[00:00:00] Speaker 04: 18-1155 real-time data LLC versus Riverbed Technology Inc. [00:00:08] Speaker 04: and again, Mr. Missouri. [00:00:14] Speaker 04: Let's wait just a moment. [00:00:22] Speaker 04: Okay. [00:00:28] Speaker 03: Thank you, Your Honors Kevan Neruzi, for real-time data. [00:00:31] Speaker 03: This appeal involves a single question of law and challenges a single error of law. [00:00:39] Speaker 03: The board concluded. [00:00:41] Speaker 04: Well, weighing it at step 102 shows unidentified data, right? [00:00:50] Speaker 04: And it tells you what to do with that. [00:00:52] Speaker 04: So why isn't there a combination between that [00:00:57] Speaker 04: in a frenazack that satisfies the claim? [00:01:03] Speaker 03: I was hoping you would ask me that. [00:01:05] Speaker 03: And the answer is, there was. [00:01:07] Speaker 03: And it failed. [00:01:09] Speaker 03: And they can't win on that basis. [00:01:11] Speaker 03: And that's why they never urge it on appeal. [00:01:12] Speaker 04: Well, you'd have to explain yourself. [00:01:14] Speaker 03: I will. [00:01:14] Speaker 03: And the reason is because the claims at issue require something very specific. [00:01:19] Speaker 03: They don't only require analyzing whether the data that has come to you [00:01:27] Speaker 03: I'm sorry, they don't simply require the analyzing and then a determination of content dependent or content independent. [00:01:33] Speaker 03: They require that the analyzing not be done solely based on a descriptor. [00:01:38] Speaker 03: And at step 102, the way that Wang is determining whether that file is a known file type or not is based solely on the descriptors of the file type, and the board found that. [00:01:50] Speaker 03: We address this point [00:01:53] Speaker 03: In the very last pages of our opening brief, and if you'd like, I could cite you to the portion of the decision where the board covers this, but the, but the trucks of it is the theory based on the combination of Franichak at step 102 is one that they had, but we defeated it because we demonstrated it cannot meet the claim limitations because at that stage, the analysis is only based on a descriptor, which the claim limitations expressly forbid. [00:02:21] Speaker 04: Where do the claim limitations forbid that? [00:02:24] Speaker 03: So if we go to our opening brief, we've set forth the claim language. [00:02:37] Speaker 04: Well, let's look at the patent itself. [00:02:40] Speaker 03: Sure. [00:02:46] Speaker 03: It is either the last or second to last limitation of the claim. [00:02:50] Speaker 04: Well, we're talking about the 513 of the 992 here. [00:02:55] Speaker 03: Both of them have it in all of the relevant claims. [00:02:57] Speaker 04: Are you referring to the 513? [00:02:58] Speaker 04: I am referring to the 513. [00:03:01] Speaker 04: Okay. [00:03:02] Speaker 04: Which claim? [00:03:04] Speaker 04: Claim one. [00:03:08] Speaker 04: Okay. [00:03:16] Speaker 04: Where does it say without reference to a descriptor or whatever you say it says? [00:03:24] Speaker 01: I am pulling that up. [00:03:27] Speaker 01: This line's 45 and 46? [00:03:28] Speaker 01: Is it a very, very end? [00:03:31] Speaker 03: It states, wherein the analyzing the plurality of data blocks to recognize when the appropriate content-independent compression algorithm is to be applied excludes analyzing based only on a descriptor indicative of the any characteristic attribute or parameter. [00:03:50] Speaker 01: That's not 513. [00:03:51] Speaker 03: I might have been reading the 992 there. [00:03:56] Speaker 04: Okay and so where does step 102 say it's only based on the descriptor? [00:04:03] Speaker 03: That was the board's decision and we obtained testimony from the opposing expert. [00:04:09] Speaker 04: Where does the board say that? [00:04:11] Speaker 03: That is at [00:04:17] Speaker 04: You really should bring the hard copy of the brief and appendix with you. [00:04:22] Speaker 04: It is easier for us and easier for you if you do that. [00:04:27] Speaker 02: What page? [00:04:30] Speaker 02: Appendix 30-32 and 76-77. [00:04:39] Speaker 04: 30 to 32 and what? [00:04:42] Speaker 03: And 76 to 77. [00:04:43] Speaker 03: They're two decisions, and they are very similar to each other. [00:04:47] Speaker 04: Well, we're on 30 and 32. [00:04:48] Speaker 03: Well, if you start at the bottom of page 30, appendix 30, they talk about the arguments on both sides. [00:04:56] Speaker 03: They say patent owner argues that file extensions like those in Wang are descriptors. [00:05:01] Speaker 03: And then they go on. [00:05:03] Speaker 03: And at the bottom of 31, they say, we agree with patent owner's position. [00:05:07] Speaker 03: that the file extension names and control information in Wang appear to be descriptors of a data block, right? [00:05:13] Speaker 03: And then they say that they're persuaded by the combination of Wang, Matsubara, and Franiček, which is with respect to step 108, not step 102. [00:05:23] Speaker 03: So step 102 had no combination with Matsubara. [00:05:26] Speaker 03: It's only at step 108 that Matsubara comes in. [00:05:29] Speaker 03: And the briefing below is very clear on this point of why the combination of just [00:05:37] Speaker 03: Franiček and Wang in step 102 can't prevail since what's going on there is analysis based only on the descriptor. [00:05:45] Speaker 04: Well, I don't see that they say exactly that. [00:05:48] Speaker 04: Well, Your Honor, I... He says that the file extension names and control information in Wang are descriptors. [00:05:56] Speaker 04: So where is that saying that they analyze only based on the descriptors? [00:06:05] Speaker 03: If you look a little bit higher up on page 31, they are talking about our evidence regarding step 102 of Wang using only a descriptor. [00:06:21] Speaker 03: That's what that briefing and argument is about that they're citing. [00:06:25] Speaker 03: So if you were to pull up the cited pages of the patent owner response or the testimony that is being cited there, [00:06:32] Speaker 03: It's only with respect to step 102 and Wayne using only a descriptor. [00:06:37] Speaker 03: And I mean, I can explain this to you so that it has contextual value. [00:06:41] Speaker 03: A descriptor is exactly what's described on page 31. [00:06:44] Speaker 03: It's like when you have a word document and the end of it says dot dot, right? [00:06:48] Speaker 03: So that file name extension, that's a descriptor. [00:06:51] Speaker 03: Now, maybe that is a word document. [00:06:53] Speaker 03: Maybe it's actually a picture document, an image document. [00:06:56] Speaker 03: You don't know that for sure. [00:06:58] Speaker 03: It's sort of circumstantial evidence, right? [00:07:01] Speaker 03: What is direct evidence is to go into the file and analyze it and look at its contents and then say, this has a text and it is in a doc format, therefore it's a Word document. [00:07:13] Speaker 03: WAN does not do that at step 102. [00:07:15] Speaker 03: The claim requires that kind of analysis. [00:07:17] Speaker 03: So WAN step 102 in combination with Framichug can never prevail, and the board so found. [00:07:24] Speaker 03: What the board needed to do was rely on step 108. [00:07:30] Speaker 04: Well, unless Franiček itself uses something other than a descriptor. [00:07:37] Speaker 03: Franiček itself is itself also only using a descriptor, actually. [00:07:43] Speaker 03: So in Franiček, I will say that wasn't a theory, but to answer your point, Franiček, the way that it works is that there is a specific field in each item of data that comes in. [00:07:58] Speaker 03: that has a data type descriptor. [00:08:01] Speaker 03: And if that type descriptor is there, then Franchek uses content-dependent compression techniques that it's preselected. [00:08:07] Speaker 03: If that field is empty, therefore there is no descriptor, or if the descriptor is not recognized, then Franchek uses content-independent techniques. [00:08:18] Speaker 03: So Franchek does not do an analysis of the data [00:08:23] Speaker 03: that is not based only on the descriptor to determine whether to apply content dependent or independent. [00:08:29] Speaker 03: And that's why they didn't have either anticipation or obviousness based on frantic checks, standing alone, or even as a primary reference. [00:08:39] Speaker 03: So we have to, despite the inconvenience of it, look at step 108. [00:08:46] Speaker 03: And what happens at step 108 is that they have a motivation to combine that is based on the factual premise [00:08:53] Speaker 03: that when and if, at step 108, Wayne fails to recognize the data type, then and only then, someone would be motivated to bring in, a person of steel in the art would be motivated to bring in this teaching from Franchek of resorting to a content-independent compression technique to deal with the supposed problem of unknown data types. [00:09:17] Speaker 03: What the evidence showed, very clearly, and from both sides, [00:09:22] Speaker 03: is that at step 108 in Wayne, that problem can never occur. [00:09:27] Speaker 03: The problem of unrecognized data types is not possible at step 108. [00:09:33] Speaker 04: This last step that you referred to in claim one, where do they get that last step? [00:09:41] Speaker 04: Which reference is relied on for the wherein step? [00:09:46] Speaker 03: So they are relying on [00:09:51] Speaker 03: Matsubara to fill in the concept of direct analysis rather than analysis based on descriptor, they have a three-part combination. [00:10:01] Speaker 03: And it's Wang, Matsubara, and Franichak at step 108. [00:10:05] Speaker 04: So the last step, just to be clear, the last step, which says [00:10:16] Speaker 04: Whereas the analysis of data to recognize any characteristic attribute or parameter excludes analyzing based only on the descriptor. [00:10:24] Speaker 04: You're saying that analyzing not based only on the descriptor comes from this third reference. [00:10:33] Speaker 04: That was their contention. [00:10:34] Speaker 04: That was their theory. [00:10:36] Speaker 01: And that's the creating of histograms based on certain data. [00:10:40] Speaker 04: Yes. [00:10:40] Speaker 04: So why can't that be done in step 102? [00:10:44] Speaker 03: Why can't that be done at step 102? [00:10:48] Speaker 03: That they didn't have such a theory, Your Honor, could and would is the key distinction in this case, right? [00:10:54] Speaker 03: So like many other cases, many things could be done in many places, but the question is, would they have been done, would there have been a motivation? [00:11:01] Speaker 03: They never even put up that theory about step 102. [00:11:03] Speaker 03: Frankly, I've never thought about that. [00:11:05] Speaker 03: And I have no idea why someone would want to do that. [00:11:08] Speaker 03: There's no problem with step 102 that would cause, the method of weighing it step 102 is literally, [00:11:14] Speaker 03: Look at the file extension. [00:11:15] Speaker 03: If I know it, if it's on a list of friendlies, so to speak, pass it along from step 102 downward towards step 108. [00:11:24] Speaker 03: If it's on a list of unknowns, send it over to step 104. [00:11:28] Speaker 03: There's no reason to go into deeper analysis at step 102. [00:11:32] Speaker 03: Could someone have jammed something in there? [00:11:36] Speaker 03: Anything could be done in the field of computer science by writing code, but there has to be a reason to do it. [00:11:42] Speaker 03: And why someone would have done it. [00:11:44] Speaker 01: I may be misremembering this. [00:11:47] Speaker 01: I thought what I'm remembering as being at the core of the board's decision was to say there's evidence in the record that data types are proliferating. [00:11:59] Speaker 01: So that somebody with Wang would immediately understand that at any given moment in time, one is not likely to have all the data types. [00:12:11] Speaker 01: there's a virtue in not doing the lossless compression. [00:12:16] Speaker 01: Is that a fair description? [00:12:17] Speaker 01: Content-independent? [00:12:18] Speaker 01: Content-independent. [00:12:21] Speaker 01: That's why you have content-dependent one, because there are certain efficiencies. [00:12:24] Speaker 01: And so an ordinary skilled artisan would immediately think, let's figure out a way to see if there's a type that we're not getting out of the particular information in 108. [00:12:39] Speaker 01: And if we do have that, [00:12:41] Speaker 01: then we're going to get a benefit of using content-dependent rather than content-independent. [00:12:47] Speaker 03: Right, so this is a very, very important point, this argument about proliferating data types. [00:12:54] Speaker 03: The structure, when you look at the flow diagram of Wang and when you look at the testimony about how Wang operates, the reality is that proliferating data types would never cause a problem at Wang's step 108. [00:13:07] Speaker 03: They could maybe cause a problem. [00:13:09] Speaker 03: at steps 106 or 107, but not 108. [00:13:13] Speaker 03: And let me explain from a factual standpoint why that is. [00:13:16] Speaker 03: So what LEN does is it starts with an entire file. [00:13:20] Speaker 03: And that file may be composed of a single data type, or it may be composed of multiple data types. [00:13:26] Speaker 03: If it's a single data type, that's called a simple file. [00:13:29] Speaker 03: And if it's multiple data types, then it's a compound file. [00:13:32] Speaker 03: So let's suppose that we have a scenario with this [00:13:35] Speaker 03: you know, new kind of data that Wang has never seen before. [00:13:38] Speaker 03: It doesn't know anything about it. [00:13:40] Speaker 03: It's one of the proliferations. [00:13:42] Speaker 03: And it's a simple file. [00:13:43] Speaker 03: It's the single data type, right? [00:13:45] Speaker 03: At step 102, Wang will say, sorry, I don't know this data. [00:13:49] Speaker 03: It goes over to 104. [00:13:51] Speaker 03: Then it bypasses step 108, never reaches step 108. [00:13:54] Speaker 03: Now let's say conversely that you have a compound file. [00:13:59] Speaker 03: And within that, the file extension is known to Wang. [00:14:04] Speaker 03: But within that, there's some data type in the compound file that is unknown to Wayne. [00:14:10] Speaker 03: Now what happens? [00:14:11] Speaker 03: Well, Wayne goes down to step 106. [00:14:14] Speaker 03: So it passes step 102 because it knew the extension, so it let it through. [00:14:19] Speaker 03: Then it goes to step 106 and it makes a determination. [00:14:21] Speaker 03: Is this file simple or is it compound? [00:14:24] Speaker 03: In our instance, it's compound. [00:14:26] Speaker 03: If it is compound, it has to now first go to step 107. [00:14:29] Speaker 03: And step 107 has a really important requirement [00:14:32] Speaker 03: that again undermines their whole theory in this argument, which is that step 107 says, take all of the individual data types in this compound file, break them apart pristinely so that there's not one bit of overlapping data between the separated segmented files, so that you only have one data type per file. [00:14:51] Speaker 03: So you've taken the compound file, you've created perfect simple files, and now pass that on to step 108. [00:14:58] Speaker 03: In that process at step 107, their own expert admitted [00:15:02] Speaker 03: that Wang has to know what the data type of each piece of the compound file in Wang is. [00:15:11] Speaker 03: It has to determine that. [00:15:13] Speaker 03: So if, due to a proliferation of data types, somehow Wang were unable to grapple with some data type that were in the compound file, it would fail out at step 107. [00:15:22] Speaker 03: The file would not make its way to step 108, where Wang would then face the challenge of, what do I do about this unknown type? [00:15:29] Speaker 03: And their own expert testified on that. [00:15:32] Speaker 03: And so this notion of trying to use the idea of proliferating data types that would stymie Wang doesn't get them to the combination of 108. [00:15:41] Speaker 03: At best, it indicates a potential failure at step 107. [00:15:44] Speaker 03: It doesn't give a person of steel in the art a reason to put in a solution to unknown data types at step 108. [00:15:52] Speaker 03: And again, there's testimony on that. [00:15:54] Speaker 03: And the challenge with the board's decision is that it does not address these points. [00:15:59] Speaker 03: It doesn't consider these arguments. [00:16:01] Speaker 03: These arguments were made [00:16:02] Speaker 03: in great detail below. [00:16:03] Speaker 03: And in the first argument of the day, Judge, I heard you say that it's true that the decisions of administrative agencies don't need to spell out every single thing that they consider and address every little argument. [00:16:14] Speaker 03: But there are numerous instances where the silence as to key issues that are of significance is a basis for error. [00:16:22] Speaker 03: And our briefing has been all along that this is one of those situations. [00:16:26] Speaker 04: I think we're out of time, particularly when we start going back to the first argument. [00:16:37] Speaker 04: Mr. Bell. [00:16:40] Speaker 05: Good afternoon. [00:16:41] Speaker 05: May it please the court. [00:16:42] Speaker 05: I'd like to start with my colleague's last argument there in which he said that there was, at Wang step 108, [00:16:51] Speaker 05: It wouldn't be able to fail. [00:16:54] Speaker 05: In other words, it would always be able to recognize data types at step 108. [00:16:58] Speaker 05: And the fact is the board expressly considered that argument. [00:17:02] Speaker 05: If you look at page A22 and A23, the board walks through in meticulous detail real-time's argument, including citing the very testimony that real-time says supports its position. [00:17:19] Speaker 05: That's testimony from Dr. Prezere. [00:17:21] Speaker 05: You can see that at the very bottom line of 22, where it cites exhibit 2010. [00:17:26] Speaker 05: A few lines up, cites 2010. [00:17:28] Speaker 05: And then it goes to the next page. [00:17:31] Speaker 05: And in the second full paragraph, not only does it say we are not persuaded by patent owners' arguments at the beginning of the first full paragraph, but in the middle of the second full paragraph, it specifically calls out this argument. [00:17:44] Speaker 05: The board says, [00:17:45] Speaker 05: with respect to patent owner's argument, and I'm quoting here, that Wang never fails to recognize any data types. [00:17:51] Speaker 05: And it's talking about it at step 108 there. [00:17:54] Speaker 05: And then it goes on to give its reasons for rejecting that argument. [00:17:58] Speaker 05: And it had very good reasons for doing so. [00:18:00] Speaker 05: And the question here is not a question of law, respectfully to my colleague. [00:18:05] Speaker 05: It is a question of substantial evidence. [00:18:08] Speaker 05: And at these pages, pages A23 and A24, the board lays out very clearly [00:18:14] Speaker 05: why it saw a motivation to combine the Franchek reference with Wang to provide the fallback, if you will. [00:18:21] Speaker 05: What happens, and that's what all of these references are dealing with, is we would prefer to use a content dependent. [00:18:28] Speaker 04: Okay, does this depend on having identified data at step 108? [00:18:33] Speaker 04: Data types? [00:18:35] Speaker 05: The data types in Wang are only, the only step that talks about that is step 108. [00:18:41] Speaker 04: And so- Oh no, step 102. [00:18:43] Speaker 04: talks about recognizing file format, which is a subset of data type, right? [00:18:49] Speaker 05: Respectfully, no, Your Honor. [00:18:50] Speaker 05: So some clarity on that point might be helpful. [00:18:53] Speaker 05: A file format is a more general mechanism to look at whether you recognize something. [00:19:01] Speaker 05: So for example, the .doc or the .pdf or whatever it might be, that might give you a general sense of the file format. [00:19:10] Speaker 04: So you can be able to identify the file format without identifying the type of data, is what you're saying? [00:19:17] Speaker 05: Precisely. [00:19:18] Speaker 05: It might collapse into the same thing if the file only contains one data type. [00:19:23] Speaker 04: Okay, so it's step 102 that it hasn't been the case that all data types have been recognized, right? [00:19:33] Speaker 04: Not necessarily, no. [00:19:35] Speaker 04: So your theory is it's going up the chain. [00:19:38] Speaker 05: Exactly. [00:19:38] Speaker 04: And their theory is that [00:19:40] Speaker 04: 107 prevents it from getting to step 108. [00:19:44] Speaker 04: Is that correct? [00:19:44] Speaker 05: That is their theory, as I understand it. [00:19:47] Speaker 05: But the board considered exactly that argument. [00:19:49] Speaker 05: They made this argument below, presented it in great detail. [00:19:52] Speaker 05: The board probed it at the oral hearing and ultimately disagreed with it. [00:19:57] Speaker 05: And there is substantial evidence to show, including from our expert, Dr. Creuzer. [00:20:02] Speaker 05: So to the extent my colleague suggests that this is somehow undisputed, that our expert admitted that there would always be [00:20:09] Speaker 05: that it would never fail at step 108. [00:20:11] Speaker 05: That's simply not the case. [00:20:12] Speaker 05: And I would point the court, for example, to A 7422 and 7423. [00:20:20] Speaker 05: And at those pages, and that's one of many examples, A 7417 is another one, 7419 is another one. [00:20:31] Speaker 05: But I'll focus on 7422. [00:20:33] Speaker 05: And there, Dr. Pruzer testifies, quote, absolutely [00:20:38] Speaker 05: absolutely one of skill in the art would understand that at Wang's step 108, it might well fail to recognize data types. [00:20:49] Speaker 00: In other words, it's... What lines are you looking at on page 7422? [00:20:53] Speaker 00: I might have the right page. [00:20:54] Speaker 00: Yes, it's towards the bottom. [00:20:57] Speaker ?: Okay. [00:21:00] Speaker 05: And specifically, it is line 23. [00:21:03] Speaker 05: Okay. [00:21:05] Speaker 05: So, the question was, [00:21:07] Speaker 05: And this was in Dr. Corsair's deposition. [00:21:10] Speaker 05: So you say that a person of skill in the art would have read Wang to leave open the possibility that there might be unknown data types when you get to the bottom of that, as Your Honor said, the chain. [00:21:20] Speaker 05: The answer, absolutely yes. [00:21:23] Speaker 05: And if you go to the next page, he explains one of the reasons why that is. [00:21:27] Speaker 05: And on the next page, he talks about that proliferation of data types that Your Honor referenced a little while ago. [00:21:33] Speaker 05: Because if you're going to build a module that is trying to recognize [00:21:37] Speaker 05: data types, as Wang does in step 108, and there's a proliferation of data types, you're never going to be able to cover all data types that appear in the future. [00:21:47] Speaker 05: And so that's what he explains there. [00:21:49] Speaker 05: That's not his only reason, though. [00:21:51] Speaker 05: If you look at other pages, such as 7419, he gives a very specific example of how it could go through this entire chain and recognize the file format at step 102 [00:22:06] Speaker 05: break down the file in step 107, and nonetheless come out at 108 with unknown data types. [00:22:14] Speaker 05: And he uses the example of an HTML file. [00:22:16] Speaker 05: And I was trying to think of an analogy. [00:22:18] Speaker 05: And I sort of came up with the notion that you can recognize something as a bus. [00:22:22] Speaker 05: You can take out all the kids and line them up. [00:22:25] Speaker 05: You don't necessarily recognize each of the kids. [00:22:28] Speaker 05: I think that's kind of what he's getting at there. [00:22:31] Speaker 05: And certainly, when you step back to the level of review we're dealing with here, [00:22:34] Speaker 05: This provides more than substantial evidence for the board's decision. [00:22:38] Speaker 05: And the board went through, again, on pages A, 23, and 24, and gave at least four pieces of evidence that support its decision for why there would be unknown data types at step 108. [00:22:51] Speaker 05: The first is the Franchek reference. [00:22:55] Speaker 05: The Franchek reference itself is very clear. [00:22:58] Speaker 05: It says, if you're trying to do content-dependent compression, [00:23:04] Speaker 05: and you don't have a known data type, you could get in real trouble because it could be catastrophic if you lose one bit of data as to certain data types. [00:23:12] Speaker 05: So in other words, if you can't figure out the data type, you better use something safe. [00:23:16] Speaker 05: That's what Franchek is saying. [00:23:18] Speaker 05: And so for that reason alone, there would be a motivation to sweep that in and connect the dots to Wang step 108. [00:23:25] Speaker 05: But the board doesn't stop there. [00:23:27] Speaker 05: It goes on and says, and Dr. Zeiger [00:23:30] Speaker 05: a patent owner's expert testifies it. [00:23:33] Speaker 05: There's a proliferation of data types. [00:23:35] Speaker 05: That's the point we were discussing earlier. [00:23:37] Speaker 05: Goes on to say, Wang itself, in a general sense, that's step 102, although not directly applicable, in a general sense, Wang itself recognizes that there's this problem. [00:23:48] Speaker 05: If you encounter something unknown, there it's talking about file format, but the underlying principle is the same. [00:23:54] Speaker 05: If it's something unknown, you better use a safe compression method. [00:23:58] Speaker 05: That provides further [00:23:59] Speaker 05: support. [00:24:00] Speaker 05: And then finally, the testimony of Dr. Prisier, in particular paragraph 86 of his declaration, where he walks through all of this evidence and says at the end of the day, a person of ordinary skill would understand that you get to step 108 and there might be unknown data types. [00:24:16] Speaker 05: He even gives one example, and that is the Matsubara reference, which it's now just to be clear, it's undisputed that that combination is proper. [00:24:25] Speaker 05: That was heavily disputed below. [00:24:27] Speaker 05: But here, there's no dispute that Matsubara is used at step 108 to implement the how. [00:24:33] Speaker 05: In other words, how do you recognize the data types? [00:24:36] Speaker 05: And Dr. Kuzera here says, you might, at the end of doing that, using Matsubara, you might have an unknown data type. [00:24:45] Speaker 05: The board credited that. [00:24:47] Speaker 05: The board found that. [00:24:49] Speaker 05: And that, we submit, is ample substantial evidence. [00:24:53] Speaker 01: Take it that in Wang itself, which is not a very long document, there is nothing that says when you get to what has now been labeled 107, if the decomposition process doesn't recognize some child in the bus, that I think Mr. Neruzzi said the [00:25:24] Speaker 01: the process ends rather than goes to 108, that there's nothing in this description that says it ends rather than continues? [00:25:31] Speaker 01: No, absolutely not. [00:25:35] Speaker 05: Unless the court has further questions. [00:25:39] Speaker 04: OK, thank you, Mr. Mell. [00:25:41] Speaker 05: Thank you. [00:25:51] Speaker 01: Is that right, by the way, Mr. Nouriz? [00:25:54] Speaker 01: Just at Wang itself, there's nothing that says if box 107 fails to recognize some portion of the overall file, the whole process stops. [00:26:07] Speaker 03: So Wang does not contemplate any of its steps failing, and it doesn't discuss what happens if any of them do. [00:26:14] Speaker 03: But the testimony from the opposing experts was that Wang would have to know how to recognize the data type to succeed at step 107. [00:26:23] Speaker 03: Wayne is not performing step 107 as Todd and Wayne, then how can we go to step 108 and Wayne? [00:26:29] Speaker 03: We're in a different world now. [00:26:30] Speaker 03: We're in a different algorithm. [00:26:31] Speaker 03: It's a different flow chart. [00:26:32] Speaker 04: But he also testified that going on to step 108, you were going to have unrecognized data types, right? [00:26:41] Speaker 04: Their contention was that... No, no, his testimony, his testimony. [00:26:44] Speaker 03: Whose testimony? [00:26:45] Speaker 04: Their expert. [00:26:46] Speaker 03: Their expert did not say that. [00:26:48] Speaker 03: their experts said the opposite. [00:26:49] Speaker 03: Their contention was that our experts said that, that Dr. Seeger on our side admitted that, and that's incorrect. [00:26:55] Speaker 03: The testimony they pointed you to in the record at appendix 7422, I believe it was, from our expert, is just talking about whether there can be a situation in Wang where Wang does not recognize a data type, but he is talking about step 107 where the decomposition is happening [00:27:16] Speaker 03: or step 102 where the file format is being recognized. [00:27:19] Speaker 03: If you look at that testimony, there's no mention of step 108, nor is that in any of the sort of preface to the question. [00:27:27] Speaker 03: And that's the whole point. [00:27:29] Speaker 03: And on this, I would direct you to the record below where our expert, Dr. Zegar, put in a declaration afterwards explaining this and explaining that it's never been his position that it's step 108 that can be a failure of data type recognition. [00:27:46] Speaker 00: I don't know, their experts declaration at paragraphs 84, 85, and 86. [00:27:52] Speaker 00: Why isn't that substantial evidence to support how Wang works and what Wang is missing and how Wang works with respect to step 108 and step 102 and why it is that somebody would look to Matt Zavara? [00:28:13] Speaker 00: I see an explanation here in their expert report. [00:28:16] Speaker 00: So I want to know why you think that isn't substantial evidence. [00:28:18] Speaker 03: Does the flaw with the explanation in their expert report is exactly what we exposed in the deposition and have been briefing, namely that all the expert says is he puts forth the predicate condition, the predicate condition that if and when Wang were to fail at step 108, then someone would be motivated to look to this approach and frame a check. [00:28:41] Speaker 03: But he never gives any evidence that in any instance, Wang would fail at step 108. [00:28:48] Speaker 03: And then on deposition, he admitted Wang itself never suggests that. [00:28:51] Speaker 03: A person of steel in the arc would have no reason to believe that. [00:28:54] Speaker 03: And that if Wang had a data type recognition problem, it would happen at step 107 before step 108. [00:29:00] Speaker 03: And so there's no evidence in his declaration that step 108 could actually lead to a failure. [00:29:05] Speaker 03: I just want to add one critical last point, which is [00:29:09] Speaker 03: My colleague pointed to Appendix 68 in the board's decision where he said that the board addressed our argument that this data type recognition can't happen at step 108. [00:29:19] Speaker 03: The board did not address our argument. [00:29:21] Speaker 03: The board said, quote, with respect to Pat and owner's argument that Wang never fails to recognize any data types, Toma, Franichek recognizes the issue and teaches the use of a default compression algorithm if a data type is not recognized. [00:29:34] Speaker 03: That's the entire basis of our appeal, that the holding of the board's decision. [00:29:39] Speaker 04: OK, I think we're out of time. [00:29:42] Speaker 04: Thank both counsel. [00:29:43] Speaker 04: The case is submitted, and that concludes our session.