that comment was tongue in cheek. The only way Google could do anything more awesome with WebM is to release it under an even more relaxed license and pay everyone to use it.
The fact that that would be utterly ridiculous further enhances the point: Google spent $133m to acquire On2 and has given away the most advanced version of one of it’s primary technologies to the world.
So much of the focus is on HTML5 at the moment, but that’s just the tip of the iceberg. The big boys can now make millions selling WebM based HD software and potentially save money incorporating it into their existing products and never have to pay royalties, open source their products, or give anything away themselves. All they have to do is not sue Google over the patents included in WebM.
Microsoft (or Apple, Sony, etc.): “Hmm… should we take the money cow we don’t have to pay for, work for, or look bad for – no strings attached, or should we *try* to kill the free cow, spend millions, face a public backlash, potentially have valuable patents invalidated, etc… then we can spend millions of dollars in royalties for an ever so slightly better cow.”
… then again, knowing some of the players (*coughMicrosoftcoughApplecough*)… I wouldn’t really be surprised if they chose the latter.
Ya aint grown up until you understand the GPL licence.
Apparently Google understand the GPL just fine since they didn’t use it.
The fact that this comment reached a score of 14 looks pretty bad for the future of free software, considering that in the most frequent case (freewares), GPL’s major particularity is that the author asks for some kind of retribution for his/her work, contrary to BSD where he/she works for free…
But it’s very convenient to ignore this kind of facts and only look at the restrictive aspect of the thing. Just like it’s convenient to go in the bus without buying some kind of ticket because the money saved this way could have some better use…
Edited 2010-06-06 06:34 UTC
We would not have WebKit without LGPL — Apple would not have given any of KHTML code back the community unless they were forced to!
I don’t think that Linux would be so widespread with out GPL. How may of the commercial companies would have decided it was not in there best interests to share improvements they made to the kernel?
You mean like Apple does with LLVM and Clang?
Oh, no. Seems you’re wrong. LLVM and Clang are BSD-licensed and Apple still develops it fully in the open and does not create a closed source fork.
From the overall WebKit code base, only WebCore and JSCore are LGPLed. All the rest is BSD-licensed.
Apple owns the entire source code to CUPS. Apple is not forced to produce an open source version. Same with Darwin Streaming Server, Calendar Server, launchd, Bonjour, …
I disagree.
Webkits performance results in a superior user experience over other browsers. The last thing Apple would want is for their competitors to be apple to provide comparable performance, it is just something apple has to live with.
LLVM and CLANG began as university research projects, had they originated with apple they might note be open source.
Even if they had originated at apple, they do not provide any benefit that is visible to the end-user, so from the Apple perspective they don’t lose anything by sharing the source code.
Exactly. An open-source software which originated with Apple still remains to be seen. On the other hand, things like the App store…
Edited 2010-06-07 14:16 UTC
It’s just a matter of point of view: GPL ensures the code itself always stays free, while BSD license ensures the one using the code is always free (to do as (sh)he pleases). I personally prefer the latter one, it’s not real freedom if you’re restricted by some arcane clauses in the license.
And in this particular case BSD license makes a whole lot more sense: there is a real need for Google to push WebM around and GPL license just doesn’t appeal to many of the larger companies.
Maybe. However, the OP’s was implying that BSD licensing was somehow intrinsically better. Which I disagree with. It’s the author’s right to decide that he doesn’t want to give code to big closed-source companies for free. BSD-style licensing does not offer this protection, though it has the benefit of offering more freedom to the users of the product. So it’s not intrinsically better, it’s just different.
Edited 2010-06-06 10:17 UTC
Google wanted to encourage as many parties as possible to become involved in Webm in any and every way (except via suing Google).
For this purpose, the BSD license is indeed intrinsically better than the GPL license.
The GPL license is very good for ensuring that the code will always stay open, and that all ongoing downstream recipients continue to receive the same assured freedoms with the code.
That wasn’t Google’s purpose. Google wanted to allow companies to embed Webm into closed (hardware or software) products if they wanted to. BSD suits that aim much better than the GPL does.
Remember, it is the owner of the IP who determines what their purpose for creating/acquiring that IP actually is. It isn’t about a license popularity contest.
Edited 2010-06-06 11:30 UTC
You two agree. Seems like you misunderstood the way `intrinsically’ was used.
No, I implied that GPL was obviously not the right choice in this case.
Okay, sorry for the misunderstanding ^^ That’s the drawback of short posts like this, their advantage being that they are more often actually read than longer ones…
And in this particular case BSD license makes a whole lot more sense: there is a real need for Google to push WebM around and GPL license just doesn’t appeal to many of the larger companies.
No, you got it completely wrong. By going to a “pure” BSD license, Google was pretty much taken OSI out of the picture completely.
Remember, it was Raymond who started this whole mess with the license, by sticking his nose in where it wasn’t wanted or needed, just so he could promote himself and his worthless organization once again.
Except that if it was GPL, and you were willing to use just one line of code, you would have to make all your millions lines of code GPL too.
Google wants to spread the WebM use, no matter if it’s on IE, Safari, Opera, Firefox, etc. If a license blocks all this, there is no reason to have WebM at all.
Okay, so let me explain my point again.
1/GPL is better than BSD in some cases
2/BSD is hence not something intrinsically better. Though it is indeed better in the WebM case.
Anyway, the parent poster just clarified his point, so I think that we should stop the GPL vs BSD war. For “public domain” projects started by big companies and aiming for large industry adoption, BSD is better. For most free software projects, especially those from devs who work for free in their spare time, GPL is arguably better, though this can be the subject of endless discussion and your mileage may vary. End of the story.
Edited 2010-06-06 19:32 UTC
WTF?
If your project used 1 line of GPL code, and one million lines of your own code, then it would (in theory) at worst be considered a “derived work” (this is a legal term, Google it if you are curious).
Ownership of the derived work would (in theory) be 99.9999% yours and .0001% of the authors of the other line.
This is such a ludicrous split that it would be of no consequence, so the work would in effect be all yours. License it however you please.
Oh, and while you are at it, it would be appreciated if you stopped spreading BS about the GPL.
Edited 2010-06-06 23:42 UTC
“The GPL is the first and foremost copyleft license, which means that it requires derived works to be available under the same license terms.”
http://en.wikipedia.org/wiki/GNU_General_Public_License
GNU zealot that doesn’t even know GPL. Nice.
That is just one option for the derived work. The derived work can also be released under the GPL (i.e. released with all source code), and everything is sweet. If your “project” was actually 99% GPL code written by someone else, and only 1% your work, this might be the option that you choose.
There are however at least three other options that have nothing to do with the GPL.
(1) Dual licensed code. With this option, the author of the derived work would purchase a separate commercial licensed for the code which he/she had not written. MySQL would be a good example here, you can get a commercial license for that. Qt is another example.
http://en.wikipedia.org/wiki/Multi-licensing
(2) Re-write the GPL parts (this is what you would do for the silly case of 1 line of GPL code vs a million lines of your own code. Just re-write the GPL code line(s) in a different way that is not copy & paste, and now you have one million and one lines of your own code, and no lines of GPL code. The GPL gives you permission to study the code, find out how it works, but not copy and paste it and sell it as your own work. Re-writing a work-alike of the GPL’d code is doing your own work.
(3) Get permission form the original GPL authors. They might just give you permission to use their code in your project, or they may ask for a little money.
(4) Dynamic linking … this is OK for LGPL code (many libraries). With dynamic linking, the linked-in code is not considered to be part of your product, and so your work is not a derived work.
Lots of options. All you have to do is sort out the shared ownership of the derived works, and to give the appropriate consideration to the authors of the bits which you did not write. There are multiple ways to do this, all perfectly legal.
Your one million lines are under no threat of “being forced to become GPL”.
Anti-freedomer sprouting rubbish about the GPL. Typical.
Edited 2010-06-07 02:51 UTC
But the fact is… if you use one line of GPL code in your millions_lines_of_non-GPL code, all your software must become GPL, AS I ALREADY WROTE. So, I was right.
No. You would have a potentially “derived work” which is 99.9999% your own work.
In that particular (ludicrous) case, it would be effectively considered your own work.
So, how would a court decide? How can I say that?
http://en.wikipedia.org/wiki/Derivative_work
The one line of GPL’d code would not constiute a major element of your million-and-one line work.
You are good to go, without having to worrry about that one copied line of GPL code. This is basic copyright law, it is the law of the land.
So you were wrong. Please do not sprout rubbish concerning subjects about which you obviously know nothing.
Edited 2010-06-07 03:01 UTC
Some years ago, in 2002, I saw a talk where a Professor, working in a software, needed 20 lines of bash code in much bigger BSD-licensed project he was in.
So, he sent a e-mail to Richard Stallman, which said that if he (Professor) was willing to use, even a small fraction, of GPL code, all the software must turn to GPL.
Sorry, you’re right. Wikipedia is wrong. And since you’re a GPL specialist, I recommend you to go there and edit the page. I also recommend you to send a e-mail to Stallman to clarify these points, since he probably don’t know about GPL either.
Here are a couple of attempts at the definition of “derived work”:
http://en.wikipedia.org/wiki/Derivative_work
http://www.america.gov/st/econ-english/2008/April/20080429233718eai…
They agree that the original work must constitue a significant part of the new derived work. If the original work doesn’t constitue a significant part of the new work, then the new work is not “based on” the original at all, and is therefore not a derivative work.
This is the basic principle of copyright which allows new authors to have a go at writing their own versions of the “Robin Hood” story, for example. The new versions are all going to have the line: “Arise, Sir Robin” near the end of the story, and no-one is going to claim copyright violation over that.
This is basic copyright law. It has nothing whatsoever to do with what someone like Stallman may or may not think.
Detailed discussion here:
http://www.linuxjournal.com/article/6366
Even though the direct quote from the law is a little vague, it clearly says that the derivative work must be “based on” the original work.
Copying one line of GPL’d code into a million lines of your own work just simply doesn’t make your million line “based on” that one line.
Not in any sane defition. Not on any planet in the universe.
Another article, fyi:
http://www.chillingeffects.org/derivative/
Edited 2010-06-07 03:32 UTC
I don’t care what the definition says. But I suggest you, a GPL specialist, to send an e-mail to Richard Stallman. He probably doesn’t know this.
We were talking about GPL, and GPL works this way. Period. Get yourself informed.
Pfft. Utter rubbish.
Here is the way it “works”. This was the outcome of the first actual case in the US involving the GPL.
http://www.linux-watch.com/news/NS3761924232.html
It is all about a company making a slightly modified version of an entire open source program (in this case, BusyBox) and releasing that as closed source within its own product. That is a real-life, honest-to-goodness, actual-court-case GPL violation.
Copying just one line of GPL code into a million lines of your own code just isn’t. The law doesn’t say that at all (I even quoted the actual text of the law to you). If you tried to take such a case to court as an alleged violation of the GPL license, the court would simply say to you “go away you silly little boy”.
Perhaps just that is what I too should say to you.
Edited 2010-06-07 03:52 UTC
Thank you. Just proves what I already said.
How so? Are you slow, or what?
In this case, one of the very few ever taken to court, the company called Monsoon Multimedia had copied all of the code from the GPL BusyBox program, modified it very slightly (just a few lines), and re-sold it as closed source. The product that they sold had an almost-unmodified version of GPL BusyBox within it. Monsoon Multimedia did not write the vast majority of that program, the original BusyBox authors did. BusyBox GPL code constituded a major part of a program (a derived work) that was included in the Monsoon Multimedia product.
That is nothing like what you said. You claim was that “just one line” of GPL code would require “a million lines” of your code to become GPL. The wrong way around entirely. This case involved more like “a million lines of BusyBox GPL code” plus “just one new line written by Monsoon Multimedia”.
Monsoon Multimedia quickly decided that it was far easier for them to release their one additional line of source code as GPL code also, rather than re-write the million lines of BusyBox that wasn’t their code to sell in the first place. After all, the million lines of BusyBox code, that Monsoon Multimedia didn’t write, were already open source anyway. It isn’t like Monsoon Multimedia had to give away any of their own real work. All that was required of them was to release as open source (GPL) the very few lines of BusyBox that they had modified. That’s it.
Where is the downside for Monsoon Multimedia in simply complying with the GPL license of BusyBox code?
PS: Any other bits of the Monsoon Multimedia product are not affected by this case. It involved only the GPL BusyBox code as it was used within their product, the vast majority of which Monsoon Multimedia did not write in the first place.
Edited 2010-06-07 06:15 UTC
…. And this is why the good programmers dont use GPL. Good Programmers != Suicidal.
It would be a good reason to not use GPL if it were true, and if you wanted to create a closed-source software product.
The thing is, it isn’t true.
Edited 2010-06-07 03:05 UTC
I was thinking of replying something about KDE, GIMP, Firefox, Emesene, GCC, Blender… Then no. This post does not require an argumented response.
Keep trolling, please. You’re bringing great enrichment to this topic.
Nice trolling. Try developing for a living for once
No, sorry, I don’t think that this post qualifies as nice trolling. I’d call “nice trolling” something like : “If you develop for a living, you use closed-source licensing anyway, because you don’t want to give your work for free to competitors” =p
In my opinion, it’s fair to say that if some software is awesome, well-accepted as such, and manages to remain so for a long time, its developers must be good too. But I may be biased since I’m one of those hippie public research advocates who think that good/profitable is almost necessarily a trade-off, and that the need for profitability is the root of all evil in an industrialized society ^^
(OT : About developing for a living, I thought about it some times since I entered university. Choosing between computer science and physics is a hard choice when you like both. But each time, after spending some hours thinking, I always end up preferring physics as a job and computer science as a hobby. And diving into CS-related applied physics subjects (photonic logic components, micro-generators…) that please my geek instincts ^^)
Edited 2010-06-07 06:10 UTC
You mean the future of GPL’d software, not all of us buy into Stallman’s own politically motivated definition of free.
Funny I’ve used plenty of GPL software without providing any retribution. There are corporations that save millions through GPL software without giving a dime. If you have a problem with people or corporations using your work without payment then you shouldn’t open source your software.
The BSD license allows the player to be included in proprietary browsers and players. Google is looking for adoption and the BSD is a better fit.
Ok, let’s take a fairly simple definition of free software for this post, okay ? Free = open-source. The source is available. GPL and BSD fans will both try to explain that this definition is false (that free either meets Stallman’s definition or means that the user of the software does whatever he wants), but it’s the part where both licenses agree.
You provide an implicit retribution, you keep the source open. When those corporations encounter a bug in their Linux boxes and don’t feel like waiting for the kernel team to fix it, they tell 2 coders to write a patch and submit it. BSD-style licensing does not enforce the “you tweak it, you publish it” mentality. It relies on people to be honest for software to improve. If this assumption could be reliably made, we wouldn’t have a need for tribunals, which would be cool but AFAIK there’s still a lot of thieves, rapers, and monopolistic corporations in the wild so it’s not the case.
As an example : suppose that one day, Apple decides that they don’t feel like helping Google to write a web browser anymore, and want to kill the Webkit project. If all parts of Webkit were under BSD, they only have to move webkit to closed source, and then it’s a matter of months before the two browsers have become distinct and incompatible at the core, before we get a new browser war with proprietary tags everywhere. If Webkit is LGPL-licensed, to make a closed-source Safari, they have to write a new web engine from scratch, or start over from a BSD-licensed web engine which basically means the same thing. This, plus the fundamental “devs are lazy” principle, ensures that they won’t do it.
You’re right that BSD is a better fit here.
Edited 2010-06-07 05:39 UTC
Why not just call it open source?
How do I keep the source open by running binary? The GPL prevents against proprietary forks but it does not prevent against profiting from the code without retribution. Corporations have saved billions by not having to pay for commercial Unix and those savings were not re-directed towards Linux related projects. Corporations can thank GPL programmers of the world for cutting down their IT costs so more profit can be sent to shareholders.
Most users and corporations use Linux without submitting anything. Linux can be tweaked without publishing the changes if those changes are kept internally. Google for example has their own highly modified Linux kernel that they keep from the public.
Why would that lead to a browser war with proprietary tags everywhere? What would stop them from adding proprietary extensions to an open source base?
GPL advocates for years have telling us that strong protection is needed for open source to prevent proprietary forks that could undermine the original project. But we have seen that proprietary forks are not only rare but typically involve creating a derived or combined work that competes against proprietary software, not the original project.
The doomsday scenarios that GPL advocates warned about never materialized. We don’t have a hundred incompatible proprietary forms of OpenSSH. Apache is still the top web server and competes with IIS, not a fork. The fear-mongering by GPL advocates has really gotten old.
Ignoring a risk because it didn’t materialized yet in the very tiny world of current free software is not making it disappear, you know ? Otherwise, we go back to the “no one knows about the dangers of cliffs before jumping” mentality…
Edited 2010-06-07 14:12 UTC
Lets face it, if I can delve in it, code, and then do what i want with the results then thats real freedom.
Well done Google for avoiding the ‘all you code is belong to us’ madness.
http://news.cnet.com/8301-30685_3-20006245-264.html
There is some also discussion in this article about MPEG LA’s noises (veiled threats, or FUD, if you will) about assembling a “patent pool” for VP8. This discussion seems to totally miss the point that MPEG LA offering to sell paid-for permission (licenses) for people to use Google’s technology (WebM) would be leave MPEG LA wide open to accusations of tortious interference.
MPEG LA would unquestionably face a lawsuit if they tried this “VP8 patent pool” tactic, one would think.
http://en.wikipedia.org/wiki/Tortious_interference
http://www.lectlaw.com/def2/t061.htm
Edited 2010-06-06 13:22 UTC
What if Apple decide they like WebM but they want to add their crap feature, call it a new codec and sell it closed? And what if other manufacturers like the Apple idea and they all implement the “improved” codec that is purposely not compatible? We’re back at point 0. The BSD license does not ensure it’s children will be free.
The whole point about watching video is compatibility.
If Apple (or whoever) adds a feature, but makes the resulting codec purposely incompatible, then Google can still stop Apple from being able to call their forked codec Webm (via trademark law).
We would then have open, widely available Webm, and Apple’s closed, incompatible fork that cannot be called Webm. There should be absolutely no chance of the latter becoming the standard.
http://www.webmproject.org/about/supporters/
This is a large group of industry players to be incompatible with.
This should be enough to stop incompatible forks. There is no point in making something that won’t work with the Webm group’s codecs, and also paying Apple (or whoever) for the privilege of making that broken thing.
Compatible forks (say with better performance or additional features) will only add to the strength of the Webm codec.
Edited 2010-06-06 14:21 UTC
Jep.
You wish…
Remember kerberos? MS did an interesting job of undermining that with their incompatible implementation.
Anyway, I’d still say BSD is a sane choice over GPL in this particular case, though I usually license any client software I write myself under the GPL. Different licenses for different situations.
How did they undermine kerberos? By adding their own proprietary extensions? If that same software was GPLd they wouldn’t be using it at all. Would that be a win for open source? What would be gained?
Yes, because those extensions made their kerberos implementation incompatible with the standard implementation in unix. It’s not a perfect analogy to what lemur2 claimed would be impossible, but a pretty close match.
I am not claiming that at all. As said, I agree the BSD license is a sane choice in this case. But it does not provide the safeties lemur2 talks about.
I didn’t talk at all about the BSD license providing any safeties.
Trying to make myself a bit clearer: what I said was this:
The desire for correct inter-operation between all the Webm players is the only thing that should prevent incompatibility:
http://www.webmproject.org/about/supporters/
Anyone making an incopmaptible codec would not be allowed to call it Webm (which has to do with trademark law, not copyright, and this has absolutely nothing to do with the BSD license, which is a copyright license). Their new forked codec would be a different name, and it would be incompatible with the Webm stuff produced by all of the players shown on the above-linked page.
This alone is what should/could stop incompatible forks being widely used and becoming the standard. The BSD (copyright) license itself would do absolutely nothing to stop an incompatible fork.
Edited 2010-06-06 23:59 UTC
How was kerberos undermined? They used kerberos + proprietary extensions to make it difficult for competing products to connect to Windows 2000. Kind of lame yes but their actions didn’t undermine kerberos.
If kerberos was GPL’d they would have used something else or added some additional handshake.
Like a lot of software companies MS can certainly be accused of interoperability shenanigans, and I’ve had the joy of dealing with some of those shenanigans first hand but they haven’t caused any harm to the kerberos project.
There is only one reason for Apple to make an incompatible fork: DRM. This would only affect people relying on DRM video, which is not free and open in any way. The only reason for Apple (or others) to do that: not to pay royalties to MPEG LA and use an acceptable codec.
I don’t think this is the case, but if so, it’s not a big loose. Just a moral one for Apple.
The main advantage of BSD license is the fact that companies that otherwise would not touch it, could build there closed source optimize versions of encoder/decoder.
For Google this is not about money. It is about image: imposing an open standard and so being seen as a protector of open web. This is why they to push WebM so hard.
Sell it to who? To do this they would have to make it a better value than existing commercial codecs, namely H.264. So what is the worst case scenario here? They provide competition for MPEG-LA?
The GNU GPL was not designed to be “open source”. — Richard Stallman
http://en.wikiquote.org/wiki/Richard_Stallman#Sourced
Full quote:
The GPL was designed to be more than just open source. It was designed so that the code released under this license would be freedom software, it would give recipients the nominated four freedoms, and it would not only be open source, it would forever remain open source.
NOTE: This purpose for the GPL does not suit the purpose of every author/owner of source code. If it dosen’t suit your purpose, then you should do what Google did, and choose a different license which better suits what you are trying to achieve.