[00:00:00] Speaker 02: 122-1600 Mirror Worlds versus Meta Platforms. [00:00:06] Speaker 02: Mr. Liddell. [00:00:09] Speaker 00: Thank you, Your Honor. [00:00:12] Speaker 00: Good morning, and may it please the Court, Brian Liddell, for Mirror Worlds Technologies. [00:00:19] Speaker 00: We're here for the second time in this specific case because the District Court [00:00:25] Speaker 00: twice made the same error. [00:00:27] Speaker 00: It misapplied the standards for summary judgment mandated by the Supreme Court, by this court, by Rule 56, and frankly by the Seventh Amendment. [00:00:38] Speaker 00: That it improperly waived the evidence and it failed to draw inferences in favor of the non-movement and instead [00:00:47] Speaker 00: expressly drew inferences and construed the evidence in the light provided by the movement and the movement's witnesses and credited their evidence without, in many instances, even considering evidence provided by mirror worlds and opposition. [00:01:01] Speaker 04: Can you tell me what part of your argument is independent of your challenge to the claim construction on data unit, independent of the collection challenge and of the glance [00:01:17] Speaker 00: Absolutely, your honor. [00:01:18] Speaker 00: So there are two accused systems. [00:01:20] Speaker 00: There's what's referred to in the papers as the timeline system and the multi-feed system. [00:01:26] Speaker 02: Those are all the back-end systems. [00:01:27] Speaker 00: Those are both back-end systems, correct. [00:01:30] Speaker 00: So there's the timeline back-end system and the multi-feed back-end system. [00:01:33] Speaker 00: As to the timeline back-end system, our argument about summary judgment is there are also claim construction issues, of course, but [00:01:46] Speaker 00: the summary judgment issues are sufficient to reverse the district court's decision on that issue. [00:01:51] Speaker 04: But do I understand correctly that on News Feed, your only argument is dependent on your data unit claim construction? [00:02:00] Speaker 00: That's correct, Your Honor. [00:02:01] Speaker 04: As to the News Feed... But for Timeline, you say that the three pieces of information that [00:02:10] Speaker 04: I think it's three, but you'll correct me. [00:02:12] Speaker 04: The district court said, Facebook said, these come into the system, but they don't go to timeline database. [00:02:19] Speaker 04: You say that was a factual, that raised a triumphal factual issue. [00:02:24] Speaker 02: That's correct, Your Honor. [00:02:28] Speaker 02: wrong but is that tied to the mainstream claim construction or is that? [00:02:33] Speaker 00: It is not because as to the timeline system there's a factual dispute as to the factual question of whether these pieces of information are in fact received at all by the back end system. [00:02:45] Speaker 00: That was a disputed question of fact. [00:02:49] Speaker 00: There's evidence both ways. [00:02:50] Speaker 00: The district court we submit improperly basically chose to credit Facebook's evidence and not [00:02:58] Speaker 00: addressed properly. [00:02:59] Speaker 04: And where does your queries don't count contention fit into this? [00:03:07] Speaker 00: Sure. [00:03:08] Speaker 00: So I would characterize that as an independent reason that would support reversal, certainly as a timeline where even if you were to credit Facebook's evidence, which, as I said, was disputed, [00:03:23] Speaker 00: all of that evidence that they point to is query evidence. [00:03:26] Speaker 00: And so that would be an independent basis for reversal as to the timeline. [00:03:30] Speaker 04: Is that a claim construction point? [00:03:32] Speaker 03: The queries is part of the data unit construction, right? [00:03:35] Speaker 00: That's correct. [00:03:35] Speaker 03: It's part of the data unit construction. [00:03:37] Speaker 03: So if, again, if we were to, just for the Sega argument, agree with the dispersed construction of data unit, then the queries thing doesn't save you even on timeline. [00:03:48] Speaker 03: Is that right? [00:03:51] Speaker 00: Let me see if I can clarify that a little bit. [00:03:56] Speaker 00: Our position would be that independent of the construction of data unit, queries and interest of the user, there is a [00:04:05] Speaker 00: the summary judgment was improper as to the timeline system. [00:04:10] Speaker 00: I would say that if you were to agree with the district court on the data unit construction, that would not require reversal of the multi-feet system non-infringement finding, but the timeline system. [00:04:23] Speaker 03: you could still prevail and when reversal on timeline isn't because of queries, it's because of those other types of information that you say. [00:04:32] Speaker 03: That's correct. [00:04:33] Speaker 00: It's because of other factual disputes about whether that information and whether it's really received by the timeline backend system, the aggregator, as Facebook intends. [00:04:42] Speaker 00: That was an evidentiary and factual dispute. [00:04:45] Speaker 04: And so my recollection, and I don't, I think, have all the details in my mind at the moment, is that when I looked at the evidentiary dispute about these very specific things, so I'm now putting aside all the claim construction and plans for you and all of that, that Facebook had evidence that said, here are three things [00:05:16] Speaker 04: that come into the aggregator that don't get put into the timeline database, the chronologically organized thing. [00:05:29] Speaker 04: And what you had in response is, I guess I have two categories in my mind. [00:05:35] Speaker 04: One is some internal Facebook documents that are [00:05:42] Speaker 04: not focused on those individual things, but saying, here's what generally the timeline database does. [00:05:50] Speaker 04: And you have a second thing, which is some paragraphs in your expert's declaration that gives some references to pieces of software without those places exactly saying, [00:06:06] Speaker 04: those establish that these items either don't go to the aggregator or do go into the timeline database. [00:06:14] Speaker 04: And then there's a kind of conclusory sentence that says, everything [00:06:19] Speaker 04: that comes into the aggregator, goes into the timeline database. [00:06:22] Speaker 04: And so it feels like maybe there's a mismatch between the specificity of Facebook's evidence and the generality of yours. [00:06:30] Speaker 00: Sure. [00:06:30] Speaker 00: And I think I guess I would have two responses to that, Your Honor. [00:06:34] Speaker 00: The first is that the evidence that Mirror Worlds presented, even if we accept the premise about that, what you've characterized as a sort of disconnect on the specificity, [00:06:45] Speaker 00: I don't think that the summary judgment standard requires that kind of specificity that you're talking about when you talk about what the evidence is. [00:06:55] Speaker 00: So the evidence, we presented a number of pieces of evidence that all the information that comes to the backend system is stored in the timeline database. [00:07:04] Speaker 00: All can mean all. [00:07:05] Speaker 00: and it's appropriate for a jury to credit that and a jury gets to weigh the question of should we interpret that and believe what these documents that were created before litigation by Facebook's engineers about how the system actually works [00:07:21] Speaker 00: Should we as a jury credit that? [00:07:23] Speaker 00: Or should we take this testimony which seems inconsistent with those documents and credit that? [00:07:29] Speaker 00: That's a fair dispute for a jury to resolve, but it's not a fair dispute for a court to resolve on summary judgment because you have to draw the inferences about the evidence in mirror world's favor. [00:07:41] Speaker 00: And drawing appropriate inference when the document says all the data goes into the database, [00:07:47] Speaker 00: it's an appropriate inference to say all means all, and that a jury can rely on that. [00:07:53] Speaker 00: And similarly, I would say that the testimony of the expert, Dr. Koskinen from Mirror Worlds, was pretty extensive on everything that comes to the backend system is stored there. [00:08:08] Speaker 00: And I don't think it needs to sort of then [00:08:11] Speaker 00: say, well, everything includes this piece and this piece and this piece that they're later going to point to one summary judgment. [00:08:18] Speaker 04: Why isn't it incumbent in a summary judgment context on him to have done just that? [00:08:26] Speaker 04: The other side goes first. [00:08:27] Speaker 04: They move for summary judgment, and they say there's this requirement that everything that comes into the aggregator go into the system. [00:08:35] Speaker 04: I'm just using aggregator as a substitute for that. [00:08:39] Speaker 04: has to go into the timeline database. [00:08:42] Speaker 04: Here are three very specific things. [00:08:44] Speaker 04: Why isn't it incumbent on your expert to say, I'm now going to talk about these three things, and I will tell you why they do either not come into the aggregator or do go into the timeline database, as opposed to resting on the generalities? [00:09:00] Speaker 00: Well, I think, I don't think he, so this gets to the second part of your question. [00:09:06] Speaker 04: The question, my premise, right? [00:09:07] Speaker 00: Right. [00:09:08] Speaker 00: Which is, [00:09:10] Speaker 00: Did he actually have evidence and was there evidence that was specific? [00:09:14] Speaker 00: So what Facebook talks about in these pieces of information is it talks about, again, the query to the timeline backend system and suggests that these pieces of information are included in it. [00:09:28] Speaker 00: But its own documents rebut that assertion that those are pieces of information included. [00:09:34] Speaker 00: So if we look at, for example, page 14282 of the appendix, [00:09:40] Speaker 00: Unfortunately to yes So this is this is a Facebook presentation about the timeline system, and it's architecture And it is one title front end yes, not that this page refers to the front end But I I think when we get to the content you'll see why it's important, okay, so [00:10:03] Speaker 00: So at the top here, it talks about the front end generating a request. [00:10:08] Speaker 00: Now these are the requests that we're talking about that go to the back end. [00:10:12] Speaker 00: What does it say? [00:10:13] Speaker 00: It says, generates a request for user, start date, and end date. [00:10:17] Speaker 00: So Facebook's own documents say that the information that goes to the back end is essentially just the start date, the end date, and the identity of the user. [00:10:27] Speaker 00: Facebook's witnesses argue, no, there's other information. [00:10:30] Speaker 00: But this says something different and creates a conflict in the evidence about whether what Facebook is asserting is a true assertion of fact. [00:10:41] Speaker 00: Now, they can certainly make that assertion, but it's not something that the district court can simply just choose to accept when there's conflicting evidence. [00:10:51] Speaker 04: I guess maybe you can explain a little bit more about the context of this. [00:10:55] Speaker 04: This is a PowerPoint thing. [00:10:56] Speaker 04: Yes. [00:10:57] Speaker 04: Why would you read this as saying, these are the only requests that can be made, as opposed to, here's an example of what's generated. [00:11:10] Speaker 04: Requests for user start date, end date. [00:11:12] Speaker 04: There could be lots of other things. [00:11:13] Speaker 04: I'm just trying to illustrate these. [00:11:16] Speaker 00: So if we go, I think it's [00:11:20] Speaker 00: Couple pages further to 14285. [00:11:22] Speaker 00: This is the same presentation. [00:11:27] Speaker 00: You'll see that this same request that we talked about just a second ago, this is where it talks about that's the request to the back end. [00:11:36] Speaker 00: And it's the same thing, the start time, the end time, the user ID. [00:11:40] Speaker 00: This is evidence that directly contradicts what Facebook contends about these requests. [00:11:47] Speaker 00: Now, as I said, [00:11:50] Speaker 00: Parties can disagree and have disagreements about how the system works, but it's not the district court's role to resolve those disagreements at summary judgment when there's conflicting evidence. [00:12:00] Speaker 03: I want to make sure I don't misunderstand your answer to one of Judge Toronto's earlier questions. [00:12:04] Speaker 03: He characterized it as Facebook points to maybe three types of information that don't get to the back end. [00:12:12] Speaker 03: And are you acknowledging that your expert fails to show how a reasonable juror could find that those do get to where they need to get? [00:12:25] Speaker 00: I think two things. [00:12:26] Speaker 00: So one, [00:12:28] Speaker 00: I think, yes, that there's a conflict in the evidence insofar as what our experts showed was that to the extent that Facebook is suggesting these things come from the user database or the TAU system, there's no showing that there's any pathway for that connection to exist and that the architecture doesn't support that factually. [00:12:50] Speaker 00: Second, I don't think there's a showing adequately, although this is less something that our expert focused on, that that's something that's not in the timeline database. [00:13:01] Speaker 02: On the first point, was that a point that was made by your expert? [00:13:05] Speaker 00: Yes. [00:13:06] Speaker 00: He talked about the pathways and these very architecture documents in his report. [00:13:11] Speaker 00: So the excerpts that he talked about this, you could find at pages [00:13:16] Speaker 00: 13291 to 292 of the appendix and continuing Basically the whole discussion from 13301 to 310. [00:13:27] Speaker 00: He had a number of lengthy paragraphs discussing these these very points but I think the the piece and part of [00:13:37] Speaker 00: the evidence that we have here in these various presentations, to the extent that Facebook is suggesting that this information is somehow passed from the TAU system to the back end, this court already ruled in the prior appeal. [00:13:50] Speaker 04: And by the back end, you mean to the aggregator? [00:13:52] Speaker 00: Yes, to the timeline aggregator and the timeline back end system. [00:13:56] Speaker 04: Do I have the terminology right? [00:13:58] Speaker 04: The aggregator and the timeline database are both part of the back end system. [00:14:03] Speaker 00: That's correct. [00:14:03] Speaker 04: Yes, Your Honor. [00:14:04] Speaker 04: Do they fully constitute the back end system? [00:14:06] Speaker 00: I would say that they are the primary, obviously there's some connectivity and things of that nature, but yes, they are the sort of real components. [00:14:16] Speaker 04: And everybody agrees Tau is not part of the backend system. [00:14:19] Speaker 04: That's correct, your honor. [00:14:19] Speaker 00: It's just another source of... It's a different database, yes. [00:14:22] Speaker 00: And we presented evidence and the court actually agreed with us that in the prior appeal there was evidence showing that Tau didn't communicate with the back end system. [00:14:34] Speaker 00: There might be a conflict, Facebook might have some evidence, but there was contrary evidence and that that was a genuine issue of fact the district court failed to recognize in deciding the prior summary judgment. [00:14:47] Speaker 00: the district court basically makes the same error again here, but also with respect to the UDB, the user database, the evidence showed that it's not clear that there's different information stored in that than the timeline back end because there's a lot of information we presented about the right path that shows the synchronization of the two. [00:15:07] Speaker 00: And so there are [00:15:09] Speaker 00: There are flaws both in the, is this information that they're contending really passed to the back end system, and is it really true that it's not stored in the back end system to the extent that pointing to particular pieces. [00:15:22] Speaker 02: I think I'm... Yeah, you used all your time. [00:15:25] Speaker 02: You always serve somebody, but I'll just one quick question. [00:15:27] Speaker 02: Sure. [00:15:28] Speaker 02: You agree that if we were to agree with some portion of what you're saying, that compels us to reach the alternative basis for affirmings, which would be the 101 issue. [00:15:39] Speaker 00: I would agree that at that point it would be appropriate to reach the 101 question, yes. [00:15:46] Speaker 00: Obviously we agree with the district court's conclusion. [00:15:49] Speaker 00: We disagree with Facebook as to the 101. [00:15:52] Speaker 04: This is also a 101 related question. [00:15:56] Speaker 04: Certainly. [00:15:57] Speaker 04: I think it maybe is addressed to both sides. [00:16:00] Speaker 04: So these patents have expired. [00:16:03] Speaker 04: And in fact, the last of them expired more than six years ago. [00:16:07] Speaker 00: That is correct, Your Honor. [00:16:08] Speaker 04: April of 2024, I think, as I calculated. [00:16:11] Speaker 00: That is my understanding. [00:16:12] Speaker 00: I don't have the precise statement. [00:16:15] Speaker 04: If there was non-infringement here on [00:16:20] Speaker 04: you know, true non-infringement, non-infringement because of invalidity. [00:16:26] Speaker 04: Is there any stake, from your point of view as the patent owner, that Facebook has in the validity or invalidity of these patents? [00:16:38] Speaker 00: No, Your Honor, I would agree that that's, and I'm not certain my friend is going to disagree with you on that, but I'll let her speak to it. [00:16:45] Speaker 04: You can't sue them anymore on these patents. [00:16:48] Speaker 00: That is correct. [00:16:49] Speaker 00: That is correct. [00:16:50] Speaker 00: There could not be a new lawsuit on these patents. [00:16:53] Speaker 02: Is there any other pending litigation involving these patents? [00:16:57] Speaker 00: I certainly am not aware of any. [00:16:58] Speaker 00: I believe the answer is no. [00:17:01] Speaker 00: And I'll be happy, I don't know how much my friend wants to speak to 101, but I'll be happy to address it on rebuttal if necessary. [00:17:08] Speaker 04: Thank you. [00:17:15] Speaker 01: Good morning. [00:17:15] Speaker 01: May it please the court, Heidi Keefe on behalf of Metta. [00:17:19] Speaker 01: I'll address your last question first, because it's very easy. [00:17:21] Speaker 01: We don't have any stake. [00:17:23] Speaker 01: If you find that there is no infringement in a firm on that ground, you don't need to address 101. [00:17:27] Speaker 01: because these patents are not able to be asserted anymore against anyone. [00:17:32] Speaker 02: And it was raised as an affirmative defense and not a counterflame in the original suit. [00:17:37] Speaker 01: That's correct. [00:17:38] Speaker 01: And here, it's just because there is an alternative ground. [00:17:42] Speaker 01: And so we made sure that we had it in the cross appeal. [00:17:46] Speaker 01: With respect to the remainder of Your Honor's arguments, I want to make sure that I kind of help us reframe, because I absolutely agree with the question, which is, [00:17:57] Speaker 01: Why didn't their expert tie this all up? [00:17:59] Speaker 01: Why isn't there some evidence tying these things together? [00:18:03] Speaker 01: What we have here is nothing more than a conglomeration of attorney argument about broad, high-level documents. [00:18:11] Speaker 01: In every instance, the documents that are being referred to, they say, oh, that shows you everything that's there. [00:18:18] Speaker 01: That doesn't show any other paths. [00:18:20] Speaker 01: But these documents are high-level documents describing [00:18:24] Speaker 01: at a sense of generality. [00:18:26] Speaker 01: I mean, the title of the document, if you go back to Appendix 14279, is Timeline in a Nutshell. [00:18:34] Speaker 01: In other words, it's not every single thing that ever happens here or that we're talking about. [00:18:39] Speaker 03: Well, but that's attorney argument, what you just said. [00:18:41] Speaker 03: I mean, you're here to argue. [00:18:42] Speaker 03: I get it. [00:18:43] Speaker 03: But why isn't this a jury issue as to whether or not Timeline in a Nutshell is comprehensive or is merely general? [00:18:51] Speaker 01: So I have two answers to that, Your Honor. [00:18:53] Speaker 01: The first answer is that you're not supposed to give every single inference of anything that could ever happen. [00:19:00] Speaker 03: Every reasonable inference. [00:19:01] Speaker 01: Every reasonable inference. [00:19:02] Speaker 01: And they have to be reasonable inferences, and they still have to be supported by something other than attorney argument. [00:19:09] Speaker 01: All we've heard regarding these documents is attorney argument. [00:19:13] Speaker 01: The expert did not say, and therefore that shows that every single item is there. [00:19:19] Speaker 01: They focused only on [00:19:20] Speaker 01: the words in this document. [00:19:22] Speaker 01: The biggest problem is this document never once addresses what's coming into the aggregator writ large. [00:19:30] Speaker 01: Never once is coefficient described anywhere in this document at all. [00:19:35] Speaker 01: Coefficient is one of the items of information that comes into the aggregator. [00:19:40] Speaker 01: It's undisputed that it comes into the aggregator. [00:19:42] Speaker 01: Their only argument was that it's not a unit of direct interest to the party or that somehow it's a query. [00:19:49] Speaker 01: It's not a query. [00:19:50] Speaker 01: even if you were to reach that, the evidence that they pointed to was a deposition testimony that was highly ellipsed. [00:19:59] Speaker 01: And in its ellipsing out, and this is on page 14161, the information that was ellipses out shows that the aggregator receives coefficient from coefficient 2 system. [00:20:17] Speaker 01: among other information, and it uses that information to construct. [00:20:22] Speaker 03: Where are you on 14161? [00:20:24] Speaker 01: I'm on 14161, page 71, which is the bottom right-hand quadrant. [00:20:32] Speaker 01: Page 77, so 1461, page 77. [00:20:35] Speaker 01: And if you see starting at, for example, line 12, the quote was, [00:20:44] Speaker 01: that we will get a request which includes the ellipseed out, the above information, the above mentioned user data, and it will list, that was ellipseed, back to all of major life events we get from tau, the coefficient, the ellipseed out we get from the coefficient two system, which is outside of the aggregator and the timeline database, and a bunch of other necessary data, this was ellipseed out, to help the aggregator construct the final request, [00:21:13] Speaker 01: Sorry, the final response. [00:21:15] Speaker 01: And this information will be sent to Timeline Aggregator, also ellipsed out, and some of the information there will be then used to input to the query. [00:21:24] Speaker 01: So that statement was the witness saying all of the coefficient data, the major life events, these are information that is received by the aggregator in order to create this query. [00:21:37] Speaker 04: This is a Facebook witness? [00:21:39] Speaker 01: This is a Facebook witness. [00:21:40] Speaker 01: That is correct. [00:21:42] Speaker 01: And the only way that they could try to argue that somehow this was just information coming from somewhere else was to ellipses out the appropriate facts. [00:21:52] Speaker 01: That's even worse than just attorney argument. [00:21:56] Speaker 01: But I would also point, Your Honors, because I think it's very important to the intellectual science and technology case. [00:22:03] Speaker 01: This is intellectual science and technology versus Sony, 589, F3, 1179, and specifically at 1184, [00:22:12] Speaker 01: In that case, in order to attempt to defend against a grant of summary judgment of non-infringement, the plaintiff in that case had put forward a copy of a diagram depicting a circuit. [00:22:39] Speaker 01: The attorney argument was simply that circuit shows that all the circuits are going to be like the circuit that we're accusing. [00:22:47] Speaker 01: The problem is the attorney never tied anything about what was actually in that document back to the claim language of each and every. [00:22:56] Speaker 01: That case is highly on point to this case. [00:23:00] Speaker 01: And the court specifically held that simply pointing to a multiplexer alone was insufficient [00:23:06] Speaker 01: to then infer that all circuits like that would have the required claim limitations. [00:23:12] Speaker 01: Here we have the exact same problem. [00:23:14] Speaker 01: They're pointing to documents that don't have all of the information. [00:23:18] Speaker 01: None of these documents overcome the fact that coefficient data is received into the aggregator and never falls over into the timeline database. [00:23:28] Speaker 01: Never. [00:23:28] Speaker 01: That's an unrebutted fact produced through Facebook's witness [00:23:33] Speaker 01: And it's also important to note that the Facebook witnesses who testified, in this case, coefficient data, you can find at appendix 8158, paragraph 12, as well as for newsfeed. [00:23:47] Speaker 01: Actually, you probably don't even need that. [00:23:49] Speaker 01: So the coefficient data is unrebutted, that it shows up in the aggregator and never shows up in either the leaves or the timeline database. [00:24:00] Speaker 01: And that's in none of the documents that were [00:24:04] Speaker 01: proposed to add inferences. [00:24:06] Speaker 01: So right away, there's an absolute missing hole for everything. [00:24:13] Speaker 02: The testimony you cited to us is listed as confidential in the appendix. [00:24:20] Speaker 02: All those confidential markings. [00:24:23] Speaker 01: So the confidentiality is just because it's discussing the rest of that deposition is discussing source code. [00:24:29] Speaker 01: But the portions that you relied on are not confidential. [00:24:32] Speaker 02: I don't believe so, no, Your Honor. [00:24:34] Speaker 04: And it's your witness. [00:24:36] Speaker 03: That's fine, yes. [00:24:39] Speaker 03: The yellow brief directs us to 14273. [00:24:42] Speaker 03: I don't think this is a document we've talked about before. [00:24:46] Speaker 03: They argue that yellow at 11, a reasonable juror could conclude that the references to timeline data in that document means all data for timeline and that it persisted in timeline database. [00:24:59] Speaker 03: means that it was stored in the timeline back end. [00:25:03] Speaker 03: You understand better than I do, but why is that not a possible reasonable inference a juror could draw? [00:25:09] Speaker 03: Okay, so just as... Can you just give the ASIN? [00:25:14] Speaker 01: 14273. [00:25:14] Speaker 01: 14273, and it appears as the stock image has a figure on the top and then storage. [00:25:19] Speaker 01: And Judge Post mentions this is also more confidential. [00:25:22] Speaker 01: Can we talk about it? [00:25:23] Speaker 01: We are absolutely okay to talk about it, Your Honor. [00:25:25] Speaker 01: And again, because most of these documents are relatively high level, some include some snippets of source code, but we haven't touched any of that, so no problem. [00:25:33] Speaker 01: So the reason that that is not an appropriate inference here is that timeline, unlike what my learned colleague said, the timeline backend system is larger than simply the timeline database and the aggregator. [00:25:47] Speaker 01: They chose only the timeline aggregator and the timeline database to be the computer system that they were trying to find. [00:25:56] Speaker 01: But the timeline back end receives information from numerous other places. [00:26:00] Speaker 01: I mean, even here in the document above, it shows Webby, Memcache, Tau. [00:26:05] Speaker 01: And that's not everything. [00:26:07] Speaker 01: There also is a UDB. [00:26:09] Speaker 01: There are numerous other sources of information, including coefficient, which we've talked about. [00:26:14] Speaker 01: And it's unrebutted. [00:26:15] Speaker 04: So it's true that there are these other bubbles, Memcache, Webby. [00:26:23] Speaker 04: What tells us that they are part of [00:26:26] Speaker 04: timeline, as opposed to somewhere else that timeline takes information from. [00:26:33] Speaker 01: Yes, that's correct. [00:26:34] Speaker 01: But see, what happens with, so the timeline database is only one piece of what goes into your timeline. [00:26:39] Speaker 01: As was discussed in the declaration, timeline is a system that in order to construct a user's timeline, the front end says, hi, I want my timeline, sends a request that we talked about just a minute ago saying who the user is, and that request includes the user. [00:26:55] Speaker 01: The aggregator receives that information and then uses that information to construct numerous requests. [00:27:02] Speaker 01: Some to go get information, for example, from Tau, which might include photos. [00:27:06] Speaker 01: To get information from Coefficient about what things are more important to you. [00:27:10] Speaker 01: To get all sorts of other information, which is then brought together, and this is the quote that I was directing you to from the deposition, that is then used to create a request which is forwarded to the timeline database. [00:27:22] Speaker 01: And so when they simply say, [00:27:24] Speaker 01: timeline data is persisted in timeline database, we don't dispute that. [00:27:29] Speaker 01: Timeline data is persisted in the timeline database. [00:27:32] Speaker 01: But there's also coefficient data that is not persisted in the timeline database. [00:27:38] Speaker 01: If the document had said all information used by the backend timeline system is persisted, they may have an argument. [00:27:46] Speaker 01: But they don't. [00:27:47] Speaker 01: Timeline data is in the timeline database, not disputed. [00:27:51] Speaker 01: But coefficient data, which also goes into the aggregator, is not in the timeline database. [00:27:58] Speaker 01: Major life events, which also goes into the aggregator, not a part of the timeline database. [00:28:04] Speaker 01: And so that's what Judge Codel took exception with, was that they were seemingly having to rewrite the document to say all information in the timeline back end is persisted in timeline DB, and that doesn't happen. [00:28:18] Speaker 01: The judge addresses this at appendix 50. [00:28:21] Speaker 01: in his order down below. [00:28:23] Speaker 01: He wasn't weighing evidence, he was just saying it doesn't say what it says. [00:28:26] Speaker 04: Just say again why, for example, coefficient data is not, oh, you said it is timeline data? [00:28:33] Speaker 01: No, coefficient data is coefficient data. [00:28:36] Speaker 01: It is not timeline data. [00:28:37] Speaker 04: Even though the timeline and backend system makes use of it? [00:28:42] Speaker 01: Absolutely. [00:28:43] Speaker 01: You can use data that is not itself timeline data. [00:28:46] Speaker 01: For example, the way the Facebook system works, and I think undisputed, is [00:28:51] Speaker 01: Facebook users can take actions. [00:28:54] Speaker 01: When I log into Facebook and post something, that becomes timeline data. [00:28:59] Speaker 01: That's something that I've done. [00:29:00] Speaker 01: That's an action that is stored in the timeline database. [00:29:03] Speaker 01: However, there are other things that happen where I'm not interacting with Facebook. [00:29:07] Speaker 01: For example, if I got married, Facebook doesn't know what I did off of Facebook. [00:29:12] Speaker 01: There's a whole separate section within Facebook called Life Events, where you can input what happened to you that day. [00:29:17] Speaker 01: That's a separate section of Facebook. [00:29:20] Speaker 01: You can input those major life events. [00:29:22] Speaker 01: They're not part of the timeline database. [00:29:24] Speaker 01: But when Facebook constructs a timeline, you wouldn't want the fact of your marriage to be missing just because it didn't happen on Facebook. [00:29:32] Speaker 01: And so it reconstructs it that way. [00:29:34] Speaker 01: Coefficient, similarly, is a system that is used by multiple parts of Facebook. [00:29:39] Speaker 01: The whole point of coefficient is to understand how you're interacting with all of your other friends. [00:29:46] Speaker 01: So that if Facebook wants you to stay on the site longer, [00:29:49] Speaker 01: it's going to show you more information with the people you interact with the most. [00:29:52] Speaker 01: That's what coefficient does. [00:29:54] Speaker 01: It calculates those scores, and it uses those scores to make sure it gets the right information. [00:29:59] Speaker 01: Those are all fed into the aggregator so that it can build the appropriate machine. [00:30:04] Speaker 01: That's all undisputed. [00:30:06] Speaker 01: Their only argument against coefficient was that it somehow wasn't a data unit because it was just a score, or that it wasn't of user interest, or somehow it's a query. [00:30:16] Speaker 01: None of those are true. [00:30:18] Speaker 01: Unrefuted deposition testing. [00:30:19] Speaker 04: And you agree that even the query part, let alone the user direct interest, is part of the argument about claim construction on data unit. [00:30:30] Speaker 04: And if we agree with you and disagree with Mr. Lindell on that, then there's no separate argument about the facts regarding coefficient data. [00:30:44] Speaker 01: That's absolutely correct, Your Honor. [00:30:46] Speaker 01: And in fact, their only argument [00:30:47] Speaker 01: And on coefficient, I don't believe they have any arguments other than, I think I heard him say, our expert talks a lot about code. [00:30:55] Speaker 01: But again, there, he never, in that report or any citations that are in any of the briefs, ties that back to the claim language to actually say, and this is why, that's not there. [00:31:06] Speaker 01: Because the truth is coefficient is in the aggregator, and they can't refute that. [00:31:10] Speaker 01: They haven't refuted that. [00:31:12] Speaker 01: They just say, but it's not the right type. [00:31:14] Speaker 01: of the information. [00:31:15] Speaker 04: So the various paragraphs, like 13, 10, 26, and so on, with the usual longish names of code components, none of that, you say, would be capable of being reasonably read to cover the coefficient data? [00:31:41] Speaker 01: It does not. [00:31:41] Speaker 01: And similarly, Your Honor, [00:31:43] Speaker 01: even if it were, it is not explained. [00:31:46] Speaker 01: Therefore, they have no one who can explain why that would be the case. [00:31:51] Speaker 01: And so there's no reasonable inference to be drawn because the expert can't just rattle off a bunch of information and then not tie it back to the issue. [00:31:59] Speaker 01: And that becomes attorney argument then, and that's the fairing case that we cite also. [00:32:05] Speaker 03: The district court has statements throughout the opinion that certain evidence is unequivocal and sentences, at least, that I fear seem like they're sort of fact-finding, as opposed to just saying, could a reasonable fact-finder come out this way? [00:32:23] Speaker 03: Do you have anything to say about that? [00:32:25] Speaker 01: I understand what Your Honor is saying, but I actually think if you read those all in context, they're actually [00:32:32] Speaker 01: not weighing the facts at all. [00:32:34] Speaker 01: They're just literally looking at the words of the document and seeing whether the words of the document say what they want it to say in terms of the documents that were presented. [00:32:42] Speaker 01: What you noticed in the district court's opinion, especially on page 50 and I think 51, he comments that they had to change the language of the document in order to make the argument that they were making. [00:32:54] Speaker 01: And those were the types of points. [00:32:55] Speaker 01: I don't think that's weighing evidence. [00:32:56] Speaker 01: That's just saying what the evidence actually says. [00:33:00] Speaker 01: And in terms of crediting our witnesses, there was simply nothing discrediting those witnesses. [00:33:07] Speaker 01: They didn't have anyone come in and say, but I see this in that database. [00:33:11] Speaker 01: And the coefficient data does show up here in the timeline database. [00:33:16] Speaker 01: There is zero evidence of that. [00:33:18] Speaker 01: And I note that I'm out of time. [00:33:20] Speaker 01: I don't know if Your Honors wanted me to discuss 101 or not. [00:33:24] Speaker 02: No. [00:33:24] Speaker 02: We'll take it on the papers. [00:33:26] Speaker 01: Very much appreciate it, Your Honor. [00:33:39] Speaker 00: I want to return to a couple of points that my friend addressed during the argument. [00:33:49] Speaker 00: A question that Judge Stark posed a few moments ago about some of the district court's comments about the evidence being unequivocal and suggesting that it might reflect fact-finding and weighing of evidence. [00:34:00] Speaker 00: I think there's a portion of the opinion that makes that extremely clear. [00:34:05] Speaker 00: The district court, in discussing this court's prior decision in this case, [00:34:11] Speaker 00: talked about how this court remanded after finding that there was evidence in the record from which a reasonable trier of fact could conclude that the Tau system did not provide information to the timeline back end system and that as a practical matter that meant summary judgment was inappropriate. [00:34:31] Speaker 00: The district court interpreted that and said it was interpreting that to basically mean [00:34:36] Speaker 04: Well... What are you referring to? [00:34:39] Speaker 00: You have a... A citation? [00:34:42] Speaker 00: Yes, let me find a page. [00:34:44] Speaker 00: It was... Sorry, it's in a different... [00:35:08] Speaker 00: So I think it's at page 55 of the appendix, where the district court basically said that it characterized this court's decision as having found that Facebook failed to provide, quote unquote, definitive evidence. [00:35:31] Speaker 00: There's no such thing as definitive evidence on summary judgment. [00:35:34] Speaker 00: If there's a conflict or a dispute that goes to a fact issue, you don't get to resolve that by saying, well, if you put in another declaration, [00:35:43] Speaker 00: I get to decide that two declarations is better than one, or that those two pieces of evidence outweigh the other one. [00:35:49] Speaker 00: That's the wrong summary judgment inquiry, but that's what the district court admitted it was doing here, was looking for the wrong kind of material and applying the wrong standard. [00:35:59] Speaker 03: What's your response on the coefficient argument? [00:36:02] Speaker ?: Sure. [00:36:02] Speaker 00: So, I think two things. [00:36:04] Speaker 00: First, the evidence we talked about and the portions of the experts declaration, starting at page 13, I think it's two [00:36:19] Speaker 00: 13291 and carrying over to 292 starting at paragraph 82 of the Experts Report and then there are a number of subsequent paragraphs where he talks about all of the things that come to the back end system and I think part of the problem we're having here is the same one the court addressed the last time which is the improper sort of mixing of the front end and the back end and suggesting that somehow we treat them the same or that Facebook would prefer to treat them the same. [00:36:48] Speaker 00: In reality [00:36:49] Speaker 00: The front end system does a lot of the things that my friend talked about. [00:36:55] Speaker 00: That's not to say that the back end does it, and the evidence doesn't show that the back end is what's doing it. [00:37:01] Speaker 03: If I want to understand what your expert says about what happens to the coefficient at the back end, [00:37:06] Speaker 03: Do I look at 13.291 or somewhere else? [00:37:08] Speaker 00: That and I would say, regrettably, several pages following that. [00:37:13] Speaker 00: He doesn't talk about coefficient in the specific way that I think you're asking. [00:37:18] Speaker 00: He talks about the system broadly and how the system doesn't receive other information than what is stored in the database. [00:37:25] Speaker 03: So then just at a high level, what's your response to, well, how is he going to explain it if we let you put him in front of a jury? [00:37:32] Speaker 00: Sure. [00:37:33] Speaker 00: So he's going to explain it that [00:37:36] Speaker 00: What they're saying is, coefficient gives this information to formulate the request back to the database. [00:37:42] Speaker 00: But the documents don't show that. [00:37:44] Speaker 00: The documents he points to, which are the same ones I was discussing during the earlier portion of the argument, the presentations, the design documents, et cetera, don't say that coefficient information is coming to the system. [00:37:57] Speaker 00: Now, we can't necessarily expect that design documents are going to say all the things that don't happen. [00:38:03] Speaker 00: That's kind of an odd [00:38:05] Speaker 00: thing to ask for in a design document, but they say what does happen, and they don't point to the things that Facebook's arguing here, and that's evidence that the jury could accept and say, well, if that was really happening, wouldn't they have put it in their design documents? [00:38:19] Speaker 00: Wouldn't they have put it in this material that purports to show how the system works? [00:38:24] Speaker 00: I would also add that the testimony that my friend referred to from their witness at page 8158, if you look at that and that declaration as a whole, one thing is very notable. [00:38:36] Speaker 00: There are no citations to Facebook code, to Facebook design documents, to Facebook material of any kind. [00:38:42] Speaker 00: It's a declaration standing alone that just says, this is what happens. [00:38:47] Speaker 00: I'm not going to show you any evidence that it's actually true. [00:38:53] Speaker 02: Thank you. [00:38:54] Speaker 02: We thank both sides in the course of this committee.