Daring Fireball: On Submarine Patents vis-à-vis H.264 and Ogg Theora(daringfireball.net)
daringfireball.net
Daring Fireball: On Submarine Patents vis-à-vis H.264 and Ogg Theora
http://daringfireball.net/2010/03/on_submarine_patents
6 comments
HTML5 video is effectively a tech demo at the moment. Nothing relies on it. The HTML5 video situation doesn't really feel worth discussing until Google plays their hand in regard to On2 VP8. If they freely license the codec and promise to use their patent arsenal to defend licensees against lawsuits, VP8 will become the most practical license, except for low-powered devices like mobile phones until chips are developed to encode and decode the format via hardware.
Firefox's current H.264 stance isn't practical, but it doesn't have to be. HTML5 video won't matter for years (I'd estimate two years after the release of IE9), and by then, the landscape will have changed.
Firefox's current H.264 stance isn't practical, but it doesn't have to be. HTML5 video won't matter for years (I'd estimate two years after the release of IE9), and by then, the landscape will have changed.
A main design consideration of VP8 was fast decode on low powered mobile ARM chips.
This is similar to Windows users being confused that Mac OS X or Linux doesn't need anti-virus, and actually thinking this is a fault with these strange new operating systems. H.264 is just a bit of a pig and yet its resulting requirement for special considerations to be made to compensate is interpreted as a failing in others that don't need it.
This is similar to Windows users being confused that Mac OS X or Linux doesn't need anti-virus, and actually thinking this is a fault with these strange new operating systems. H.264 is just a bit of a pig and yet its resulting requirement for special considerations to be made to compensate is interpreted as a failing in others that don't need it.
I found the linked-to article at Streaming Media more interesting than Gruber's Theora fear mongering. http://www.streamingmedia.com/article.asp?id=11746
Specifically this:
For 15 years, Xiph.Org has carefully "played by the rules," fully within the bounds, intent, and letter of intellectual property and patent law. For the past ten years we've informed the entire world, including MPEG LA, of our specifications and algorithms in detail. We've requested in open letters that any group believing we are infringing to inform us so that we make take immediate corrective action.
I predict that MPEG LA may counter that they know groups have been pressured into licensing patents in order to use Theora. This has been a recent back-room assertion. You might want to ask point blank if MPEG LA itself or any of its constituent members has engaged in this practice, thus manufacturing the evidence that "vindicates" their patent allegations.
Specifically this:
For 15 years, Xiph.Org has carefully "played by the rules," fully within the bounds, intent, and letter of intellectual property and patent law. For the past ten years we've informed the entire world, including MPEG LA, of our specifications and algorithms in detail. We've requested in open letters that any group believing we are infringing to inform us so that we make take immediate corrective action.
I predict that MPEG LA may counter that they know groups have been pressured into licensing patents in order to use Theora. This has been a recent back-room assertion. You might want to ask point blank if MPEG LA itself or any of its constituent members has engaged in this practice, thus manufacturing the evidence that "vindicates" their patent allegations.
More interesting is the MPEG-LA quote itself. Gruber summarises it as " MPEG-LA has stated that they believe otherwise" i.e. MPEG-LA have claimed that Theora uses their patents. What the CEO actually says is intended to give that impression, but never actually says it:
"In addition, no one in the market should be under the misimpression that other codecs such as Theora are patent-free."
This is a beautifully crafted sentence because Theora isn't patent free, it just so happens that all the patents we know about are licenced freely for anyone to use. (Similarly VP6, 7, 8 are all presumably patented but Google now owns those patents and can do what they like with them.)
"Virtually all codecs are based on patented technology"
Again, whose patents are we talking about here? On2/Google's, MPEG-LAs?, some random trolls? It is left unspecified. And which codecs? Not all of them, only "virtually" all.
"and many of the essential patents may be the same as those that are essential to AVC/H.264."
Keyword "may", as in "might not", or as in I'm very carefully not saying what you think I'm saying by adding qualifiers.
"Therefore, users should be aware that a license and payment of applicable royalties is likely required to use these technologies developed by others, too."
There's a "likely" which you'd think is stronger than a "may" but only applies if you are using technologies developed by others. So this "likely" depends on you accepting the "may" in the previous sentence.
"MPEG LA would consider offering on additional licenses that would make these rights conveniently available to the market under a single license as an alternative to negotiating separate licenses with individual patent holders."
This is just stating the MPEG-LA business model, batching up patent rights for easier licencing. It does not say that the hypothetical patent owners are currently part of any MPEG patent pool.
"In addition, no one in the market should be under the misimpression that other codecs such as Theora are patent-free."
This is a beautifully crafted sentence because Theora isn't patent free, it just so happens that all the patents we know about are licenced freely for anyone to use. (Similarly VP6, 7, 8 are all presumably patented but Google now owns those patents and can do what they like with them.)
"Virtually all codecs are based on patented technology"
Again, whose patents are we talking about here? On2/Google's, MPEG-LAs?, some random trolls? It is left unspecified. And which codecs? Not all of them, only "virtually" all.
"and many of the essential patents may be the same as those that are essential to AVC/H.264."
Keyword "may", as in "might not", or as in I'm very carefully not saying what you think I'm saying by adding qualifiers.
"Therefore, users should be aware that a license and payment of applicable royalties is likely required to use these technologies developed by others, too."
There's a "likely" which you'd think is stronger than a "may" but only applies if you are using technologies developed by others. So this "likely" depends on you accepting the "may" in the previous sentence.
"MPEG LA would consider offering on additional licenses that would make these rights conveniently available to the market under a single license as an alternative to negotiating separate licenses with individual patent holders."
This is just stating the MPEG-LA business model, batching up patent rights for easier licencing. It does not say that the hypothetical patent owners are currently part of any MPEG patent pool.
I'd actually missed the so-called "clarification" which, ihmo, makes it clear he's trying very hard to create a false impression:
"Ozer: It sounds like you are saying that some of your patent holders own patents that are used in Ogg. Is that correct?
Horn: We believe that there are patent holders who do.
Ozer: It sounds like you’ll be coming out and basically saying that to use Ogg, you need to license it from MPEG LA. Is that correct?
Horn: That is not what we said. We said no one in the market should be under the misimpression that other codecs such as Theora are patent-free"
So specifically asked about the patents in the H.264 pool, he responds with the non-specific "there are patent holders who do" (yes, On2 now Google) and this is couched with a "we believe".
And when asked directly, he repeats his non-statement about Theora having patents on it. Yes, we know that. What about patents held by people who are looking for money for them?
"Ozer: It sounds like you are saying that some of your patent holders own patents that are used in Ogg. Is that correct?
Horn: We believe that there are patent holders who do.
Ozer: It sounds like you’ll be coming out and basically saying that to use Ogg, you need to license it from MPEG LA. Is that correct?
Horn: That is not what we said. We said no one in the market should be under the misimpression that other codecs such as Theora are patent-free"
So specifically asked about the patents in the H.264 pool, he responds with the non-specific "there are patent holders who do" (yes, On2 now Google) and this is couched with a "we believe".
And when asked directly, he repeats his non-statement about Theora having patents on it. Yes, we know that. What about patents held by people who are looking for money for them?
"We've requested in open letters that any group believing we are infringing to inform us so that we make take immediate corrective action."
This is nice, but does it really protect them? It seems that a patent holder has more incentive to take a wait-and-see approach.
This is nice, but does it really protect them? It seems that a patent holder has more incentive to take a wait-and-see approach.
It seems that a patent holder has more incentive to take a wait-and-see approach.
This runs the risk of allowing a laches defense, which patent holders probably would like to avoid.
This runs the risk of allowing a laches defense, which patent holders probably would like to avoid.
Are there actually any cases in which laches had been successfully invoked in a patent dispute?
For what I understand, unlike trademarks, patents don't have a provision that you should protect them or enforce them. In fact that's why defensive patent portfolios work, isn't it?
For what I understand, unlike trademarks, patents don't have a provision that you should protect them or enforce them. In fact that's why defensive patent portfolios work, isn't it?
Laches doesn't invalidate a patent; it simply allows you to defend yourself in the particular case you raise it in.
Laches is a form of estoppel, which is geared around the basic idea that if you do something I could sue you over, and you can show that:
1. I did something which gave you a reasonable belief that I wouldn't sue you, and
2. That was the only reason you went ahead and did it, then
3. I don't get to turn around and go after you, because it was really my actions which created the situation.
In a laches defense, you show that the person suing you for patent infringement knew you were doing or were going to do something which infringed, and deliberately waited to come forward with their claims so as to do the most possible damage to you. You argue that, had you known about their claims earlier, you would have done something different, and so it was their inaction which led to the current situation.
This doesn't stop the patent holder from going after other people, so long as they do it in a timely manner; the patent is still valid and can still be enforced. It also doesn't automatically stop future claims against you; it's expected, now that you're aware of their patent, that you'll work out some licensing deal to avoid future infringement. But it can throw out their current case against you, or at least throw out the massive back damages you'd owe from infringement over the period when the patent holder didn't bother doing anything.
Laches is a form of estoppel, which is geared around the basic idea that if you do something I could sue you over, and you can show that:
1. I did something which gave you a reasonable belief that I wouldn't sue you, and
2. That was the only reason you went ahead and did it, then
3. I don't get to turn around and go after you, because it was really my actions which created the situation.
In a laches defense, you show that the person suing you for patent infringement knew you were doing or were going to do something which infringed, and deliberately waited to come forward with their claims so as to do the most possible damage to you. You argue that, had you known about their claims earlier, you would have done something different, and so it was their inaction which led to the current situation.
This doesn't stop the patent holder from going after other people, so long as they do it in a timely manner; the patent is still valid and can still be enforced. It also doesn't automatically stop future claims against you; it's expected, now that you're aware of their patent, that you'll work out some licensing deal to avoid future infringement. But it can throw out their current case against you, or at least throw out the massive back damages you'd owe from infringement over the period when the patent holder didn't bother doing anything.
"Laches doesn't invalidate a patent"
I disagree. In Symbol Technologies Inc. v. Lemelson Medical, the courts declared Lemelson's submarine patents unenforceable under the doctrine of prosecution laches.
I disagree. In Symbol Technologies Inc. v. Lemelson Medical, the courts declared Lemelson's submarine patents unenforceable under the doctrine of prosecution laches.
The Lemelson case was a rather extreme example; normally, a laches defense only applies to a specific defendant or defendants in a specific case. With the Lemelson case, the court explicitly noted that it was doing something extraordinary, and went to some lengths to explain how that situation differed from a typical laches case.
IANAL, but a quick googling turned up a couple of cases. http://www.patentlyo.com/patent/2008/10/laches-and-equi.html
http://www.patentlyo.com/patent/2008/08/inventorship-di.html
Apparently cases called Aukerman and Scimed established a precedent of 6 year for invoking a patent claim. "[A] delay of more than six years after the omitted inventor knew or should have known of the issuance of the patent will produce a rebuttable presumption of laches."
Apparently cases called Aukerman and Scimed established a precedent of 6 year for invoking a patent claim. "[A] delay of more than six years after the omitted inventor knew or should have known of the issuance of the patent will produce a rebuttable presumption of laches."
While "submarine patents" is a catchy term, they aren't actually relevant to this discussion.
Interestingly enough, the article cites a wikipedia discussion that explains why. (Short version - submarine patents, secret applications that turn into issued patents years later, don't exist these days because the relevant agencies publish applications within 2 years after application. Yes, there's an exception in US law, but it makes the invention unpatentable in other jurisdictions.)
Interestingly enough, the article cites a wikipedia discussion that explains why. (Short version - submarine patents, secret applications that turn into issued patents years later, don't exist these days because the relevant agencies publish applications within 2 years after application. Yes, there's an exception in US law, but it makes the invention unpatentable in other jurisdictions.)
If Theora gets sued over a patent that H.264 also violates wouldn't MPEG LA want to back up Theora as to not give credibility to the patent? If they didn't the suit against Theora would just strengthen the patent against H.264. I guess, H.264 can counter-sue to protect themselves, maybe.
He's posted an update saying Google only uses Theora because they already have an H.264 licence, so they're safe.
2 counterpoints: 1) so does Apple yet they're not using Theora because of alledged patent risk, 2) Google only provided Theora with the open source Chromium. Under this theory, why can they ship Theora but not H.264?
I think we can conclude that he's just making this up as he goes along.
2 counterpoints: 1) so does Apple yet they're not using Theora because of alledged patent risk, 2) Google only provided Theora with the open source Chromium. Under this theory, why can they ship Theora but not H.264?
I think we can conclude that he's just making this up as he goes along.
And if you’re not willing to concede that Ogg Theora is vulnerable to patent claims, well, then I doubt you’re reading this. It appears that Gruber thinks it's so obvious to any reasonable person that Ogg Theora is "vulnerable to patent claims" that the only people who would disagree are those closed-minded enough that they would avoid reading a different opinion.
Of course "vulnerable to patent claims" is weasel language. Almost everything is "vulnerable to patent claims" in the sense that you can't know for sure that at any moment someone won't emerge from the woodwork and claim they've got a patent on it. But: for H.264 it's known that there are people claiming to have patents on the technology, namely the MPEG LA; and it's known that they demand payment for (certain kinds of) using it. For Theora no such thing is known, its creators have disclaimed patent rights, and it's been around for a long time without anyone making such claims. For both H.264 and Theora it's also possible that at some point in the future someone might come forward and say "oh, by the way, I own this and this and this; please pay me $1M/year or get sued".
But if someone is trying to choose betwen Theora and H.264 on the basis of their patent situation, it's daft to say (as Gruber does) that the only consideration is that possibility of getting torpedoed by submarine patents. The big difference in the situation with known patents is obviously relevant too.