Oral argument — Experian Information Solutions v. Nationwide Marketing Services

Before we get started, I noticed that your brief is marked as under seal, and I just wanted to make sure we don't edge into anything that's sealed, or if we need to close the courtroom or anything. I don't believe that we intend to get into anything under seal. I didn't pause the brief to figure out what was and what wasn't, so if any of the counsel get close to something that they know is under seal, I guess we can make arrangements. Thank you, Your Honor. I appreciate that. We're also, you understand, on streaming. Yes. So this is not just a courtroom. It's going out to the world. No, I understand, Your Honor, and I appreciate the sensitivity to that. You know, again, I don't know what it is, but I have the big thing that says under seal. Okay. We missed out the clock. Thank you, Your Honor. Good morning. May it please the Court. I'm Robert Unichel on behalf of the Appellant Experian Information Solutions. This is an appeal from the Arizona District Court's grant of summary judgment on copyright and trade secret claims. There are four errors that the district court made that I'd like to address today, Your Honors. The first, the district court erred by ruling that a compilation consisting of names and addresses cannot be copyrightable because names and addresses are uncopyrightable facts, regardless of the process used to select the information that goes into those names and addresses. But the Copyright Act, the text of Section 101 of the Copyright Act, actually defines compilations which are protectable as, quote, formed by the collection and assembling of preexisting materials of data that are selected in such a way that the resulting work as a whole constitutes an original work of authorship. The district court here specifically said that whether or not a selection is a factual compilation possesses the modicum of creativity as needed to be original, quote, this is from the district court's opinion, it does not look to the actions of the compiler selecting the facts, unquote. That's on ER 16. But this court has made clear in, for example, CDN versus CAPAS, that in fact you have to look at the process that is used to select the information that constitutes the facts in order to determine whether or not the selection process bears the hallmarks of originality and creativity. Yeah. Well, but the most famous case here is Feist, which is the phone book. Correct. So I guess the question, and the district court obviously thought that this was like a phone book. So I guess the question is in lay terms, why is what we have here different from a phone book? So the phone book in Feist, or any white pages phone book, is a mechanical reprinting of the information that, in the case of Feist, rural telephone got from its customers about their name, about their phone numbers, and then it just reprinted that into a single directory. There was no selection of information. There was no resolution of conflicting information or data. There was no election of what information to put in. It was exactly what was obtained from the customers. Here, Experian is actually making its own judgments from a huge amount of conflicting data points, exactly what information it's going to put into every field. So, for example, Experian will vet. Well, they're all names and addresses. What you're going to put in is still a name and address. That's true, Your Honor, but there's a couple points about that. Number one is, for every person, there's actually a huge amount of name and address information that's available from multiple sources that will, for example, if you've moved in the last year, if you own a residential property as well as a second home, if you are an investor in a landlord property, all of these will come up in the data as addresses attributable to you. What Experian does not do is reprint all of that information into its database. What it does is it tries to pick a single address that it believes is the best, the most valuable to its marketers. And in doing so, there are 16 of them. Kagan. Oh, it exercises judgment as to what source it uses. Not just. Yes, Your Honor. First, as to what source it's going to use. Uh-huh. Then, and in fact, it uses multiple sources. When you have conflicting data from multiple sources, it then has to pick which of the information from the multiple sources it's going to put in to each of the fields. So, for example, the first thing it's going to decide is, is one of these a business address? Is one of these an old address? Is one of these a residential or a second home property? And it's going to make a determination it would like to get to somebody's primary residence so that the mail gets delivered to them. Along with that, for example, Your Honor, there's elections about what goes into the name fields and the household fields. So for a name, my name is Robert. For any of the multiple data entries, you're going to have some Robs, some Roberts, some Bobs, some Robbies. There's an election that has to be made for each one of the people that's going into the database. What name gets included? For people who have hyphenated last names, there's an election of you're going to do both names. Is there going to be only the first, only the last? Because differing records will have differing entries, and what Experian does not do is simply take the records that it gets and reprint them and collect them in a database. It takes the multiple pieces of information that it gets about every person and every address, and it makes its own value judgments as to which pieces of information should go into that database. I might point out, number two, Your Honor, that the district court actually erred by resolving this on summary judgment rather than presenting it to the FactFinder. In North Coast v. Maxwell, this ---- You didn't cross-move for summary judgment? We did not cross-move for summary judgment, Your Honor. Okay. In North Coast v. Maxwell, this Court held that the decision as to whether or not creativity involved in creating a work was sufficient or was trivial is an issue for the FactFinder. This is consistent with rulings from other courts. Matthew Bender v. Westbrook ---- I'm sorry. You know, you have your own strategy. I'm just wondering what it is that the FactFinder would find. So in this case, Your Honor, what the ---- I mean, that's sort of the basic facts. Does Experian do the things that you ---- Do they dispute that, what you say? I believe there are disputes as to the level of decisionmaking that goes into the various elements of the name and address fields. For example, there's a dispute as to the degree of hygiene, what's called, that's applied to the names and addresses. Hygiene is ---- I see. So the things you've been reciting to us that Experian does are things that your client claims it does, and they're not undisputed. Not all of them are undisputed. I do not believe that all of them are undisputed. I do believe that it is agreed ---- And the trial presumably would resolve any dispute as to those things. Correct, Your Honor. And that ---- But once that those disputes were resolved, it would seem to me the question of copyrightability is a question of law, isn't it? Yes, Your Honor. The ultimate determination of copyrightability is a question of law. According to the Second Circuit in Matthew Bender v. West, quote, whether the particular elements of a work demonstrate sufficient originality and creativity to warrant copyright protection are a question for the factfinder. The ultimate determination of copyrightability is for the court. But the question of whether there is sufficient creativity or originality in the choices that are being made by Experian is one that should be presented to the factfinder in the first instance. Well, for purposes of the ---- their motion for summary judgment, they must assume that what you are saying is correct. Is that right? So ---- I don't believe that they fully assumed that what we said is correct, Your Honor. I think that they presented their argument that the choices that we are making are not sufficiently original or creative. I don't believe that they ---- As a matter of law. As a matter of law, that's what they suggested, Your Honor. And the court said that the facts themselves were not copyrightable, which is, of course, true under the Copyright Act. But the court said that it did not need to look to the process that we used to make the selections. It did not need to look to the evidence, for example, showing that Experian's database statistically is different than the other databases that independent studies have compared it with. This is not a situation, for example, like with multiplication tables. With multiplication tables, there would be one possible answer that would have to go into any compilation of multiplication tables. Here, there are multiple choices, multiple ways of expressing an address to make it valuable in your estimation. And that's verified by the fact that the independent studies actually show that when you compare Experian's databases to other databases, there are ---- This is all related to your copyright claim. Do you want to address your trademark claim, too? Yes. Thank you, Your Honor. Because you're running short of time. Agreed. The district court erred in the trade secret claim by ruling that a compilation consisting of names and addresses can never be entitled to trade secret protection because publicly available individual names and addresses are supposedly readily available. But Experian never, ever claimed that individual names and addresses by themselves are a trade secret. Experian asserted that its compilation of more than 250 million names and addresses were its consumer view database. I thought the argument was that there is a considerable overlap between Experian's database and non-secret public databases, you know, 50, 60, 70, 80 percent overlap. I thought that was the argument. I don't believe that's the argument. It's certainly not the argument that the district court relied on, Your Honor. What the district court said was if you disaggregate the compilation and look at each individual name and address, you could go out and find that name and address all on your own through publicly available records. And, therefore, you cannot say that when you put all of that publicly available material together that the resulting database is somehow not readily ascertainable by proper means. But is it Experian's argument with regard to the trade secret that really it's the process by which you get to this compilation that's the trade secret? No, Your Honor. Okay. The process does in fact give rise to a trade secret, the compilation itself, but it's the value of the compilation itself. So the statutory definition of trade secret focuses on the economic value of the compilation due to its secrecy. And here it's the compilation, not the process, but what results from the process, that we license out and get millions of dollars a year for because it is valuable to consumers, because it is not readily available by other means. The reason they license it from us at considerable cost and under strict regulations as to how they can use it is precisely because we have created, through considerable effort, a database that is economically valuable due to its secrecy. What was the basis for the district court's ruling on the trade secret? On the trade secret count, this is approximately at E.R. 20, the Court said that because the individual names and addresses which make up the database are publicly available on their own for the most part, that the compilation is not a trade secret. Correct. And your position is that that's incorrect as a matter of law? That is incorrect as a matter of law. And as well, I might add, Your Honor, that it is up to the fact finder to determine whether or not there is a trade secret based on the statutory definitions of a trade secret, and there's ample evidence in the record here to show that we've met all of those. Let me ask you this. Your client has many licensees who have access to the database. Is it tens of thousands? I believe it is tens of thousands, Your Honor. Why isn't that a wide enough disclosure that it's no longer a trade secret? To use the classic example, if Coca-Cola gave its formula to 10,000 people, at some point you'd say it's no longer a trade secret. Assuming it still is, Coca-Cola is still a trade secret. I don't know that. But just to use it as an example. Your Honor, if Coca-Cola was to give its formula out without any restriction or confidentiality obligations, it may very well eventually stop being a trade secret. But that's not what Experian does. With everybody who licenses it or uses any portion of it, there are very strict I mean, let's say there are, but you don't think at some point there get to be so many that even with restrictions you can't really consider it to be private or secret. You know, it's 40,000 people having the same secret. It's not really a secret anymore. But, Your Honor, from the trade secret perspective, it's a question about the secrecy as creating a valuable asset. And here the valuable asset is the fact that other than coming to Experian and taking a license agreement subject to specific restrictions, you cannot otherwise get this information. So it does not cease to be a trade secret as to Experian, merely because Experian has had a lot of people who want to license it. So what do licensees get when they get a license? Do they get a list of names? Do they just get a database that they are able to search without seeing the whole? What is it exactly that they get? Every license is obviously subject to its own unique terms and negotiations, but any customer can license portions of the Consumer View database that are relevant to its needs. So that could be a huge number of fields that go beyond names and addresses and include all sorts of other information. But when you get a list, do you get to look at the whole list or do you just get to do a get a searchable database? You will get, I believe you get both a list and the database, Your Honor, but it will be restricted to the fields that you have actually taken a license to. I take it in order to prevail on your trade secret, you will have to show that they knowingly, wrongfully acquired this. That's correct, Your Honor. And so we would, what you're asking us to do is to say that the district court erred as a matter of law in saying that this compilation can't be a trade secret and then send it back for everything else? Yes, Your Honor. The same with respect to the copyright claim? You just want us to say the district court erred as a matter of law and send it back and everything is there for further proceedings? Without reaching any question of infringement or anything else? Correct. I don't believe that this court can or should make those determinations based on this record, Your Honor. Thank you. We'll hear from the other side. Good morning. May it please the Court, Mark Fuller with my colleague Christopher Thompson on behalf of the appellee. I'd like to start, one of my opportunities this morning is to make certain that we have focused narrowly on the copyright issue because my sense is, and particularly in the reply brief, there may have been a subtle shift of arguments. The argument below was that Experian uses its judgment and expertise to coordinate the right name with the right address, and it says that it was this arrangement that ultimately results in the CVD being a compilation that, quote, glows bright with the creative spark that warrants protection. That's the issue. It doesn't have anything to do with it. That's pretty good. Well, the creative spark language comes directly from the case law. I would suggest that it doesn't glow very brightly in this case, but that's the argument. It doesn't have to glow very brightly. It just has to glimmer. Feist says a modicum doesn't have to be a lot of creativity. It just has to be a little tiny bit of creativity. So a little glimmer is enough. Well, I think when you are engaged in the discovery of preexisting facts, facts that exist in the public domain regardless of whether you are compiling them or not. All facts are preexisting facts. You know, gravity is a preexisting fact. You know, the fact that you can put airplanes heavier than air objects up in the air and have them fly around and not fall down. You know, all those things are preexisting facts. That's correct, Your Honor. But public domain. But when you invent a device that actually does it, it gets patented and, you know, it's. . . Well, yes. There's nothing that's created out of thin cloth, out of whole cloth. Well, Your Honor, I beg to disagree because when you get in, for example, to the copyright cases and you deal with things such as the CDN case where you are estimating something. For example, you are taking public domain data and making inferences or hypotheses and coming up with a creative original authorship of estimates. Those are not really preexisting historical facts, I don't believe. Sure. The probability existed whether or not you determined it. All right. But I think we're mincing words here. I mean, the fact is what opposing counsel has presented is a process that, I don't know, strikes me at least as being involving creative choices. Maybe not what we would, you know. . . maybe not on a Van Gogh level or Degas, but enough under Feist to say they make decisions as to who's included and who's not included, what kind of name is included and what kind of name is not included for people who have multiple names, which of multiple addresses is the one that's going to be the address that's listed. That sounds much more like a creative process than happened in Feist. To me, Your Honor, that sounds much like the census example that Feist quoted as a paradigm of non-copyrightable kind of process, that the resulting list, the resulting compilation, if you will, as we pointed out in our brief, what the census does in terms of trying to determine the most accurate names paired with the most accurate addresses, and indeed for a particular day, which involves more in terms of estimates and opinions, is the paradigm example of something that does not warrant copyright protection. And I think the Eleventh Circuit. . . I don't remember what did Feist say about the census. I missed it. I mean, of course, the census is not copyrightable because it's a public work, just like our opinions. Correct, Your Honor. Creative as they be are not copyrightable. I've often regretted that, but there it is. Correct, Your Honor. Let's turn to Feist. Where are you reading from? Wait one moment, Your Honor. I'll pull it up here. Sorry, Your Honor, if you'll give me one moment. I mean, Feist talks about there's nothing remotely creative in arranging names alphabetically, and that's an original practice and so on, and that's right. I mean, arranging names alphabetically doesn't. But that's not certainly what Experian is doing, right? No, it isn't, Your Honor. It's not. Maybe your co-counsel can figure out where on Feist you're referring to and slip you a note. Well, I will find it here in one moment, Your Honor. Since you were relying on the census. Your co-counsel is sort of leaving you hanging there. I thought maybe you could talk about something else about co-counsel. He's coming up. He's behind you. That's excellent. He's giving you something, taking away something. Okay. Let me turn. Feist does, in fact, refer to the census, as do a number of its progeny, and we'll find that. Well, the question is what it says about the census. I don't remember the discussion. I believe that it talks about it, but I don't remember it being a key part of the discussion. Well, it's used as an example of something that would not. Yes, thank you. It's actually highlighted here, along with a good deal of the rest of the case. And so we're looking at page 1289. I have page 355. Or 347 of the U.S. reports. That's what we usually refer to, you know, the official reports. 347, okay. This is because facts do not as the last paragraph on that page. Last paragraph of that page. The last paragraph begins, it's a better principle. Yes, that's right, that paragraph. Facts do not owe their origin to an act of authorship. The distinction is one between creation and discovery. Census takers, for example, do not, quote, create the population figures that emerge from their efforts. They copy these figures from the world around them. Census data, therefore, do not trigger copyright because these data are not original in the constitutional sense. Okay. So how does that help you? Everybody agrees that the names and addresses themselves are facts that aren't copyrightable. Well, correct. The Court certainly didn't say that when the census — again, it has no reason to talk about copyrightability of census data because it's a government publication and can be copyrighted. But it's certainly not saying that when the Census Bureau decides it's going to look for certain kinds of data and not other kinds of data, as it asks all sorts of questions, that that compilation is not copyrightable. It certainly didn't say that. Well, Your Honor, I — I'm sorry. So your reliance on the census is a little misplaced. Well, I beg to disagree, Your Honor, but I can see that I'm not going to persuade you about that. I believe that the Eleventh Circuit later also, and I would point to the Bell South case, which applies the feist. And I think this addresses the very issue before the Court. To be sure, BAPCO employed a set of strategies or techniques for discovering the data. Any useful collection of facts will be structured by a number of decisions regarding the optimal manner in which to collect the pertinent data in the most efficient and accurate manner. If this were sufficient, then the protection of copyright would extend in census data. But they're claiming more than the manner of collecting, which would be, so do you go out on the street and look at the street address, or do you go on the — they're saying we call the data and make judgments about which things to include and which things to exclude. Oh, and I think that's — that's important. I don't think that is the argument. The argument is we go out and we search from public domain sources to try to find the most accurate information. They call it deliverability. But not choosing from sources in some other creative sense. They're trying to find what is, they believe, accurate. Precisely the same thing the census does. Well, the data — here's what I'm not understanding. Yeah. The data itself is not copyrightable. Correct. So the names and addresses themselves are not copyrightable. Correct. But what Feist seems to be saying is that a compilation of these things, if there is some judgment exercised in deciding what's in the compilation, is — can be copyrightable. A compilation could be. That's true. If all you are taking — So why is this one not? Because it is not a selection of particular, for example, consumers. It is not a subset. It's not the creative spark in taking information and culling it down into something else. It is, we are going to go out to the public domain. What do you mean by something else? Something else? In any of the other cases. For example, you can take the case about the Chinese-American businesses. You can take the one about CDN, which they rely heavily on in coming up, taking public domain data and coming up with estimates of value. That's where the creativity is. It is not going out and just collecting raw names and raw addresses. And I — I mean, that's my reading of the cases and my reading of Feist, but I see my time is running down and I want to — So, counsel. Yes. If we were to decide that this compilation was copyrightable, would it be Nadamark's that we should also decide on the infringement issue as a matter of law rather than sending it back based on the record? I believe it can be, and in particular in connection with the virtual identity standard, and we've briefed that. I do, and I hope that's answered your question. If it has, I want to move on quickly to something that is case dispositive regardless. Well, the district court didn't reach that, right? The district court did not reach that. It's a separate basis to affirm, an independent basis. I want to — I want to get to the Antonick airframe set of cases in Dine, and that is it is undisputed on this record, and it's a threshold issue because it resolves copyright. It's dispositive, is that they did not compare the registered copyrighted work to the allegedly infringing work. Antonick, Judge Hurwitz's opinion in that case says that's a requirement, and every other case that's looked at this, I believe, has said that. And the airframe case out of the First Circuit, which Judge Hurwitz cited, is directly on point. In Dine, the In Dine case out of Florida, is also almost exactly on point. They involve subsequent versions of computer code databases and so forth. That is, as we said in our brief, an unyielding legal requirement. You cannot compare. And I want to be clear about what the objective facts and the objective record in this case. The registered copyright is an — is a version of the database. It was called X149, version 149 of this database. That's the allegedly infringed work. Experience in-house comparison used version 183. Thirty-four versions later, during which, if we use the 1.5 million people moving, consumers moving per month, roughly 52 million consumers had moved. And that's the comparison they rely on. Antonik, airframe — This presupposes that this is copyrightable, and this goes to infringement? This goes to whether you can prove a — no. This goes to substantial similarity as a matter of law. And this is resolved and decided time and time again as a matter of law. The courts look and they say, if you did not compare the — the registered work with the allegedly infringing work, you lose, bar none. And what's interesting is the — as I said, the airframe case is directly on point. And in that case, there was actual evidence that the defendant had copied. It's an — it's an interesting case. Not only was there plainly access, but when the — when the defendant was copying or some employee was copying, there was actually in the margin in the code comments to the effect of, I don't know what this means, but I'm leaving it in. And yet the First Circuit said, you compared — you compared a much later version. And it was updated. It had changed. It was updated in the ordinary course of business. What do we have in this case? We have a — we have a database that is updated on a 24-hour-a-day, 7-day-a-week basis according to experience. Thirty-four versions later, 52 million people have moved. And that's their only evidence of similarity. That's their only evidence of infringement in this case. And I would suggest to you, and as I said, Antonick cites a number of those cases, including the Airframe case. So this is another issue that this Court didn't address that you think is in the record? It is absolutely in the record, and it's case dispositive as it relates to the copyright issue. Are you going to say — can you say 10 seconds' worth on the trade secret? Yes. It's — these cases are almost always customer list cases. They deal with certain attributes. But I think, Judge Ellis, your question was directly related to a point I wanted to make, which is the argument here is that the process — there are public domain sources out there. There are phone directors. There's a national change-of-address database. There are utility records. There's various public sources. And we've come up internally with some system or method, and we plug that data in, and it spits out what we think is the most accurate data. And I think what experience has done is conflate the process with the result. The process itself might be trade secret, because you could then go out and use that to cull from public domain sources and continue to generate these kinds of accurate lists. But I think then you go to the Phoenix case — well, you go to Enterprise Leasing Case first, which says that a comprehensive listing, at least the way I read it, of public domain data doesn't suffice. Public domain data in general is not subject to trade secret. The method of selecting out from that might be trade secret. I'm not sure. Yeah. I mean, they claim that — I mean, lots of stuff's in the public domain. Every ingredient on the Coca-Cola formula, as best I can tell, is in the public domain. Sugar, you know, I don't know what all the stuff is in the formula. But they claim that what the secret is is the compilation, is the formula. So the fact that the stuff is culled from the public domain just strikes me as being — Well, no. I mean, I think we're — I mean, and I'm not sure I disagree. It's the formula. Right, the formula. The formula is — Their claim is we culled this. We used a secret process to take out names, leave certain names in, and come up with this thing that is not a reflection in itself. I mean, it has stuff that's in the public domain, but the thing as we have culled it doesn't exist anywhere except in our secret database. So why isn't that a perfectly good trade secret? They may be wrong. They may not be able to prove it, but why isn't it perfectly good? Because I think the process is secret. The process, not the end result. It's not the Coca-Cola bottle that's sitting over on counsel table. It's the process of taking those and the process by which you come up with that. Right? That's the process. And you can't, from the Coca-Cola — well, actually, an engineer or somebody trained in that probably could. Why if you keep — I mean, the problem with Coca-Cola is you have to send it out for people to drink, so you can't really keep the result secret. But in this case, they're saying we create a result and we keep it a secret. We're licensing it to people, and each of those licensees is subject to strict confidentiality. Two points. The process is not loaned out to anyone. The process isn't. A list of names and addresses may be, but that's just the ultimate list. That tells you nothing about the process that goes into it. Nothing at all. And I will tell you that — You just made an irrelevant answer. You just made an irrelevant comment. Nobody claims that the process is secret. Their claiming is what we have created is a secret database that we have then handed out to licensees who are then bound to keep it secret. Not the process. They specifically said not the process. No, I understand. I think that is the — No, you didn't understand because you said the process isn't — they are talking about the database that is a result of the process. I understand. And my distinction is to say that the process might well be a trade secret and that a list of public domain names and addresses is not. I agree that's not the argument they're making. I understand that's your position, but why? Because it is the process that is kept secret. It is the process. In fact, their nondisclosure agreements exclude public domain data. But they can't protect the result of the process. That's what you're saying. Well, I think they could as to 99% of this database. Attributes, I mean, you get into the customer list cases. If you call out certain attributes and so forth, we're not talking about any of that. I'm saying that if you go out into the public domain and you collect names and addresses, how you do that, if you come up with some special algorithms and so forth, how you do that may well be a trade secret. Let's talk about customer lists. If a company has a list of customers, all of those names can be found on the Internet or Yellow Pages or something because they're customers. And yet, customer lists are traditionally viewed as subject to trade secret law. Well, I think the Khaleesi case. People get prosecuted for stealing them. Well, I believe the Khaleesi case and the enterprise leasing case indicate that if the customer list is simply a list of some names and addresses that are basically available to competitors, then those are not trade secret. It's a selective listing of attributes. For example, a particular policyholder. A list of policyholders, I believe, is not a trade secret under Arizona law. But if you have a list of policyholders and then a series of attributes about them, about perhaps what particular kind of insurance do they buy, what kind of premiums do they pay, what are their coverage limits and so forth, now you're getting into something that I think may well, under Arizona law, suffice as a trade secret. I don't believe that a fair— I think you're out of time. Oh, I have been out of time for a long time. I was just trying to respond to your questions, Your Honor. Thank you. Thank you. We'll give you two minutes for rebuttal. Thank you, Your Honor. I'll be brief. I'd like to address just two points from opposing counsel's discussion. First, as to the Antonin case, the airframe cases, these are, in fact, related to the infringement analysis. The substantial similarity analysis is infringement. There is no record on infringement here because the district court, having found that this was not a copyrightable database, did not engage in that. That doesn't mean that there's no record on it, right? There's no record. Either there's no rulings from the district court, and the only time these cases were raised were in the reply brief on the motion for summary judgment. There was no opportunity for us to actually engage or go through any of the relevant discussion. I will note, however, that in the Data East versus EPICS case, this court, the Ninth Circuit, said that what is required is not necessarily a comparison of the original work to the alleged infringing work. It's that there must be sufficient evidence of the contents of the work to make a, quote, fair comparison. Here, there is — there is, in fact, three groups of comparisons. Is that a compilation case? The Data East was no. That was a source — that was a code case, Your Honor. Here, there is evidence that it was presented below on summary judgment and in the broader record that shows there are three comparisons which have been made to the actual consumer view database, not to some recreation of it or simulation of it. In Antonick, I should note, there was no evidence whatsoever of any comparison. The only claim in that case was that if you could prove access, even without showing what the access was to, that that would suffice to establish substantial similarity. And here, the court said, no, you can't just prove access. You have to show some comparison. We have comparisons here, and that's why this is a fact issue that should be dealt with at the district court level. However, the comparison was not involving. So the copyright went through September 30, 2011. And Dr. Hinman didn't compare the database from 2011, but instead compared it to an Experian file from 2012, and then two NatiMark files against an Experian file from 2014. So the fact that Experian continually updates its database, right, so that changes are going to be made over time when you get different information and you want to prove the accuracy of that information, because that's the value of the database, that it's necessarily going to change over time, I think is what counsel's point was. Therefore, if you're comparing it to something later, to not the copyrighted work, then it's kind of irrelevant as to how similar or dissimilar it is, right? Yes and no, Your Honor. First of all, I should note that in addition to Dr. Hinman's analysis, there was the original comparison which was done of the NatiMark sample in May of 2012, which is much closer in time to the date of the registration, which showed a 97.81 percent similarity or match rate between the nationwide sample and the Experian sample. So the record is not just reliant on Dr. Hinman's analysis, but moreover, the complaint to the extent that there would be a complaint then about that comparison invokes the best evidence rule, which was an objection that was not made or ruled on on an adequate record below, where, for example, in Airframe, the case that they're relying on, the footnote number nine of the decision, which is on page 107, specifically says, quote, if the best evidence rule is satisfied, evidence other than the original may be sufficient to establish the content of the copyrighted work. Here there are too many fact questions about what the differences are or are not, what the availability of the original database is or is not, and those are not available for this Court to rule on on this present record. All right. Thank you. Thank you, Your Honor. The case you saw here stands submitted. We are now going to hear argument in the last case on the count of Booth v. United States.