https://www.courtlistener.com/api/rest/v4/audio/105315/?format=xml105315/audio/105315/google-llc-v-parus-holdings-inc/https://www.courtlistener.com/api/rest/v4/dockets/73447008/?format=xml2026-06-05T10:33:12.762912-07:002026-06-05T10:36:49.898103-07:00CGoogle LLC v. Parus Holdings, Inc.9c342ce281de93bfb0eee651459e469dccb62baahttps://www.cafc.uscourts.gov/oral-arguments/24-2381_06052026.mp3mp3/2026/06/05/google_llc_v._parus_holdings_inc._cl.mp3mp3/2026/06/05/google_llc_v._parus_holdings_inc..mp32153TrueFalse11The next case is number 242381, Google LLC v. Parus Holdings, Inc. Okay, Mr. Haber. May it please the Court. Benjamin Haber for Appellant, Google. This is the fourth time that related issues have come before this Court. Each prior time, the Board found the challenge claims invalid, and this Court affirmed. Different claims. Different claims, true. Substantially similar claims. Well, maybe. So, can I just bring you to what I guess I'm coming into the argument most focused on? Yes. One-I. The one-I element about unexpected response. That language and concept, I think, is not in the earlier case, the 941 claim. That was about change generally, modification generally. The fact that there was either uncontested or even an adverse decision to the patent owner on the broader notion of any change does not imply and is not inconsistent with an opposite conclusion for the narrower one-I element of unexpected response. That, as I say, is what I'm thinking about in the most focused way as I come into the argument. Okay. So, let me address that point directly. So, element one-I, as Your Honor noted, requires reducing rank number in the event of an unexpected response. And so this term, unexpected response, doesn't appear in the patent. You're just going to combine one-I and one-J, and I understand that they kind of go together. So, when you look at the term unexpected response, you have to look at what the specification says. You didn't ask for a claim construction on that, right? No, Your Honors. We did not. So, the board gets to take the notion kind of at face value and make any-take any understanding of that term that is reasonable under that language, which has not been construed. I don't think that that is correct, Your Honor. I think that the board has to understand that term in view of the specification. I think understanding that term in view of the specification informs the plain and ordinary meaning. And what the board did, essentially, was look at the term unexpected and say, well, if something somewhere somehow is expected, then it's not unexpected. And I think that is a claim construction position that is not consistent with the specification. And so- Could we back up just a second? If I understand what the petition did, it assumed that the claim construction that you're arguing for now, that is, that any change is an unexpected result, was sort of the correct construction, correct? I think we read the specification, and that is the understanding that the specification provides. Yeah, and I don't think the patent owner response disputed that. Am I correct about that? That is true. In fact, the patent owner did not contest Element 1i as to the Kovach Burroughs grounds. The first time we saw the board's position on Element 1i is in the final written decision, where they took this view of the claim that is not consistent with the specification. So are you arguing that we're in an exonic situation where the board has adopted a new claim construction without notice? Yes. Yes, I think that's true. I think Google, the echo factor also applies. The board's decision affected the scope of the claims, and they did it without notice and an opportunity to respond. And I think that is something that this Court reviews de novo. And when the Court looks at the specification, I want to direct Your Honor's attention to Appendix 56. At the bottom, Appendix 56, column 16, around line 55, going on to column 17 around line 5, describes kind of what the patent believes is expected versus unexpected. And they talk about expected formats. Expected formats specified by the content descriptor file. The content descriptor file is a piece of software that is looking for particular content on a website, and if that content is not in the format that it expects, that is an unexpected format. And then at the bottom, and this is important, column 17 around line 4. This is like the alarm? Yes, so there is a mechanism that generates a warning message or alarm indicating that the specified website has been modified or is not responsive. And so it can be changed. It can have a modified format. It can have modified content. The issue is that whatever it's looking for, it doesn't find, and that is the unexpected response. And so it may find it because it's changed. It may find it because the format is unexpected. There's a number of reasons why it might be unexpected. Right, and if there's a website with the weather, you absolutely expect change every five minutes. It cannot be that any change is unexpected. You know, a lot of very common websites that are designed to be keeping information up to date. I think that's a fair point, and one of the points that the board makes is that obviously websites can change, and the web evolves and develops, but the issue is whether or not change is expected. And there are certainly cases where you're looking for a particular paragraph, and the paragraph is no longer there. And the software that is looking for it knows where to find it, and it knows what it's looking for, and if it doesn't find it because it's changed or for whatever reason, that is an unexpected response. And so that's what the specification describes, and that's what… So what was the position that you took in your petition about the scope of unexpected response? Now I think you're saying, well, yes, of course, not any change could fall into that. It's some narrower set of changes. Let me pull up the petition onto this element for you. I thought the petition argued the broader construction. I'm sorry, I didn't… I thought the petition assumed the broader construction that any change fell within the scope of unexpected. Well, so this is a computer-based piece of software, and so there's a certain piece of software called a content descriptor file that goes out and fetches the content. And let me find what we said here. So we talk about… What page? This is appendix 153. Thanks. So the browser periodically searches the web to determine the index page status. The response is not as expected because the information has changed. A modified page would decrease in rank. Its handle has a delete and add operation. And so the changes that we're talking about… You say at the bottom, right, at 153, after the semicolon, the response is not as expected because its information has changed. If your position there is the very broad one, any change comes within unexpected response, that can't be right. So a departure from that position, what are we supposed to do when the board says, well, that obviously isn't right, and you haven't made any narrower position, asserted any narrower position about what subset of changes would be unexpected? Right, and I think that the patent is instructive here. The subset of changes that are unexpected is where the software is looking for information, and the information is not found, and it's not found for any number of reasons. It may be changed. It may be changed in location. It may be changed in formatting. There are particular changes to what the software, the content descriptor file expects that cause this unexpected response. It's not… Suppose that the board had said to you, we think your implied claim construction is wrong. Unexpected results means what we think it means and the way they described it. But we'll give you the opportunity to file a reply that addresses that question of whether the prior art meets this claim construction. Did you have evidence that you could have presented that the prior art showed unexpected results under the board's construction? So I think we do. I think even… So a few things. I think you have to consider what it is that has these expectations. And what the board said essentially is that everyone knows the Internet can change, and so it's not unexpected because everyone knows it can change. You have to focus on the software. The software is programmed. It's programmed in a certain way, and what we would have explained, citing to the specification and citing to extrinsic evidence. We have extrinsic evidence, for example, exhibit… I'm sorry. What about… Not the claim construction evidence, extrinsic or intrinsic, but the prior art that you rely on. What in the prior art did you or would you have relied on or did you rely on to say this prior art teaches a particular kind of change that might qualify under the less than fully broad understanding of unexpected response? Yes. There's a portion of Burroughs. That's exhibit 1006. I'm looking at appendix 459. Appendix? 459. That talks about deleting entries. Yeah, the column 15. Yeah, column 15. There is another portion that describes… And where does that say something about, I don't know, a kind of change that would… The particular part of… Would be a change that we did not expect, but we see. Yeah. There is a portion that talks about the maintenance module, and the maintenance module treats changes as a delete and odd operation. That's on page 454, column 5. Is that the one called maintaining? I'm looking at page 458, appendix 458. Okay. It may periodically be necessary to omit modified or new entries and expunge deleted entries, and so there is a particular teaching in Burroughs about what to do with these modified entries. When there is information that has changed, or it's not what you expect, you essentially remove it and then add it back in, and it gets re-ranked at that point. And that's what we identified as teaching claim element 1i. The idea that there is these changes that perhaps happen, and you need to respond to it. You need to respond to it by updating the ranking information. And so we would have pointed to Burroughs. We would have pointed to the specification. We would have pointed to our expert testimony. We have a paper that was put out by IBM that specifically talks about the fact that the web is dynamic and changes, and you need to update search engines and other issues in response to these changes. The key point that I wanted to stress is that there are two grounds. There's Kovach Burroughs, and there's Wise Burroughs. And the key issues that we have on appeal here are not substantively disputed. There's no dispute that the Board did not set out its opinion with respect to element 1c, which was the key disputed element all throughout below. There's no dispute that, as to the Wise Burroughs grounds, the Board's decision is not consistent with what they said in the 941 and 402 appeals. There's perhaps disputes on claim construction, but you will not hear a patent owner make an argument that somehow these decisions can be squared. It is not in their brief, and it is not actually accurate. And just to be clear, if the Board was fine on 1i, 1j, none of that would matter? No, Your Honor. I don't agree with that. The other issue, and this pervades both grounds, is that the Burroughs reference was considered only in isolation. There are findings about Wise and Kovach that bring into the combination very important aspects, kind of the precursors to H, I, and J. And specifically, they define how the ranking list is, they define the rank number that is used by these softwares. And Burroughs provides specific teachings on what to do with that rank number, but very importantly, you have to consider those teachings combined with the primary reference. Your Honor, I'm not understanding this. Why is it that if we were to agree with the Board on 1i and j, that you could still win? Because if you look at what the Board said about, and I'll direct you to Appendix 23, I think this is kind of a summary of what the Board did with respect to Burroughs. They went through a very long discussion of what Burroughs does and doesn't do, specific to Burroughs' implementation. By the time they got to the end, where they were actually going to focus on the combination, as KSR tells them to do, the combination is Kovach, Kovach's figure 4, the destination tree, implemented with Burroughs' teachings. And they just didn't consider that combination. They never considered how the teachings to raise and lower rankings from Burroughs would have been applied to the infrastructure and the database tree of Kovach. And that is an obviousness problem. It is not consistent with KSR. It looks at the references only in isolation. So putting aside the errors that the Board didn't set out its findings with regard to key disputes. I'm sorry. I'm still not understanding how that helps you on 1i and j. Let me direct you to Appendix 404. Appendix 404 is Kovach's figure 4, the destination tree. And what Appendix 404 shows is that there are a number of websites. In this example, under books, there are two websites, Amazon and Barnes & Noble. And what we established, all the way getting through the claim elements up to element H, is that these websites would be ranked with a particular number. And then, when the ranked number needs to be changed, it's these particular websites and these particular bits and pieces of software that would do that. All the Board did was look at Burroughs' specific implementation of its teachings. It didn't look at how those teachings would have been applied to this infrastructure in Kovach. And that's the key problem with the Board's decision. Because we don't know what the Board thought with respect to Kovach's figure 4, other than the rehearing request that was granted that said Kovach taught all of the claim elements that essentially resolve the majority of Parrish's arguments. And the only thing left for the Board to do is to apply Burroughs' teachings in the context of Kovach. And that's why we've argued that this case should be remanded, so that the Board can actually do that analysis. Okay. I think we're out of time. We'll give you two minutes to rebuttal. Thank you. Mr. Fowler. Good morning, Your Honors. May it please the Court. Charles Fowler for Parrish Holdings. The Board's decision here is replete with factual and credibility findings, any one of which is inconsistent. Well, there is this inconsistency argument, as Judge Geronto pointed out, at least with respect to the unexpected response element, the claim language is clearly different, the scope of the— Okay, but putting that one aside, you've got a problem, right? I don't think we have a problem, Your Honor. Why not? A couple of things. One, and I think at the appeal stage, we're focused on the 402 and the 941 IPRs. There's really no issue left with respect to the 431 IPR. So we're talking about 402, 941. And the Board, there's this notion that the Board didn't address these prior IPRs. It's not a question of whether they addressed them. It's a question of whether there's inconsistency. And, Your Honor, the typical way the Board would deal with a purported inconsistency is under the rubric of collateral estoppel. And that's exactly what Google— No, that's not the only way. Yes, Your Honor. There's this Vicor— And there's Brunetti, which either one of you bothers to discuss, which addresses this specifically. Your Honor, and, you know, I think the key—well, two points. I mean, first, I mention collateral estoppel because that's how Google presented it in its petition. Google—not petition, because the 941 and 402s hadn't been decided yet at that point. They raised it for the first time, and they reply, fine. They were new decisions at that point, but they presented it as collateral estoppel, and they explicitly withdrew that at the hearing. And so the Board, in its decision, Appendix 21, Split Note 7, says Google presented a collateral estoppel argument based on 402, 941. Okay, it's not collateral estoppel, but that's not the only ground. Right, but that's the only ground that was—that was the only ground that was raised. So we would say the only ground that was raised was withdrawn, and so there wasn't a separate inconsistency point. Just so that I'm orienting myself, do I remember right that there was a colloquy at the—whatever they call it, the oral hearing or something, the oral argument in front of the Board in which they said something that they want to interpret as asserting inconsistency, and you want to interpret as just saying, you know, we were right then. You should draw the same conclusion without invoking, you know, inconsistency as a separate—is this—is that the state of— That is all correct, John. Okay, so suppose that the question of inconsistency in a Vicor, Brunetti sense is, you know, not forfeited. Putting aside, again, 1I, 1J for now, what is your argument about what to do about that issue? Sure. Key distinction is separate records. These whole line of cases, Vicor, are identical issues on identical records. Some of these unpublished cases, the same expert testimony, the key is the same record. Most cases don't suggest that the record has to be identical. I mean, what you have here is a situation in which the prior art was characterized differently, found to disclose different things in the two proceedings. And, Your Honor, again, the big distinction, the claim, the ranking, and polling elements that are disputed here are part of the independent claims. They were part of dependent claims in the prior IPRs that were not disputed. So none of the expert testimony, none of the argument or evidence in the record in this case disputing those elements was part of the record. They simply were not disputed, and the board rendered a fact finding insofar as it made findings on, again, not exactly the same. Okay, that would be a perfectly sound reason for distinguishing the prior decisions, but we don't have that. They didn't do it. And I would say they don't need to, Your Honor. And if, again, the patentee here, you know, these IPRs, you end up with all of these challenged elements, and the patentee was more than at liberty to make an election in those prior IPRs. We're going to focus on the independent claims. We're not going to focus on these other elements in the dependent claims. And it would, if the board were there, then it were incumbent upon the board to specifically engage with past board findings on related patents. I guess maybe one way to make the point is that it might well be enough for the board to say, in this previous matter, we read this prior art in the following way. We now have better airing of that issue, and we now draw a different conclusion. And agencies do get to change their conclusions when there are new arguments, new evidence, whatnot. But the board didn't engage in that discussion. I mean, even in Vicor, you know, we sent it back and say, you have to address this. You can't just be silent about it. I think, Your Honor, it's a critical distinction that Vicor was dealing with the same disputed issue. It wasn't a past finding on an undisputed record. The board references it in Appendix 3 and Appendix 21. Clearly, it's aware of these prior decisions. Google didn't present. All the more reason it should have addressed them. You know, I don't think there's any precedent whatsoever that would suggest under these circumstances with a prior undisputed finding that there's a – I mean, essentially what Google is asking for is a firm rule, an obligation on the board to say a sentence. I mean, I think Google would say, look, if we were – if the board had put the one sentence that Judge Johnson suggests. Or, you know, some kind of – well, we considered and we recognized that that was – some kind of acknowledgement that there's a difference here because it's disputed here, it wasn't disputed there. That would be enough. But that's new law. I mean, that would require a new rule that this court has not announced, not even as – by extension of I-Corps. Can I turn you to 1I and 1J, which may well be kind of independent of what we've just been discussing in the last seven minutes? Yes. I'd like to start with 1J because I think 1J is actually the cleanest, clearest, easiest path for resolving this case because Google's art, as the board correctly found, has, as the board said at Appendix 29, quote, nothing to do with the requirements of the claim. 1J is the claim element that requires polling a website and decreasing its rank if its response time is longer relative to another website's response time. And the feature of the Burroughs reference that Google relies on here – and to be clear, Google relies entirely on Burroughs for the functionality of comparative response times – is this notion that Burroughs will optimize its index at the query stage. So it will keep a log of how much processing it takes to search for certain words. Burroughs – To search what, the index? To search the index for words. I thought that was the heart of the distinction. Right. One is about response time of the index and the other is about response time of the website. Those are different things. Exactly, Your Honor. Okay. Exactly. And trying to speed up searching Burroughs' word index, which is just Burroughs goes out and crawls and documents what's there. It's just a typical web crawling search engine. Speeding up searching within that word location index, the board found, as a matter of fact, has nothing to do with comparing response times of websites or resources on an external network. They are just not related. The board made a credibility finding, discrediting Google's expert on this point, made a credibility finding, crediting Paris' expert, resolved this is a matter of fact, and that we would submit is an independently dispositive factual finding, clearly supported by substantial evidence, including Google's references themselves, and that is an easy way to take care of this case. Turning, again, I mean, similar points on H and I, the functionality in Burroughs that Google points to for teaching decreasing a rank is this add, delete operation in Burroughs. That's the only feature of Burroughs. We have a situation in which the board's construction of unexpected results is not consistent with the implied construction that the petition gave it, a construction which you didn't dispute. So why shouldn't the board have given the petitioner an opportunity to address argument and evidence to that new construction? I disagree that Google urged a implicit construction. I think both sides agree. Google urged implicit construction. Okay. I think the way the argument was framed before the board and the way we addressed it and the way the board addressed it was this term, unexpected response, gets its plain and ordinary meaning, does or does not, the feature that Google points to, teach it, the board held no. I think that is just a garden variety application of a claim term's plain and ordinary meaning. The board correctly held that Burroughs has no expectation. It has nothing comparable to this content descriptor file that the 705 patent claims. It just goes out and crawls the web and documents what's there. It doesn't expect anything. And as Judge Taranto pointed out, the 705, not every modification for the 705 is unexpected. Modifications that Burroughs would require an add-delete operation for to update its index would not be unexpected. I agree that the implied plain construction was wrong and that either the board's construction or some other construction was right. But shouldn't they have an opportunity to address that? Again, Your Honor, I think it's just a plain and ordinary. You didn't raise the patent owner response. You didn't raise this construction issue, right? Again, I don't think it's a construction issue. I agree with Your Honor that the patent owner response focused on the decreasing rent piece. But I think this is a garden variety applying plain and ordinary meaning. The art doesn't teach it. So that's how I believe that issue should be resolved. Do I understand correctly? You started with 1J on the premise that the board was right about that. And that one doesn't, I think, even arguably involve this claim construction process issue that 1I at least presents. Yes, Your Honor. It doesn't involve claim construction. It doesn't involve isolating references. It doesn't involve any of the arguments that Google's raised. But in my last minute, let me touch really quick on this notion of isolating Burroughs from Kovacs. That is simply incorrect. The board was very clear about what it was doing with element 1C. The board said in Appendix 21, footnote 6, we are deciding the Kovacs-Burroughs combination based on H, I, and J. We don't need to resolve the other disputes. So the board did not make an explicit determination one way or the other on 1C. It made the determination that because these other elements are dispositive, we don't need to. As to H, I, J, the board explicitly considered and rejected as a matter of fact and credibility, the notion that combining Burroughs' teachings into Kovacs would, quote, change ranking or placement in Kovacs' destination tree. That's Appendix 23, 1H, on 1I, similar finding as applied to Kovacs. That's a quote. This teaching of Burroughs would not, quote, result in modified entries moving to lower-ranked positions in Kovacs' hierarchy. That's Appendix 26, 27, explicitly considering the alleged combination as presented by the board. Okay. Thank you. Mr. Abel. Thank you, Your Honors. I want to start where my colleagues left off, footnote 6. I think it's very clear what the board did. They did not consider the Kovacs-based disclosures and disputes, and they needed to. If you look at the claim, all of the elements leading up to element 1H define the system, and they define the system based on Kovacs. They define the content descriptor file. They define how the system works. They define the destination tree. They define the ranking. Then the later elements, H through J, modify that ranking. What is being modified is the system that is defined by Kovacs. It's clear that the board did not consider those teachings in the combination, and that is a problem. It's a problem because what the board did was picked apart the Burroughs-specific implementation. They saw these deleted entries. They saw the modifications. They saw these processing cost entries, and they said, the way that Burroughs does this does not satisfy elements H, I, and J. That is not responsive to the obviousness combination. The obviousness combination is Burroughs' teachings implemented in Kovacs, and they just did not address it. They said, for example, at Appendix 23, we don't agree, but they didn't explain. They didn't explain anything about Kovacs. They didn't explain anything about the expert testimony other than disregarding it. The problem with their statements about Burroughs and about the expert testimony is that it's not based on substantial evidence. They said that the expert did not provide support for his opinions, and that is just not true. They did not cite the reply. They did not cite any paragraph other than, for example, here on Appendix 23, Paragraph 172. But there's substantial analysis and evidence that goes directly to this combination, and the board just didn't address it. And one final point I want to get at is that the board stated that as to Element J, the disclosure has nothing to do with time. That has nothing to do with time, the time it takes to access a website. That is just not consistent with the evidence. We have expert testimony from their own expert where their expert admitted at Appendix 2235, the way that Burroughs is using the term cost is in terms of processing time or processing cycles, not monetary costs, right? That's correct. I'm sorry, but is that processing time to search the index or to search the web? So it's processing time in Burroughs, in Burroughs' specific implementation to search the index. But as applied to Kovacs, when we're talking about the Kovacs destination tree and minimizing processing cycles, what Kovacs is talking about is searching the Internet. And so as applied to Kovacs, we have a disclosure specifically about minimizing processing time and altering results based on processing time that needs to be viewed in the context of Kovacs' disclosure. And that just wasn't done here. And that is a reversible error because it is error to view the references in isolation like this. And that's exactly what the board did. Okay. Thank you. Thank you, Ron. Thank you.