Google open codec wins OSI love after patent shield rethink Google has rejiggered the license on its open-source VP8 video codec after complaints that it wasn't really open source. Today, with a blog post, Google's Chris DiBona announced that the company had switched to a pure BSD license, ending a bit of a feud with various open source stalwarts over the codec. On May 19, after …
It's going to be an interesting next couple of years. Methinks Google had better do something quick though, before this Microsoft-inspired antitrust suit potentially makes them unable to just say "Youtube is now in VP8. You use VP8 or you don't use Youtube."
That'd certainly futz things up a bit for Microsoft and Apple.
I'm very sure that Apple and Microsoft (well, at least Apple) have or will have VP8 add-ons at the ready - all it would take is one automatic download later (or Safari/IE8 coughing up a pop-up asking if you'd like the plugin, then reaching out to their respective motherships, or even Google, to get it).
>But at least we can all now agree that VP8 is the way forward — with the exception of Apple, Microsoft, and the MPEG-LA.
Microsoft said that if you have VP8 installed then IE9 will use it.
I believe the same is the default behaviour in Apple Macs.
Firefox and Opera are on board...
Drinks all round?
Oh an hahaha MPEG-LA, hahaha :-)
A genuine royalty free, open-source, high-quality video codec would be a winner. But can anyone conclusively say that it isn't open to accusation that the developers have 'borrowed' from patented ideas/technologies? Patents have been a major headache for some time now, but the idea isn't all bad. If you've invested a huge amount of time and money developing something, I think it's only reasonable to have a fair opportunity to make a return on it. It won't happen if others just run off with your work and sell it as their own.
I'm confident this is the case with MPEG-LA - I don't doubt they have spent a truck-load of money putting together H.264 (and putting up barriers against patent trolls). But can we just take Google's word for it that there's no cause for concern? After some of the decisions they've made, I'm not so sure. Are they providing legal protection for companies that use VP8?
@Random_Walk: "If we sat around and waited until all the patent trolls were satisfied, we'd never get anything done."
Disappointing - I hoped for a more enlightened answer than that. I certainly hope there is one, or VP8 will remain on the fringes. The companies driving H.264 have made a huge investment to mitigate the risk of legal action by patent trolls. We're all unhappy this has to be done, but it's a commercial reality that can't be ignored. And having gone to this effort, why should investors in H264 abandon it for VP8 if they are only exposed to more risk?
Seriously - you expressed a hope that every software house alive (including, I suspect, the former On2) does their level best to satisfy, if only for their own fiscal preservation.
Unless you're looking for some sort of iron-clad guarantee that On2 paid its danegeld to MPEG-LA, or that they're holding various execs' children as ransom, there's no way to answer your implied question to anyone's satisfaction.
The reason MPEG-LA is cozy is that they've got buy-in from (nearly) all the big players, and have a litigation war chest the size of Siberia... and nothing else.
Patents are not supposed to be issued for "mere methodology." An algorithm, though, is far more complicated than a method. To use your example: There are dozens of ways of solving any algebra problem, even something as simple as x=2+2; however, an algorithm defines a single set of instructions or calculations that must be followed verbatim for it to work properly. The definition is often written at a higher level than simple algebra because some of the details are not relevent to the algorithm at a patent level, but only at a copyright level.
A physical example could be given for Eli Whitney's cotton gin. Very useful, as it allowed cotton to be processed by machine instead of by hand. The patent here needs to specify how the parts interact to allow the machine to seperate the seeds from the useful cotton. However, if the patent were to specify every detail, then someone else could simply scale the plans up slightly and circumvent the patent.
This is exactly what a good algorithm does: it specifies exactly how a problem is solved by specifying a required set of steps to solve it. Algebra does not specify a set of steps, only a set of tools that may be applied to an equation. Algebra doesn't even recognize a difference between a problem and a solution.
Unfortunately, yes, there are many patents that are issued that are outside the bounds of actual patent law, but this is not an indication that the law is broken. Only that the patent office is as overworked and poorly staffed as ever. There are countless patents for physical processes that are no better than some of the software patents that have been issued.
"A physical example could be given for Eli Whitney's cotton gin. Very useful, as it allowed cotton to be processed by machine instead of by hand. The patent here needs to specify how the parts interact to allow the machine to seperate the seeds from the useful cotton. However, if the patent were to specify every detail, then someone else could simply scale the plans up slightly and circumvent the patent."
Patents on machines are somewhat different from those being foisted on algorithms: generally, the design of a machine gets a patent because even if the design were copyrighted, it wouldn't stop anyone from using the design to make the actual machine. In short, the "protection" for such physical artifacts is moderated by patents.
Meanwhile, granting a patent on an algorithm prevents the expression of that algorithm - frequently as demanding as devising the algorithm in the first place - in any form, and where vagueness creeps in, often deliberately, such patents often seek to forbid any activity in the domain of that algorithm. Taking such vague patents back into the physical domain, it would be like a patent on a machine that monopolises *any* processing of cotton. And there are, of course, ethical reservations about granting monopolies on thought processes, which is what patents on algorithms are.
"Unfortunately, yes, there are many patents that are issued that are outside the bounds of actual patent law, but this is not an indication that the law is broken. Only that the patent office is as overworked and poorly staffed as ever."
So the solution is to sink even more money into the patent bureaucracy rather than spread it around amongst the people actually discovering new ways of doing things? I don't think so somehow.
"However, if the patent were to specify every detail, then someone else could simply scale the plans up slightly and circumvent the patent."
That is nonsense ! Simply scaling-up would be 'obvious' in the strict patent sense.
Patents describe inventions in sufficient detail that 'one skilled in the art' can understand/repeat the process/construction or whatever
So what if Google responds to the MPEG-LA action by saying, "I'll pit MY patents against YOUR patents." That means MPEG-LA would run the risk of some of their patents being invalidated or submarined--a patent war. Google isn't taking licensing fees from its patents, so MPEG-LA stands to lose more from a patent war.
I think Charles has it right - Google know exactly what they are doing, and they are forcing the hand of the MPEG-LA group.
Up until now, the MPEG-LA group have been able to stifle competition by the well known techniques of FUD - ie by telling all and sundry that they have the **ONLY** "safe" option available, and that anyone who thinks differently will get their house burned down by legal costs. Hmm, sound just like those old fashioned "protection" fees extorted by gangsters of old - pay the "insurance" or your property won't be safe from burning down.
Now the MPEG-LA group have to "pup up or shut up". If they ignore this new codec, then any future patent enforcement action is doomed - they have to take action against any known infringement of their patents or they lose their rights to do so. So they now have to take action against Google - who have enough money to fight them.
Since we can assume Google has a few patent cards in it's hand, I expect to see a few rounds of very public mud slinging, followed by sue and countersue in the courts, followed by an agreement that Google won't shutdown the MPEG-LA groupt's H.264 codec provided MPEG-LA don't shutdown Google's. Mutual Assured Destruction in the digital age - but just like the cold war equivalent, it only works if both sides have powerful weapons.
And already pointed out, Google (via Youtube) have the means to force vendors to include their codec as standard - even Microsoft and Apple couldn't get away with shipping systems that can't view Youtube videos of cats dancing or whatever. So in practical terms, vendors of all sizes will have to ship both codecs in their systems.
But, if MPEG-LA try to screw people on the licensing (ie put the price up once it's commonplace), then they run the risk of people just telling them where to go - after all there's a free alternative. So MPEG-LA will have to keep the costs down below the threshold where people ditch them and they lose their ubiquity, and their ability to charge is even further eroded.
I have huge reservations about Google and the power it wields on the internet, but this action I definitely approve of.
"If they ignore this new codec, then any future patent enforcement action is doomed - they have to take action against any known infringement of their patents or they lose their rights to do so."
Whilst that does apply to Trademarks, it doesn't to patents. Would make the system a lot better and slightly less prone to trolling if it did though!
While I'm sick of this software patent mayhem US of A has wreaked upon us, I concur that VP8 is frighteningly similar to H.264 in the patent-covered areas, meaning google's probably got some interesting times ahead.
With the amount of money MPEG-LA are looking to lose out on VP8 if it succeeds (and when they finally raise interweb H.264 license fees like the arseholes they are), M$ et al. are unlikely to back off either.
Litigatious departments report in full gear!
Microsoft are being very careful when choosing their anti-linux battles. They have to win via private agreement because if it becomes worthwhile for the opposition to fight the whole thing become precedent. If it becomes precedent then other companies also appear to be involved and must also defend or pay by default. Even if the patent holder wins in court they may find that a later ruling overrides their claim and they would have to give all the money back.
It is not the winning or loosing that matters here but simply the going to court . It is too big a risk for the big portfolio owners to risk.
That's mainly a point about the choice of word... but...
For many users and developers, H.264 is very cheap or free till 2015, I think. MPEG LA then get to hike the price to whatever the market will bear. I assume also that while the patents grow old and wither, new, newly patented features will be added to the codec system, and their inclusion will be compulsory. It's "1984" for real!!!
Animated PNG for freedom ! ! ! ! !
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
