# Oral argument — In Re Neill > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A16282 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** March 7, 2016 ## Text Our first case today is 2015-1431 in Re Neill. Mr. Crye, am I saying that right? You are, thank you. Please proceed. Good morning, your honors. The board denied Cablevision patents for its remote DVR technology by cherry-picking elements from four unrelated prior art references. That decision was not supported by substantial evidence and the court should reverse. Cablevision developed a unique technological solution to a problem that no one had even tried to solve before. How do you take a home DVR, move it into the cloud with all the advantages that cloud computing entails while still complying with copyright law? Let me ask you a sort of housekeeping question. Certainly. At the end of their red brief, Apley says that despite alluding to considerations of expense and complexity, Neill offered no declaration evidence and no evidence of secondary considerations of non-obviousness. Is that correct? That is correct, your honor. But the court doesn't need to reach secondary considerations because even on the face of the board's own rationale That's why I said housekeeping. Cablevision solved this problem through a unique arrangement of elements, including a novel multiple buffer system. In that system, data first passes through a primary buffer. I think we all understand how the invention works. Why don't you jump right to what the board got wrong in its determinations? Certainly, your honor. Where the board went wrong is that it claimed that there were motivations to combine these four elements from unrelated prior art references and attributed those motivations to the prior art references themselves. But the references simply do not say what the board claims. And the first example of that is Walters. The board relied on Walters for a disclosure of a multiple buffer system. Walters does disclose a multiple buffer system. As the board found, it discloses a buffer and a controller and the board held that a controller was And you didn't argue that's not a multiple buffer system. You only argued the motivation to combine point, correct? Under the substantial evidence standard on appeal, that's correct. So tell me why they got motivation to combine wrong. Because the motivation that the examiner and the board found to combine those references was that using multiple buffers would reduce cost. And there's simply no support in the reference that the board referred to that supports that. The only reference to cost reduction in Walters, which is what the examiner cited, was a discussion of a completely unrelated aspect of the invention. In the Walters disclosure, there is a discussion of the buffering system on one page of the specification. What the examiner did was he took that reference to a multiple buffer system and then flipped back two pages in Walters to a completely unrelated aspect of the invention where Walters explains that if you transmit programs in 30-minute segments rather than entire programs, that can be more cost effective. And then the examiner said that that cost reduction rationale for a completely unrelated aspect of the invention was a reason to combine Walters' multiple buffer system and only that multiple buffer system with the RSDVR. And your honors, that just doesn't make sense. You can't attribute a motivation relating to a completely unrelated aspect of a prior art reference as a reason to combine some different and separate aspect of that prior art with the invention in suit. And just as a matter of common sense, the theory that you can combine or line up multiple buffers in series and thereby reduce cost doesn't make sense. It's counterintuitive and it's certainly not supported by Walters. The reason that CableVision's remote DVR... Hold on for a sec. What the quote says from Walters is such an embodiment reduces the cost of the receiver at the subscriber's location by reducing the amount of storage required. Wouldn't that also be true in a multiple buffer system? We're talking about the receiver. We're not talking about... Wouldn't it reduce storage on the receiver in the CableVision system? So the receiver includes both the buffers and the storage. And what Walters is referring to, the reason that embodiment reduces storage cost is that if you transmit programs in 30-minute segments rather than entire programs, you only have to store the 30-minute segment on the hard drive in the Walters system. That has nothing to do with how many buffers the data passes through on the way to that hard drive. If I can direct Your Honor's attention, the cost reference in Walters is on column 6, page 414, which is, I think, what Your Honor's referring to. And that's contrasting two different embodiments of the transmission system, one where Walters transmits entire programs, and then a different one where Walters transmits programs in 30-minute segments. And he says at lines 48 to 54 that transmitting and storing the programs on the hard drives in 30-minute segments is more cost effective. The buffer discussion, Walters, is several columns later. It's on page A416, column 9, from lines 26 to 43. And that's where Walters discusses what the board construed to be and which we're assuming on appeal to be a disclosure to a multiple buffer system. And that has nothing to do with whether Walters is transmitting entire programs or 30-minute segments. There's nothing in Walters, and there's nothing in common sense, that says having data pass through two buffers rather than one buffer on the way to the hard drive reduces cost. All Walters says is that if you transmit programs and you store them in 30-minute segments on the hard drive, that reduces cost. But that is not anything relevant to Cablevision's invention, and it's not anything that the board incorporated into the remote DVR when it found this motive to combine. And then, as if that weren't enough, your honors, the board did the exact same thing with LEAD. The board relied on LEAD for the idea that you can have de minimis fragments of programs, because those de minimis fragments are an important part of the RSDVR. What makes the RSDVR's buffer system novel is that the primary buffer, the one under the service provider's control, contains only a de minimis amount of data. And then the secondary buffer, the one under the user's control, can allow a larger amount of data to accumulate. Now, the board said that LEAD disclosed de minimis fragments, and then it incorporated LEAD, or claimed there was a motivation to combine those de minimis fragments with the primary buffer in the DVR. But there's simply no motivation to combine there. What the board cited was a passage from LEAD, where LEAD says that its overall transmission architecture allows for better bandwidth construction and lossless delivery. But that was a paragraph from an introductory description of LEAD's overall broadcast scheme. It had nothing to do with the de minimis fragments in LEAD. The only place where LEAD mentions de minimis fragments is on column 9. This is on page A431 of the appendix. This is the only thing in LEAD where he mentions de minimis fragments, and this is what the board and the examiner cited. And this is not describing anything to do with the transmission scheme, let alone a recording scheme. It's a discussion of a mathematical algorithm that LEAD discloses for computing what is the optimal segment size to transmit under LEAD's transmission scheme. This is from lines 31 to 52, and there is a single reference there. It's around line 37, where in the first step of that algorithm, what LEAD does is he divides up a program into segments, each of which only has one frame of video. And then the algorithm, what you do is mathematically, you repeatedly combine those segments until a formula tells you that you've reached the optimal transmission size. And then that is what LEAD says you should transmit through each of your streamers. That's the only place where LEAD talks about de minimis fragments. He does not say that you should transmit de minimis fragments, and he certainly doesn't say that it makes sense to have a primary recording buffer that stores only a de minimis amount of data. But he does say you can break it down into even a single frame, and then at column 4, where they talk about the fragmentor 204, I assume that's the actual portion of the system that is breaking it into these little bits, and it talks about the fragmentor, and it says each of these media objects is broken up into segments by the fragmentor, and then it talks about how they're periodically broadcast, and then when you go all the way down to line 60, it talks about them being stored in download store 112. Right. Two points on that, Your Honor. First, the fragmentor does not store the de minimis frames of data. The de minimis frames of data are separate. The mathematical algorithm starts with de minimis fragments, but then it combines them over and over again until you get the optimal segment size. So that does not disclose that it makes sense to have streamers, each of which is only streaming a single frame of data. That would be 100,000 hardware streamers for a single program, but that doesn't show up anywhere in Lee, and it doesn't make any sense. But beyond that, even if you thought that Lee disclosed a system where you stream and then store a de minimis amount of data after transmission, that's the transmission phase. Lee doesn't say anything about recording programs, because Lee is not a DVR. It's not a recording system. It is a system for optimizing the broadcast of programming, and it tells you what is the most efficient way to divide up a program and then broadcast it to the user. There's simply no motivation to combine the fragment size used in broadcasting with the fragment size that's used in the RSDVR's recording buffer, because that's a completely different stage of the process. In the remote DVR, the recording buffers are where the data goes through on its way to being recorded at the head end before the transmission size, and there's no relationship, none whatsoever, between how many buffers that data passes through on the way to the hard drive, what the size of those buffers are, and how, on the other hand, you transmit that data back to users. Lee is a transmission scheme. There was no motivation to combine the fragment size used in that transmission scheme, let alone the fragment size used in this theoretical algorithm as a preliminary step to determine that transmission scheme and incorporate that into the primary recording buffer of the RSDVR. Again, what the board did with Lee was the same thing it did with Walters. It said, oh, well, here's the element I want to combine. Then it went to a completely different page of the specification and found a motivation describing a completely unrelated aspect of the invention and said, well, that motivation for this aspect of the invention is a reason to combine just this one feature of the invention with other prior art. That's not substantial evidence. It's reasoning that doesn't even make sense on its own terms. The board did that with Walters, and the board did that with Lee, and in both cases, its findings of a motivation to combine were not supported by substantial evidence. In addition to those two references, which are by themselves grounds enough to reverse the board's decision, there was also a separate and independent error in the board's consideration of the fourth reference, Applebaum. The board held that Applebaum disclosed exclusive user control, but Applebaum does not disclose exclusive user control in the sense that Cablevision's invention uses that term. For the remote DVR to comply with copyright, it's important that users have exclusive control as against the service provider, not just as against other users, which is all that Applebaum discloses. If I may reserve the rest of my time. That's fine. Okay. How do I say your name? Rusilla? Rusilla. Rusilla. Thank you, Your Honor. May it please the Court. As directed by KSR and its progeny from this Court, the claimed subject matter here is unpatentable as obvious because it's merely a combination of familiar components according to known methods from other prior art transmission and storage systems that yield predictable results. Here, the examining board identified... If that's your argument, how in the world does multiple buffers lead to a predictable result of reducing cost? That doesn't seem very much a predictable result. You incorporate multiple buffers, you're increasing cost. You're not decreasing cost. You may be increasing some complexity and some level of cost, but what Walters teaches is that overall, you get a cost savings because you're saving in the storage system the way that it's set up in Walters. And what Walters teaches in the receiver, because that's what the language in Walters says, is that the storage system in the receiver saves costs. And he teaches a storage system that has... And how is that tied to having multiple buffers? The storage system in Walters may reduce costs by virtue of other aspects of what is disclosed in Walters, but I have a very difficult time linking that to the multiple buffer. And that's the only thing you're taking from Walters. The only thing you're taking out of Walters is the multiple buffers, and you're combining it with Ellis and Applebaum and Lee. So how in the world would you... Why would you reach into Walters and say, oh, let's take these multiple buffers? The board relied on... Well, they would do that because Walters says it'll reduce costs. But I don't see how Walters says multiple buffers will reduce costs. There's something about Walters that reduces costs, but not multiple buffers. Well, two points on that. First, the buffers do reduce costs in Walters, and I'll get back and explain that in a second. The second point is that the board also relied on increased performance as another motivation to combine. Now, going back to how the multiple buffers in the storage system reduces costs, the multiple buffer system in Walters acts very much like the dual buffer system in the claims. What Walters teaches is that you have a first buffer, the first memory, 330, where the information comes in. And that checks the bid. Isn't he correct, though, when he's explaining the layout of Walters? Column 6 is where the cost reduction. It's tied to such an embodiment reduces costs in the receiver. And this is where it has set out two separate embodiments, one that sends whole programs and one that breaks programs up into 30-minute segments. And when it says such an embodiment, this is with direct reference to the 30-minute segment in transmissions, correct? That's correct. Yes, so such an embodiment reduces costs, as opposed to the other embodiment, which doesn't break into 30 minutes, but also uses multiple buffers, right? Right. So both embodiments use multiple buffers. That's correct. But only one embodiment is articulated as reducing costs. That's correct, and that multiple buffer system allows the overwriting, which is the cost savings discussed in Column 6. So the buffer system works very much like in the claimed invention. Here you have the initial buffer, the primary buffer. The information comes in. It checks the VID to see if the program's been ordered. If the program's been ordered, then it will allow the system to save that information in storage. So once it shows that that information has been ordered, it then strips the header protocol, the transmission header protocol, out of those packets and then passes it up the line. It passes it up the line to the controller. Now that controller then aggregates the data and then Walters teaches us that that information, it checks to see if there's VID data that's embedded periodically in the video data that's stored in the storage, in the controller, in the buffers in the controller. Are you trying to make this purposefully dense? No, I'm not trying to. Just curious. So apologies if I am making it more complicated. Then the controller, that information, it holds on to the video data segment that has been streamed in and then overwrites it to you. The reason I ask my question about the density is the statement in Walters seems just unequivocally clear. The cost saving is by overwriting the program. You take a 30-minute segment and instead of needing two hours of space on a DVR to record a two-hour long program, if you're going to record 30 minutes and then the next 30 minutes, next 30 minutes, and you're just going to overwrite, isn't that all that this is, overwriting? It's like if I had an old-fashioned tape. You're too young to remember those, but like an old-fashioned cassette tape. You could record something on it and then guess what? You can actually record over it on the same thing and then that way you save storage. So isn't that all that's at issue here is recording a 30-minute segment and then the next 30-minute segment, instead of putting it in a separate spot in memory, just recording it on top of the original 30-minute segment? Which is exactly what the second buffer allows the system to do. Because the second buffer in the controller holds it because when the viewer is watching the program, that program is that first segment is stored in memory and then that segment is being played out of memory. So that second segment is downloading that information while that first segment is being played and it has to hold it in the controller until that information is provided to the viewer so then it can overwrite the system so that there's no breaks in the system and so there's no latency problem. So the way that WALTRS teaches us the primary buffer and the secondary buffer, that's how it can get the cost savings in the storage. So I would submit that the statement in column 6 where it talks about the cost savings in the storage in the receiver is directly facilitated by the two-buffer system that's taught in WALTRS. And so I think that is substantial evidence for the motivation to combine on cost. But even if you don't agree with the cost one, the board also found improved, or the examiner also found improved performance is another reason why you could use the two-buffer system here and import that into ELLIS. And you get improved performance because you are transmitting less information over the transmission line so you get less bandwidth requirements and so it's a more efficient system. And the ability to have the two buffers to allow different manipulation of the data happening in parallel also allows increased performance and that allows the users to view the data in real time as it's being played back. All of those things assist with that so you have increased performance as well which is an additional motivation to combine that's separate and distinct from the cost savings. With respect to Leigh, Counsellor Neil argues that you can't import the de minimis teaching from Leigh because it's embodied in an algorithm and that it doesn't discuss recording de minimis fragments. I think both of those arguments are incorrect. First, with respect to Leigh, what the board found, the board did not take the algorithm, did not use the algorithm and try to plug it into the system. What the board looked to Leigh for is that it's a teaching that one ordinary skill in the art would understand you could fragment video down into smaller segments and then repackage them based on whatever design considerations you have on both the receiving side and on the reception side. And that's what Leigh teaches. Leigh teaches that you can minimize bandwidth on the server side. You can also minimize bandwidth on the storage side, the client side. And what it does, it looks at both of those and tries to find what value of the number of frames in a segment will make those two bandwidths most alike. It minimizes the difference between the two, the delta between the two. And that's what it decides it's going to package the data into segments and then send it along the system. So what Leigh is actually teaching, it teaches not only minimizing bandwidth on the clients, on the transmission side, it also talks about it on the storage side. And Your Honor was absolutely correct when you were talking about what the structure of the fragmentor is. The fragmentor, as it's taught in this reference, it breaks the data down to single frames and then adds additional ones to find the optimal segment size. So it is actually storing each frame at one point and then it builds upon that. So it does teach storage of single frames. It may be just a very short period of time, but it's still storage of those single frames. And so one ordinary skill in the art looking at this would understand there's a lot of different ways that you can set up a system to fragment and minimize provided daemon in this amount of data and then transmit that over the system based on bandwidth requirements on both the sending side and the client side, the reception side, the storage side. How often do you see four obvious, four reference obviousness rejections? Because I'll be honest, I've never seen one. That's a lot of references you've got to piece together to get to obviousness. It's a good number of references, but in this case the examiner found a motivation to combine. The question I asked was actually quite personal. How often have you seen a four reference obviousness rejection? Because I never have. Not out of the PTO, never. I've seen people try to argue it in litigation, sure. But I've never seen the PTO go to four separate references and cherry pick items and then combine them together. Well, I don't know if I've seen a four reference one, but I don't believe that the board or the examiner... Or anything greater than four. Let's be clear so this deposition transcript reads right. Four or greater. That's a lot of references. It is, but the examiner here, all of those references come from the same field of art. It's all the transmission and recording of video programming data. Yeah, but the whole reason the cable version came up with this invention is to get around Sony. The whole reason they came up with this is to get around the copyright problems with DVRs in homes and try to reduce the amount of storage you need in your home. That's so different than pulling together these pieces where you say, well, this one mentions in the background cost savings. And this one mentions in the background efficient bandwidth. And this one mentions... And then you're just cherry picking random elements from these references and smooshing them together to come up with the invention. And it's... I don't know. You're awfully lucky this is a very deferential standard of review. That's for sure. Well, as to your first point, as to the motivation as to why the applicants put the known and familiar components together, copyright compliance, KSR and the project from this court teaches us that the motivation of the applicants does not control the obviousness of that analysis. I mean, it's the objective scope of the claim itself and what one of ordinary skill in the art at the time would have done. And so here, while their motivation is laudable, I mean, we don't want people infringing copyright, but the examiner and the board found motivations within the references themselves and used them to show how one of ordinary skill in the art would put these known components together in ways that everyone understood how they reacted to get predictable and expected results. So I think substantial... By looking at the motivations combined provided by the board and the examiner, there is substantial evidence to support the obviousness rejection. I might address one thing that just popped into my mind. And that is... And I'm going to ask the panel to do the same. In the blue brief towards the latter part at 44, it says it is a technological solution to the legal problem of ensuring copyright compliance during the recording system. It reminded me when you were saying that. And my question is, is Poseida in this instance, or at least one of them, a lawyer? And I'm not being facetious about that. Because I couldn't find any place where there was such a thing. I think that that strikes on the problem of the term de minimis and tying it to a legal determination. It would make the claims indefinite. If they really tried to read in this de minimis copyright construction, then you will have circuit courts that may have different decisions on what exactly de minimis is. So you would have different standards depending on where you're located in the country. So then the person of skill is the judge. And I found some cases that say the judge isn't either. That's correct. So I think that the problem with importing that and trying to use that to show some kind of patentable weight here is that first, it's not recited in the claims. And second, it would render the claims indefinite. And so then again, you have to look back to the objective analysis. And the board and the examiner pointed out motivations to combine for all of the prior references. And with respect to Alphabom, the last one, Alphabom clearly discloses in paragraph 33 that an administrative user is the only entity that's able to configure the hard drive. So therefore, one of ordinary skill in the art in reading that would understand that that could also apply to the service provider and not just to the different users. And that is the section that the board and the examiner relied upon to find that motivation to combine. So their argument with respect to the board's failure to find a service provider and only rely on the users is not correct because that passage in Alphabom is not just solely directed to the users. And one of ordinary skill in the art would understand that you could use that to provide exclusive control to the user that would prevent the service provider from configuring the hard drive as well. Okay, thank you. Let's have Mr. Cry have his rebuttal time. Mr. Cry, I wanted to ask you that question and you sat down. Sorry. So that page 44 question about the legal aspects of it. Right. This isn't a technical solution to a problem that exists because of copyright law, but the person of skill in the art doesn't need to be somebody legally trained. The problem Cablevision was trying to solve was design a remote DVR system such that no storage under the service provider... But in order to satisfy, somebody has to be legally trained. I don't agree, Your Honor. The problem they were trying to solve is how do you have a remote DVR in which no storage of more than a de minimis amount of data is ever under the service provider's control. That is a technical problem. What's de minimis mean? Well, as the examiner found, it's a term that means very small and that's a term from dictionary definitions and from other definitions. There's no suggestion in the rulings below that there would be any legal indefiniteness problem with that term. And the term is elaborated on in the prosecution history as well, which gives examiners... The suggestion to me came from me reading your brief. No, even in the specification, there's references to particular percentages of a program and so there's more than enough guidance in the specification to add content to that term. Judge Moore, you commented that the government was... The PTO was lucky to have a deferential standard of review, but I don't think that's going to be enough for them here. And the reason is that even under substantial evidence review, the rationale that the board itself gives has to make sense. If they say a reference says a certain thing, the reference actually has to say that. And I think as you put your finger on during the questioning of the PTO, that's a standard they can't meet. The board said and the examiner said the reason to combine the multiple buffer system of Walters with these other references is that it would reduce cost. And then they cited a particular passage from Walters. And there is nothing in that passage for all the reasons that I think came across clearly during questioning that says that increasing the number of buffers reduces cost. The most that Walters says is the... But Walters says this system will reduce receiver and storage side cost at the expense of multiple buffers, but it does say that. Right, but that's not an element of Walters that the RSDVR incorporates. The RSDVR cannot break programs down into 30-minute segments and then overwrite them because that would make no sense for a DVR. If you're recording a show, it doesn't do you any good to have a system that records the first half and then records the second half over top of it. That wouldn't work as a DVR, which is why, not surprisingly, the board and the examiner never said they're incorporating that aspect of Walters. All that the board and the examiner took from Walters was this multiple buffer system, and that has nothing, nothing whatsoever to do with reducing cost. So under any standard of review, however deferential, there is no evidence at all supporting that critical finding of a motivation to combine. And the same thing is true for LEAD. On LEAD... Your time is up, so if you have a final thought. My only final thought, Your Honor, would be that rejections based on four prior art references are rare for a reason, and it's not just four references, but four references combined in a very particular, very novel way to solve a problem that no one had even thought of before. Why didn't you argue secondary considerations? I know I told you your time is up, but I just, I can't help it. Why didn't you argue them? It was 23, 23 years, 25, 23 years between Sony and your invention in this case as applied for. That's a long time for people to be puttering around to find a way around Sony. Long felt need, failure bothers. Why didn't you argue any of this stuff? There no doubt are secondary considerations that we could develop in terms of the record before this court. There's more than enough basis to reverse on the basis of the board's rationale, and we might, may well, may well pursue those arguments on remand, Your Honor. Thank you. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A16282. Public record. Not legal advice.