Oral argument — Oracle America, Inc. v. Hewlett Packard Enterprise Co.

It is Gar, Your Honor. Thank you very much. Thank you, Your Honors, and may it please the Court. With the Court's permission, I'd like to reserve three minutes of my time for rebuttal. Watch the clock, okay? Yes, Your Honor. Your Honors, in this case, Oracle seeks to hold Hewlett Packard Enterprise responsible for its own role in the rampant theft of Oracle's copyrighted software in an effort to siphon off tens of millions of dollars of Oracle's valuable support business. In failing to find that the record before this Court established a single triable issue, the District Court erred in three overriding respects. First, the District Court fundamentally misapplied subtle summary judgment principles. The District Court's refusal to infer that installed on date could mean installed on date epitomizes this error, but this error is seen throughout the Court's opinion, including in the Court's disposition of the fact-intensive statute of limitations issues. Now, you can handle this however you want, but to me, this case is divided into two parts. One has to do with the statute of limitations and when that begins and ends, and then the second part is whether, if you get past that, whether a cause of action has been alleged. With respect to the statute of limitations issue, I'm really puzzled by Oracle's position. How can you plausibly argue that the document of fraudulent concealment applies here when you had sufficient to trigger a duty to reasonably investigate? Your Honor, let me answer that. Let me first make clear, though, that on the statute of limitations, it's undisputed, I believe, that we can go back three years on the copyright claims and two years on the state law claims. The only question is whether we can go back further than that because the statute of limitations was told. So there are claims... I get that, and again, I take your point that once you get past the statute of limitations, you've got a separate case and you're arguing the Court was wrong there, too, but I'm right now focusing on those cases, those instances that occurred prior to when the Court said the statute of limitations. Right, and I think fundamentally... How does fraudulent concealment help you here? Well, it helps us because it goes to our alleged constructive knowledge, Your Honor, and I think it's critical in working through this issue to appreciate that they have to show that we had knowledge of HP's role, actual or constructive, not simply knowledge of Tarek's. And what happened here, Your Honor, is what the documents show, and this is an extremely fact-intensive, disputed issue, but what the documents show is that at the very most, Oracle had suspicion that Tarek's was involved in perhaps downloading and installing Oracle's copyrighted patches. At that point, Oracle investigated Tarek's by doing what made perfect sense, looking at its MOS website, which is what it had access to, and searching as to whether or not Tarek's had made unlawful downloads there. And it didn't uncover anything because Tarek's had gone to such extraordinary lengths to uncover... With respect, Oracle's a very sophisticated company. You investigated this single entity, didn't find anything. On the other hand, you had the right, under your documentation, to audit. Apparently, it was just a Trojan horse because you didn't do that. Why didn't you check this out? Certainly, there are other places in America, and I'm not going to take up your time to cite them, where I would think that the legal department of Oracle would have red flags up everywhere about this concern. Where am I wrong about that? Your Honor, first of all, Oracle did confront Tarek's, and it confronted HPA. They denied their wrongdoing. Second of all, the notion that the customer... Is not fraudulent concealment sufficient to toll the statute? Yes, we do, Your Honor. But there was more than that. There was the additional actions that they took to conceal their fraudulent activity, using prepaid credit cards, fictitious names, and the like. But on the customer audit, I think this is really important. The fundamental question for the court is whether it was reasonable, whether Oracle's investigation was reasonable. Oracle has never, in its history, undertaken that kind of audit. It would have involved, Your Honor, going to customers and actually accessing their servers, thousands of servers located across the country. HP itself, Your Honor, resisted discovery into its provision of services, on the ground that this was unduly burdensome and time-consuming. You can see that at page 1364-65 of the record in this case. So for HP to turn around and say that we should have simply engaged in this unprecedented audit of our customers is unfounded. So you're saying that you never, never audited any of your customers? Not, Your Honor, not in the manner that they're suggesting here, which is to go after the individual servers, hundreds and thousands. Do you agree that your documentation permitted you to do that? It did, Your Honor. Of course, there would be a serious cost to that as well, in terms of the notion that we would essentially have to access individual servers. And again, it was completely burdensome. And this all goes to the reasonableness of the investigation. If you look at cases like the recent Adobe case at this site, this court's decision, Gorman v. Waldorf, talks about how these are intensely factual issues for a jury. Can I ask a question about what you did do? Yes, Your Honor. Put aside what somebody thinks you should have done. Am I wrong in thinking that the investigation was limited to finding one account that said at Tetrix, Your Honor? That's correct. What was the investigation that your client did do, assuming they were put on constructive notice? Right. Well, we did go out and we reached out to Tetrix or NHB, and we asked them whether or not they were engaged and they were respecting our contracts. They denied that. We then did what was within our capability of doing, which was to look at our MOS website, which is the gateway through which customers would get patches. And because we suspected Tetrix of wrongdoing, we searched for evidence that Tetrix was unlawfully downloading patches. Right. And I'm familiar with the record. I'm trying to figure out whether anything else was done. Your Honor, that was the focus of it. And again, this is at a time where Tetrix was denying it. Tetrix was taking extraordinary steps to conceal its wrongdoing using fictitious names, prepaid credit cards. What we learned about HP's role in the course of our litigation against Tetrix, we sued Tetrix timely in 2013, and it was during discovery and that litigation in 2014 that we learned through depositions of Comcast that HP itself was involved in actively facilitating Tetrix's scheme. And then we entered into the tolling agreement and we brought suit. Your Honor, if I could shift to the merits of the claims, which are presented as to the timely periods. On indirect infringement, our position is simply that there was ample evidence of a tribal issue that HP had engaged in direct installations. First, there's the evidence of Mr. Christian Hicks laid out at ER 13624, where he explains that the spreadsheets produced by HP show that 108 instances patches were installed on HP's customer servers that were released after the date that the underlying servers came off contract. This is undeniable evidence. Does it have to be installed? What about if it's just uploaded, copied, whatever? Isn't that sufficient for a copyright violation? It absolutely is, Your Honor. The Ninth Circuit's decision design data versus Unigate makes that clear. And that, I mean, I think that the downloads are the focus of our indirect infringement claim. And there you have admitted thousands of copies of downloads. And the defense is that the licenses somehow allowed those downloads. And they absolutely didn't. You can look at the support contract. Did Judge Tucker rule on that defense? In other words, is the issue of whether the licenses covered these downloads in front of us or not? Well, HP thinks he did. We don't think he did. There's a single sentence in the background. I'm not asking you what people think. I'm asking you whether Judge Tiger ruled on that issue. Do we have to decide in this case whether there was a license, or is that an issue that was left open? I think you could remand for Judge Tiger to address that, because it is extraordinary that he didn't deal with the terms of the licenses. And I would point you to pages 142, 148, and 130, which is the support contract, which explicitly says the customers may not download patches for use on unsupported servers. And that's exactly what Terex was doing. Terex's own CEO, Mr. Appleby, admitted that at pages 1285 to 1286. Terex's Director of Technical Services, Mr. Joyce, admitted that at pages 702, 703. That clearly creates a tribal issue that they were exceeding the scope of the licenses, and so there was a sufficient basis on the indirect infringement claims, Your Honor. And, again, this Court's decision and design data makes clear that downloads are copies. Under this Court's case law, a single copy is a sufficient basis to go to a jury on a liability issue. And then the question would be damages that the jury. So that resolves with the indirect infringement claim. Just jumping back to the direct infringement claim briefly, there's overwhelming evidence creating a tribal issue on that claim as well. First, we have the 188 patches that Mr. Hicks explained with forensic evidence as to particular servers and particular patches were installed by HP on its customer servers. That's 1324. We have the admissions of HP's own employees, Mr. Chris Elfers at 553, 554, Mr. Bill Hall at pages 1182, that HP was installing patches on Symantec servers that were not on contract. Again, this creates a tribal issue. This is a case on summary judgment, and the way that the Court avoided that was to shift the burden to draw reasonable inferences in favor of HP, not Oracle, and to resolve disputed facts. We asked this Court to reverse the grant of summary judgment on both all the claims before us, and we also have a set of state law claims, Your Honor. The fraudulent conduct in this case is tailor-made for the state law interference claims before you. HPE's own employees admitted that the scheme truly took revenue right out of Oracle's pockets and put it into HP's. That's at page 1203 of the record, Your Honor, and it did it by encouraging customers to violate the terms of their support contracts, to go along with a scheme in which you would put a single server on a support contract and then download hundreds and thousands. Mr. Hicks made clear that there were thousands of downloads at page 1325 and allowed customers to access those patches or install those patches on unsupported servers. This was clearly a violation of the copyright laws. At the very least, it was a tribal issue that precluded summary judgment, Your Honors, and I just want to stress briefly the business model here, Your Honor. Oracle's support business is extremely valuable. Oracle makes clear in its licenses and agreements that customers can only access support patches for servers for which fees are paid. So a scheme in which Terax, facilitated by HPE, is downloading thousands of patches for use on unsupported servers goes to the heart of Oracle's business models, and as Mr. Pampanella, our damages expert, explained, resulted in tens of millions of dollars of lost profits, Your Honors. So the state law claims, too, should be allowed to proceed. Your Honor, I would like to return... Do you want to return for the balance of your time? I will, Your Honor. Thank you very much. Very well. Okay, Mr. Perry, you need to unmute, and you're on. Thank you, Your Honors, and may it please the Court. Judge Hurwitz, you asked my friend, Mr. Garr, whether Judge Teigar had construed the licenses. The answer is absolutely yes. In fact, the parties... Let me... I want to tell you what's troubling me about that. Yes, Your Honor. The only place I can find a construction of the licenses is on a very brief portion of his order, ER 338 and 39, and he just basically seems to say, in an offhand manner, it was okay under the licenses. Nobody really briefed that issue to us very well, but I must say that if I had to construe the licenses, I would come at it the way your friend did. The licenses seemed to only allow single downloads for use on service on... So do you really want us to ruin the license issue, or do you want us to send it back to him to look at it? Well, let me answer it this way, Judge Hurwitz. The parties filed cross motions for summary of judgment on this. There were no disputed issues of fact. It was a legal issue of the construction of the licenses. HPE laid out in great detail its construction of the licenses, both in its own motion and in opposition to Oracle's motion. Oracle's position was only that Judge Greenwald, Magistrate Judge Greenwald, had rejected the perpetual license argument in the Tarrac case. That was the only argument... Okay, but put aside who's right or wrong. I'm trying to figure... You really... Show me where Judge Tiger ruled on this issue, other than those pages that I identified. Let's start at ER 324 in the order where Judge Tiger identifies the two licenses at issue, the BCL and the SLA, and it explains that both of them grant to the customer a non-exclusive right to use the software that is provided by Sun and later Oracle. Then on the next page, ER 325 is the first of the three license constructions at issue in this case, none of which Oracle challenged on appeal, and it appears in the middle of the page. An Oracle support contract, the license, grants the customer or an agent acting on the customer's behalf, such as a third-party maintainer, that would be my client, Your Honor, HPE, a license to download, deliver, and install Solaris patches. And then it goes on to explain downloading and delivery and installation, all as explained in the license. And not only in the license, Your Honor, but this was Oracle's testimony. And I should point the court to SER 1102, and this goes to Judge Royal's question about whether or not downloading is copy. Here's a direct quote from Oracle's testimony. Quote, during the duration of a contract, they have access to all patches, and they would have the ability to download them. At the expiration of the contract, they have access, but if they had downloaded patches during the contract, they can install them at a future date. And that's exactly what Judge Teigar rules at the end of that passage on ER 325, lines 17 to 19. And, in fact, that citation is to the document I just read you, which is an internal Oracle contract analysis. And the second… Can I switch you from license for a minute? You know there was a spreadsheet that is a matter of some importance in the case, and I'm wondering what your best argument is why a reasonable jury couldn't rely on the spreadsheet analysis from Oracle's expert to conclude that there was at least one instance of direct copyright infringement by HPE as directly caused by HPE. For two answers, Your Honor. Well, three. One is Oracle didn't make any disputed issue of fact as to the spreadsheet below. It actually moved for summary judgment in its favor on the spreadsheet, so there's that. Second, there is no evidence that HPE did the installations. The testimony is undisputed, and this is SER 564, that we don't know who made the installation. It could have been a customer. It could have been Oracle. It could have been a different third-party maintainer, or it could have been HPE. And there is no evidence as to any particular server who it was. But your client was under contract to some of these people at the time. It may have been Judge Royal who installed it too, but isn't it a reasonable inference that the person who was under contract to the client installed it? Not at all, Your Honor. Most of these contracts, and there's the evidence of the SunGuard contract is in the record, for example, where HPE provided only hardware break-fix support. HPE didn't actually provide the software support, so this spreadsheet shows everybody that the 35 companies that HPE had a support relationship with. Those are not software support companies. Those are the 35 that were analyzed for this case. So there's no evidence that for any of them HPE was providing installation support. In fact, often the customers take care of that themselves, and the standard license between HPE and its customers make clear that both the installations and the licensing remain the customer's obligation, not HPE's. So in fact, the inference that there's a… So what was your client downloading them for then? HPE was not downloading them, Your Honor. There's no evidence other than as to one client, Symantec, that HPE made a single download in this case other than in its other capacity as a marketing partner with Oracle. Most of the thousands of downloads that HPE did were Oracle licensed pursuant to a deal that HPE and Oracle have together to jointly serve customers. Yes, HPE does that, but Oracle knows every one of those because we pay Oracle back pursuant to a different contract. Those aren't in this case, but that's why Oracle keeps trying to mix up which customer, which contract, which client, which is why, by the way, Judge Tiger construed these licenses to require a server-by-server analysis because unless you get an improper installation on a particular server outside the contract date, you don't have copyright liability under these licenses. That's why the license construction is so important. To go back to your question, ER 338 to 339, that's exactly where Judge Tiger construed the licenses, and those pages of his order aren't even cited by Oracle in either of its briefs on appeal. They don't take issue with those. They don't mention that he made those rulings. They don't challenge them as unlawful. They just pretend that they don't exist. If we do not find the two-page reference to Judge Tiger's comments about the license to, in quotes, construe the licenses, and if we have questions about that, and if we have questions about whether there's one or more alleged download, would that not require us to send that back to the district court for further discovery and perhaps a trial? Well, Your Honor, it's contract interpretation, so it's a legal determination. I mean, we got here because Oracle never made any of these arguments until it got to this court. In fact, in its reply brief is where it makes the arguments for the first time. Of course, the court could send it back if it wants more information, but we would submit that Judge Tiger did a very careful approach. We also cite in our brief, in the red brief at page 35, we cite our extensive, you know, we devoted 12 pages of the summary judgment briefing to this. Oracle devoted two to the license arguments, explaining with many, many factual citations to one recorded record. And again, ER-1102, I think, answers all the license questions in this case, and that's an Oracle document giving Oracle's explanations. I'm sorry. Part of the problem with Zoom is that I thought you'd finished and you hadn't, so please finish what you were saying. Well, when I look at 338, we agree that it's 338 and 339 where Judge Tiger seems to rule on the license issue, but as I read his ruling, it's the same ruling that underlies the rest of his summary judgment order, which is that you haven't, nobody showed a particular installation of a particular patch on a particular server. Assuming, as Judge Smith said, that there was evidence enough from which a jury could infer that, is he really saying the license would cover that? Your Honor, we certainly, HPE certainly takes the position we read the order as saying, if Oracle could prove that a patch was installed on a server after the contract had expired and that that patch was not downloaded during the term of the contract, that is the middle of page 338, they could satisfy the infringement standard that the judge drew from these licenses. I agree. I agree. That's the dispute between you, but when I read Judge Tiger's licensing discussion on 338 and 339, it seems subsidiary to the premise that there was no such proof. He doesn't really say if there were such proof, there would be a license to do this. I don't think it is, Your Honor, for this reason. You have to get through the downloads and the deliveries to get to installation, and the importance of 338 to 339 is the court explains that as long as you have an active contract, the downloads and deliveries are licensed, which leaves only installations under the theory on which this case proceeded. And then as to installations, it's not enough to have installations in the air or in the abstract. You have to get to particular installations by the server because it requires a server-by-server analysis. But I take it you agree that if there were sufficient evidence to show installation on a specific server, then it would be unlicensed. If it were installed on a server that was not under contract and it was not downloaded at a time that that server was under contract, it would be unlicensed. That's exactly right. That's what Oracle had to show, and that's what they didn't show, which really brings us to what was brought up on appeal. Oracle didn't challenge that standard on appeal. That's our position. They accepted that standard. Then the only question is summary judgment, and there's three customer categories. As to the joint support customers, Oracle has abandoned its installation argument. It makes no argument with respect to installations whatsoever, so they're out of the picture under that standard. As to Symantec, Oracle proved zero protectable patches, protectable under Section 102 of the Copyright Act. Judge Tygar ruled on that. Oracle didn't challenge that appeal. Game over. We don't even have to get to installations. That leaves the direct support customers. That's the spreadsheet, right? There's 35 customers with 5,000-plus servers. All have a license from Oracle. Most of them have hundreds or thousands of servers under contract with Oracle. Some of them, the old ones, have gone off contract. Oracle's obligation was to show that installations of patches were made by HPE on an unentitled server. It took immense discovery. We had 2 million pages of documents, terabytes of data, 900 requests for admission and interrogatories, 40 depositions. They subpoenaed all these customers. They subpoenaed four of the customers, Your Honor. They didn't ask them when were the patches installed. They didn't ask them who installed the patches. They took their depositions. They didn't ask those questions. They didn't exercise their audit rights. They never checked the server logs. SER 1470 is our experts report, which explains what Oracle could have found looking at the software that it claims to own and have exclusive rights in. It could have answered all these questions. This case was litigated to the hilt. Millions and millions of dollars were spent over years and years, and they couldn't find a single server on which HPE installed a single patch outside of the contract. And that's the end of the case. That's Summary Judgment 101. You know, and as to the installed on day, Judge Smith, I didn't quite finish answering your question as to what a jury can conclude. Not only do they have the problem of who installed the patch, if any patch was installed, which they can't satisfy. And, by the way, their expert conceded that he doesn't know who at SER 367 to 68. So that would be one answer because the Copyright Act requires showing that the defendant did the copying. But even if they could show that HPE did it, we don't know when it was done. That is, whether it was under contract, because the installed on date doesn't necessarily mean the installed on date. And this isn't an ambiguity in the sense that summary judgment law speaks to. This is simply a lack of information. Counsel, help me. I thought that the spreadsheet did have some dates on it. I gather there were three potential constructions of what was there, but one of them allegedly was the installation date. Is that incorrect? That is correct, Your Honor. One of the three is the installation date. The other pertinent one is the date the file was last accessed, which would be meaningless for our purposes. There's no way to distinguish a particular server between the two. And that gets me back to my question. If the license issue were resolved against HPE, and the issue were presented to a jury, and the jury had three potential ways of construing the spreadsheet, and it concluded that the information necessary to show a violation was there, would that not be enough to show a direct violation? It would not be, Your Honor, both because they don't know who did the installation. They don't know what was installed. Mr. Jensen testified that it could be a Solaris patch or it could be some other patch. This system picks up all kinds of patches. And they don't know when it was installed because that date is indeterminate. That would not be enough. In other words, if the jury reached a verdict in Oracle's favor, the court would have to throw it out, and this court would have to throw it out because it wouldn't be legally sufficient evidence. I think we may have lost you there. See if our tech people can... I don't know whether my colleagues are having the same problem. I can't... Yes, I have the same problem. Okay. He was at about 105 on the clock, so we can... Yeah, something like that. Yeah. Sam, whoever's handling the tech thing, can you hear us? We've lost Mr. Perry. Yes, Judge. It looks like he may have had a network issue. I can try to call him and see if he's able to reconnect quickly. Would you do that? And, Mr. Garvey, just kind of stand by. We'll work with this. There he is. That's Mr. Perry. Mr. Perry's moving back there. Okay. I froze for a second, I think. Go back to what you had in time before. Go ahead, Mr. Perry. Thank you, Your Honor. I don't know exactly where I froze up, but my point was there was 188 servers out of, by the way, thousands and thousands and thousands. So, that's the maximum universe is this 188. Oracle never did the next point of looking on those servers to find out what was on them, asking the customers who they subpoenaed, who they deposed, who did any installations. This is just a failure of proof. They got some evidence, but it wasn't enough to go to the jury, and they didn't follow up. That's, you know, the reason they lost the case and the reason it should be infirmed. The last point I'd like to make before my time runs out is on this burden-shifting point. This is a license case. Judge Tiger was not construing the Copyright Act for every software download case, every patching case. He construed these licenses, and the way this burden-shifting works is once the license has been interposed, the judge construes the license, and then the copyright holder has the burden of showing that the activity was outside the scope of the license. The scope of the license is defined at ER 338 to 339, the critical passages, Judge Hurwitz. That is the scope of the license as accepted by Judge Tiger. Again, I have to keep beating on this. Oracle does not challenge that on appeal. It did not bring a challenge to that license construction, and under that construction, Oracle had the burden of showing a particular server, and it just didn't. Is that the law? I know that's the law in the Second Circuit. Have we ever said that? Your Honor, the court, the Bourne case, which both parties cited in the Second Circuit, I think is the clearest example of it. I think this court has said it in the Cadell case that's cited in the bottom-side amicus brief. It's a Ninth Circuit case on the license defenses, but it's also a logical way to do it because the scope of the authorization, once defined by the court, it remains always the copyright holder's ultimate burden of persuasion to show copying. This court has said unlicensed copying isn't part of the prima facie case, but after the license has been interposed and interpreted, it becomes part of the copyright holder's case on a going-forward basis. Okay. Well, Mr. Perry, thank you for your argument. Let me ask my colleague whether either has additional questions for Mr. Perry. Well, Mr. Perry, in looking through some of the records, and in the interest of time, I won't read much of this, but Lynn Farland's deposition, various emails, so forth and so on, that come from your company, it looks like there are people in your company that are concerned that Hewitt Packard is doing something wrong with the copyright related to Solaris, especially related to the package, the patches, and I don't understand how that fits into this. Terex is playing the word game. Oracle will come after us because we're the larger of the two. Target, how do we get some OS support? Clear vision issues and avoid $10 million lawsuits. We hope we can get the package, and this is talking about the Solaris package. We hope we can get the package from Terex, which means you're not getting it. So what does all that mean in the context of this case? Very briefly, Your Honor, I think it shows HPE being scrupulously careful about intellectual property rights. It has many copyrights and patents of its own. It had some concerns about Terex, sure. What it didn't have was concerns about its own processes. There is no evidence, not a shred of evidence, that HPE ever downloaded an improper patch or installed one. There were some questions about Terex. I think that's fair. Terex was a subcontractor. Terex was a separate company, but Terex kept HPE in the dark as well. Terex wouldn't let HPE see these presentations. Terex warranted to HPE that it had a license with Oracle, but it turned out it didn't. And most importantly, once all this came to light, HPE has severed its relationship with Terex. Oracle has gone after Terex and secured the judgment that was satisfactory to Oracle on that basis. This happens in the contracting world. I think none of it rises to the level of knowledge of copyright infringement because what those emails and so forth show, Your Honor, is that there were questions about what Terex was doing, but they also show that HPE wasn't doing anything to invade on Oracle's copyrights, as we have not, which is why this case should end now. Well, didn't Hewlett-Packard promote the Clear Vision program, the Terex office? No, Your Honor. The Clear Vision was Terex's name for its presentation that it required NDAs to even review, refused to allow HPE to see. We only, one of our employees, Ms. Farland, had a late partial preview of it late in this process, but that was the Terex side, and that was kept secret from HPE as well. Other questions? If not, thank you very much, Mr. Perry. Now we'll hear again from Mr. Garr with your rebuttal. You've got a little over two minutes. Thank you, Your Honors. There's a first on the licenses. There's nothing in the opinion in which the judge purported to engage with the license terms. This court can and should reverse on that basis, and it should adopt the interpretation that Judge Grewald adopted in the Terex case because it's clear on the faces of the licenses, and again, I would point you to 142, 148, and 130, that customers are not allowed to download patches for what Terex was doing. Those admissions are ample in this record. The notion that Oracle didn't raise the license argument is, frankly, preposterous. In our opening brief, we addressed it at 39 to 42. It's addressed in our reply brief. It was one of our central submissions, and to be clear, it was the downloads themselves that violate the terms of the licenses with respect to indirect infringement. We don't have to show the installations. Downloads constitute on impermissible copies of outside the license. The design data case shows that, and here there were thousands of downloads for unsupported servers, which violates the terms of the licenses. Second, on the installed on date, HPE's own counsel admitted below that a reasonable juror could find that installed on date meant the date that patches were installed on servers. That's at SCR 38. And HPE is really acting quite extraordinarily in this respect. This information, the spreadsheets, were produced in response to requests for all information of its data, the provision of services, including patches to its customers, and that information was ordered because HPE dug its heels on this, and that's the order is at page 227. For HPE to now turn around and say that we should disregard its own installed on dates spreadsheets is, frankly, remarkable. Now, with respect to the evidence on the direct installations, Your Honor, Mr. Hicks showed 108 patches that were released after the date that the underlying servers came off contract. That's true regardless of the installed on date field, and that's at page 1324. Mr. Elfers, an HPE employee, admitted that HPE was installing patches on unsupported servers. That's at 553, 554. Now, Judge Royal, you asked the question of the significance of those documents. The significance is, is it shows that HPE knew the tariffs was involved in an illegal scheme, and I would point you to pages 929 to 30, 931, and 930, 931, which show that it knew the tariffs was involved in illegal activity. Then HPE facilitated that. It did participate in the Clear Vision meetings. Look at pages 585 and 497. It helped customers go to tariffs to find patches. It was directly involved in the scheme. Oracle is seeking to hold HPE accountable under the copyright laws and the state laws for interference with business relations. It's at the very least entitled to present that claim to a jury. We ask the court to reverse the grant of summary judgment for HPE and return this case for a jury trial. Thank you very much. Let me ask, do either of my colleagues have additional questions for Mr. Garr? No. If not, we thank you very much to both counsel for the argument. Well argued on your part. The case just argued is submitted.