# Oral argument — Silicon Graphics v. ATI Technologies > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A9552 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** November 3, 2009 ## Text We have five cases this morning on the calendar. A patent case from the District Court, a trade case from the Court of Federal Claims, and two veterans' cases, one of which is being submitted in the briefs and won't be argued. Our first case is Silicon Graphics v. ATI Technologies, 2008, 1334, and 1353. Mr. Manning, you indicated you want to take eight minutes and save seven. Yes, Your Honor. I counted eight issues to argue. No doubt you'll pinpoint the most important and best arguments. Yes, Your Honor. May it please the Court. As to anticipation, ATI has appealed the District Court's JMOL ruling that the four references presented to the jury do not anticipate. One critical claim construction error that the District Court made in denying ATI's motion for JMOL was to limit the claims to specialized graphics hardware. The District Court said for the first time in the JMOL order that the claims were limited to a computer system operating on specialized rasterization hardware that allows for high-speed interactive rendering through a graphics pipeline. These words, specialized rasterization hardware, high-speed interactive rendering, are nowhere in the claims. There is no word in the claims that could be reasonably limited in this way, and certainly the term computer system in the preamble cannot require specialized high-speed rasterization hardware. There are no words that manifest exclusion anywhere in the specification limiting computer system. And SGI's argument importing... Didn't your experts say that the frame buffer was hardware-based? Our expert Judge Rader, Dr. Patel, said to my recollection that these claims and that the frame buffer may be hardware-based, but also would cover software. In other words, that the patent could speak about both hardware and software. In fact, two of the anticipating references, OpenGL and RenderMan, are what's called an interface, which is basically teaching hardware and software how to interact with each other and how to operate. And so Dr. Patel was always interpreting these claims as involving both hardware and software, surely never denying that they covered hardware and that hardware was being spoken about, but that the computer system was much broader than hardware alone. Another procedural error in the district court's J.M. O'Neill opinion was when the court adopted... But this is a jury case. The jury then would have had substantial evidence to find that the frame buffer was hardware. We, Your Honor, had a motion in limine, excluding any references to specialized graphics hardware, because we knew that this argument would be made. The motion in limine was held in abeyance. Numerous objections were made throughout trial to these arguments in cross-examination. Of course, that's going to be an abuse of discretion standard. You're going to tell us that this magnificent district judge abused her discretion? Your Honor, I do believe that to find for the first time in the J.M. O'Neill ruling that these claims were limited to specialized rasterization hardware, graphics hardware... But my point is that substantial evidence is against you on one point. Abuse of discretion is against you on the other. You're really rolling up upstream. I do understand that, Your Honor, but we also believe that to have such a finding on these claims with the preamble saying computer system does rise to that level. Another procedural error in the district court's J.M. O'Neill opinion was when the court adopted a construction of rasterization that was narrower than the construction provided to the jury. The court found for the first time on J.M. O'Neill that rasterization requires entirely floating point scan conversion. It was not the construction that was in front of the jury for the entire trial, nor was it the construction in the jury instructions. The new construction of rasterization requiring entirely floating point scan conversion was adopted at SGI's urging at trial and in their J.M. O'Neill briefing over ATI's repeated objections. We're talking about what? Claims 1-6? We're talking about claims 1-6, also claim 9, and really predominantly claims 1-16, Your Honor, with respect to the word rasterization. Because claim 1 says which operates on a floating point format. Yes. So what's wrong with that point construction? If I may request, which element the court is referring to in claim 1? The second element? Maybe that relates to cross appeal rather than your argument. I believe it does, but I'm clear. All right. Okay. So you're agreeing with the supposition that I was making? Yes. All right. But I guess if you could clarify for me, and obviously we're not going to get into the cross appeal at this juncture, but it's unclear to me. It seems to me that what you're saying is consistent with what your opponent is arguing in his cross appeal with respect to the construction of rasterization process. Am I wrong about that? No, I don't believe you're wrong about that, Your Honor. And I believe there is some consistency, but I believe that the court, for the first time in J.M.O.L., to find that rasterization required scan conversion entirely in floating point was wrong because the district court had previously found that rasterization involved translating and filling in. And while scan conversion was entirely in floating point, scan conversion was not a part of rasterization ever throughout the entire trial. So then to find that it was a part of rasterization in the J.M.O.L. ruling at opposing counsel's urging for the first time in the J.M.O.L. brief had some inconsistency with, which I will address in my rebuttal, had some inconsistency with arguments that had been made previously. I believe that under Hewlett-Packard v. Mustak it was improper for the district court to adopt this construction after trial to deny A.T.I.'s motion for J.M.O.L. I will now move to the issue of abandonment. A.T.I. has appealed the district court's dismissal with prejudice of the invalidity counterclaims that related to patent claims that were found not infringed on summary judgment. The district court found on summary judgment that 12 claims from the 327 patent were not infringed and there was no longer a case for controversy. The district court thereby divested itself of jurisdiction over A.T.I.'s invalidity counterclaims for the non-infringed patent claims. For counterclaims to be abandoned, the court had to have jurisdiction. The court did not have jurisdiction and therefore A.T.I.'s counterclaims were moot. Is your position that the court didn't have jurisdiction because in the absence of anyone saying anything, those particular counterclaims just were gone? With due respect, Your Honor, we believe that the court did say something very clearly in the district court's summary judgment opinion. It was a 60-page opinion but that the court very clearly said no case for controversy remained and hence it was clear that there was no jurisdiction remaining. With respect to the claims that were found non-infringed and that the only conversation... So if there's no jurisdiction, those claims weren't dismissed, right? They were still there. You didn't move, did you, to dismiss without prejudice given that you thought there was no jurisdiction? We did not formally move. Right. Someone had to do something with those claims. Even though you're saying there was no jurisdiction, they're moot, they're still sitting there on the docket and someone has to do something to resolve them, correct? That's correct, Your Honor, and that's exactly what we're asking this court to do is to dismiss without prejudice. Which is something you never asked the district court to do. Am I wrong about that? I don't think the court is wrong about that in the discussion that preceded the trial. I cannot point to the record right now where we did that after trial because the court's JMOL ruling came down and when it came down saying that these counterclaims were abandoned, I was shocked. I called the court immediately, requested briefing on the issue, and we then briefed the issue and there was a subsequent opinion on the issue in which we believe there were three serious errors that the district court made in that analysis on the abandonment issue, which we point out in some detail in our brief. One of those errors was that the district court indicated that we had not discussed prior to her ruling that we could go forward on these four claims of 17, 18, 22, and 23, that the trial was restricted to those four claims and that's just borne out completely incorrect by the record. Those four claims were discussed explicitly with the court almost in the transcript two or three pages immediately before the court ruled that we could go forward. Mr. Manning, you want to save seven minutes here. Into that time you can save it or use it. I will save it and I thank you very much, Your Honor. Mr. Bolinger. Thank you, Your Honor. Please support. I'm counsel for Silicon Graphics and I'd like to start by shifting the focus to our appeal and discuss that. Our appeal is on three claims construction rulings and a ruling regarding an implied license, a license defense that was asserted as an affirmative defense in this case. Before I go there, though, I'd like to identify two fundamental facts which I believe provide kind of the foundation of this debate and this dispute. The first one is the idea of what the core aspect of the invention is which relates to and all the claims deal with the use of floating point format data throughout an entire graphics pipeline which is comprised of three parts. Can I ask you a threshold question? Sure. Am I wrong? I understood the other side. Well, the question is was the other side arguing in the context of anticipation for the same view of rasterization process that you're arguing for on cross-appeal? Is that your understanding? It's hard for me to know because they've changed their position. They argued at the district court that rasterization required scan conversion and it required it to be entirely in floating point. That's not our position. Our position is that the district court's definition of rasterization was essentially correct. And so their expert testified at trial that that's what the definition of rasterization, including all those other limitations were, and that's why the district court included it in the Jay Maul ruling. That's all that had to do with it. Well, you agree with what the district court said on how she defined rasterization, but you disagree with how she applied rasterization process. There was a longer clause, Your Honor, a rasterization process operating in floating point format, and she construed that as requiring 100 percent floating point throughout the entire rasterization process. That's on our red brief. It's on page 22, 23, I think, for the actual clause that was in dispute. And basically there were three reasons that that was an erroneous conclusion. She construed it, the district court construed it as a whole, and that was improper because the claims don't require 100 percent. And we look at the – there's a heavy presumption that what the claims say is the correct interpretation. Their ordinary meaning should apply. In addition, the specification clearly notes at several points that rasterization can include multiple processes that can be in fixed point, which is diametrically opposed to what the lower court ruled on rasterization in that case. What I'm confused about is I know you're arguing also on scan conversion and the construction of scan conversion, but let's assume we disagree with you on scan conversion and we buy into what the summary of the invention says, which is scan conversion process is now handled entirely in floating point. Does that infect our analysis of the claims and the use of rasterization process in the claims? It doesn't. It doesn't because rasterization doesn't mandate that scan conversion operate entirely in floating point. Mr. Bollinger's column 4 says the scan conversion process is now handled entirely, entirely is the word, on a floating point basis. That's correct, Your Honor. And the reason that there's no disclaimer with that language is because if you read it in context with the rest of the summary of the invention, you can see that they're talking about a number of processes individually operating in floating point. If you go down the next line, you'll see that it talks about various other rasterization, texturing, fog, all operating in floating point. And then you read further into the specification and you see that at column 11, lines 39 and 42, that it says one or more of these rasterization processes, including scan conversion, fog, texturing, can be done in fixed point without departing from the invention. And the term entirely is not in the claim. Scan conversion was never... Well, entirely is at least in the summary of the invention. So the portion that Judge Rader read to you. That's correct. So if we construe this, if we read the claims in the context of the specification and we construe scan conversion entirely on a floating point process is requiring that scan conversion be in a floating point, what does that do to your arguments? Those claims that require scan conversion would now be redundant because they also specify floating point parameters. But they would also be very narrow, which goes to my other point, which is that they've basically changed a very trivial number and made it fixed point. The XY coordinate data... I guess I'm not clear. You would concede then if we read scan conversion entirely on floating point as being definitional. Then you would concede that the claims that include scan conversion are not infringed. A hundred percent floating point as the way the district court ruled, I think that's correct. But why wouldn't that infect our view of the other claims, which don't reference scan conversion but talk about a rasterization process if we all agree that the accused devices go through scan conversion, which is not on floating process? Because the phrasing in the other claims were specifically structured to capture the operation of other rasterization processes other than scan conversion and to require that those other processes operate in floating point. So we would read those claims that don't specifically reference scan conversion as not consisting of scan conversion. Absolutely. That would be the proper reading and view of the specification and the number of... But let me address the scan conversion disclaimer argument. Just one further point because this court has talked about in several cases... So claims 206 would be remanded then? Absolutely. And claim 1 is... But again, I would back and say that scan conversion shouldn't be... There should be no disclaimers on the sole use of Summary of the Invention paragraph, its location as a defining, because the Praxar decision and the other decisions that this court has identified on this issue where Summary of the Invention included language that could be construed as exclusionary, like the word entirely, where they went back and they looked through the rest of the specification and how the terms were used in the claims. And Praxar is a good example of this, the Voda decision also. And you see that the weight of those other statements in the rest of the specification and how the term was used in the claims, where there was a redundancy by adopting the disclaimer, the court clearly ruled there was no manifest exclusion and no intent to parse out. So if I understand this correctly, what you're really arguing is that lightning or fog, if it's appearing in a motion picture, some complex operation would need a floating point and more mundane operations, routine operations, would stay in fixed point. That's part of this invention. What the invention was about, it wasn't about putting scan conversion entirely in floating point. That's one aspect of it. There are other aspects, and that's the important factor here that brings this case into the libel Floerschend case and also in the Praxar decision and pulls it out of what happened in the C.R. Bard case. The fact is, in this case, the... The difference in Praxar is that the summary statement was later clearly contradicted in the rest of the specification. Do you have a similar indication here? Yes, because... What is it? At column 11, lines 39 through 42, the specification specifically states that the enumerated processes above, which include scan conversion and rasterization, can be done in fixed point without departing from the invention. I'm not... What line are you... 39 through 43. 39, yeah. Where does that include scan conversion? Well, I apologize. The rest of the passage above identifies the various blocks associated with that particular figure. Figure, I think it's four. And the various blocks are processes that fall within the rasterization, including texturing and things like that. The same paragraph occurs again at the end of column 12, lines... I can't remember the lines right now, but it's basically the same kind of caution that don't hold us to 100% floating point because there's so many parameters. Computer scientists would know there's too many numerical values that are being handled throughout the entire process of these computer systems. It's line 27. Again, it should be noted that one or more of the above blocks can be implemented in a fixed point format without departing from the scope of the present invention. And that, in figure 5, you'll see there are a number of rasterizing and scan conversion operations within that. But let me just clarify, even if we were to disagree with everything you've just said with respect to scan conversion, your position would still be that claims to through... well, at least the claims that don't specifically reference scan conversion would still... her claim construction with respect to rasterization process would still require a reversal. Absolutely. Her rasterization... Well, let me ask you just hypothetically a housekeeping question then with respect to that question. If we were to agree with you on that point, does that open up a revisit? Does that throw out all the questions of prevailing party, mootness, invalidity? I mean, does that all... that reopens all of those questions, including the invalidity issues, right? But, I mean, if the anticipation issue or maybe some of the other invalidity issues that fell off were based on what we've reversed as a claim construction, then all of that has to be revisited, does it not? The waiver issue doesn't because it has nothing to do with the claims that were actually presented at trial. So there's no claim construction, JMAL issue on that. And the other... even those four claims, if the construction was considered to be so bad that it infected the jury verdict, then that would have to be reversed. But we don't believe it was. We think there are plenty of other reasons. Anticipation failed in this case. Can I... I have a couple more minutes and I was going to jump to the Microsoft license, which is three errors in construing that license. And again, it's section 2.4 of the license that they asserted in the Microsoft, which immunized acts when a combination product was found to infringe the 327 patent. And it was based on whether the product Windows was actually... contributed any of the claimed elements within it by use of this term separate and apart. And the error of the lower court was first to focus only on use infringement and not get into discussions regarding... Are all the ATI cards working with Windows? No, but a vast majority... Well, what remains? There are other operating systems out there, Linux and Macintosh. But the vast majority work with Windows. But Windows has absolutely nothing to do with the graphics processing. The only evidence they showed was their expert, Dr. Wolf, who said that you can't load programs into the ATI card, third-party games and things like that, without an operating system, which is absolutely correct, undisputed. However, that's not the separate and apart analysis. The separate and apart analysis looks at the entire infringement, the panoply of possible infringing conduct, and assesses whether any of it would be still infringing if Windows didn't contribute... was not part of the combination. And the fact is that the apparatus claims, as we show on page 16 of our reply papers, there's a clear diagram of theirs, a unit that shows all of these features, the rasterization circuit, the scan converter, the fog unit, all of these things are in there and has nothing to do with Windows. So if we affirmed you on the license point and agreed with you on claim construction, would we still need to remand? I think so, Your Honor. On the way the case is developed, it still needs to go down and have a final determination on the question of infringement. I don't think the record before you is sufficient to provide a finding of infringement by ATI. We're moving these defenses. And I do have a couple... I'm going to leave a minute-thirty left for my rebuttal, unless there's anything on the Microsoft issue you want to get into. If you want to sit down, we'll give you all your rebuttal back. We'll hear from Mr. Manning, who has a good amount of time. May it please the Court. SGI, in their closing argument, and for the first time in writing in the JMOL brief, stated that scan conversion entirely in floating point was a part of rasterization. This claim construction was adopted by the District Court for the first time in the JMOL order. SGI concedes, on page 14 of their gray brief, that ATI does not infringe any 327 claim under the construction of rasterization that SGI successfully urged the District Court to adopt on JMOL. SGI should not be permitted to urge one construction as to rasterization to the District Court and argue for a different construction on that same term in this Court. It seems to me that maybe the answer, or at least a response to your point, is that there's a difference between rasterization and rasterization process, and that that explains, even under the Court's claim construction of rasterization, that would not preclude what the other side is arguing with respect to a rasterization process in the context of this comprising claim, prefaced by the word A. The District Court, Your Honor, to my understanding, found that there was no difference and only construed at Markman the word rasterization and said that it would then be used throughout the way rasterization is used and as to rasterization process. In the District Court's summary judgment opinion, the District Court amplified or clarified that construction of rasterization and said rasterization process as a whole has two parts, translating and filling in, which was the prior construction and the construction requested by SGI, translating and filling in two parts, not entirely in floating point, or that scan conversion was a part of rasterization. And when the District Court then found those two parts, translating and filling in, the process as a whole, the District Court said that meant that some aspects of translating and filling in needed to be in floating point. And because we do not do any translating or any scan conversion in floating point, but only X and Y coordinates in fixed point, the Court found no infringement and that was the basis of the ruling. But I do believe that there is an inconsistency in the argument they've made at JMOL versus what was done in the Markman hearing and we request that the Court hold them to the argument they made in JMOL to avoid anticipation and not attempt to then broaden the definition and come back to include infringement and, based on the concession they made in their brief, the case would be over. Now, if I may, I'll move to the SGI license with Microsoft and then back, if my time permits, to scan conversion. The first prong of the Microsoft license, basically it's this question. Can the ATI product infringe, separate and apart, by itself, when it is not combined with Windows? The answer to that question in the District Court was emphatically no. The 327 apparatus claims require a circuit structure for performing rasterization. Until Windows provides programming, ATI GPUs are unprogrammed general purpose processors and they cannot contain the alleged infringing structure. That was unrebutted testimony from Dr. Wolf in the District Court and any testimony that was cited by their expert, Dr. Gleicher, the District Court found did not come close to creating a genuine issue of material fact. On the second issue, the second prong of the license, is assuming intent, would Microsoft be liable for inducing infringement? And the answer there is emphatically yes. The gray brief is the first time that SGI ever argued that inducement is outside the scope of the second prong. By not raising it at the District Court, or in its red brief, SGI waived this issue. Now, assuming the court goes on and looks at it, this court is well aware of the two aspects of indirect infringement, contributory and inducement. Contributory requires knowledge and inducement requires intent. The second prong of the Microsoft license assumed knowledge and intent, thereby referencing inducement by referring to intent. Intent has meaning only in the context of inducement. We believe that the District Court correctly applied the Microsoft license to this case and to these facts and that the District Court should be upheld in that regard. Back briefly with the time remaining to scan conversion, I would note that in the patent at columns 11 and 12 in the section cited, essentially lines 4 through 46 in column 11 and lines 18 through 32 in column 12, that there is no reference to scan conversion in any of that and any talk of scan conversion in figure 5. There's no reference to scan conversion in figure 5 of the patent. So all of those comments do not relate to scan conversion. Scan conversion is spoken about in the very beginning of the patent, as this court has noted, and the parties essentially agreed on the definition of scan conversion as the process that specifies which pixels of the display screen belong to which primitives. That's column 1, lines 31 through 33, and conceded in their great brief at 8. The only disagreement was whether scan conversion should be entirely in floating point, and this court has noted what the words say. I have run out of time, and I thank you very much. Thank you, Mr. Manning. Mr. Bollinger has time for a rebuttal on your cross-appeal. Thank you, Your Honor. The first point I want to make is on the definition of rasterization, and the district court got it wrong when they ruled that it had to be 100% as a whole, and so our discussions on that are quite clear. That application... Isn't he right that scan conversion doesn't appear anywhere in columns 11 or 12? It's embraced by the term rasterization, and in fact, if you look at the functional operation in those columns, I'm not sure if the word actually appears. I don't think it does, but if you look at figure 5A, you can see specifically that there are these polygon rasterization, 501, line segment rasterization, point rasterization, and those embrace scan conversion. Certainly the polygon rasterization does because of the language... But am I wrong that figure 4, the portions of column 11 that you were citing to us before went to figure 4 and not figure 5? The column 12 dealt with figure 5. Right, but the column 4... 4 dealt with figure 4, and you see rasterization here is broadly identified in 406, and next to it you have texture, fog, and anti-aliasing, and collectively all of those are described as rasterization. The first block, 408, clearly encaptures and embraces scan conversion because all of these talk about the scan conversion, and it was good that Mr. Manning identified the definition of scan conversion in the patent because it defines it precisely as we presented in the district court at Markman without the requirement of entirely floating point. So when you define something in the specification and you exclude something like that, I think that negates the effect of the summary of the invention in the statement there. I'd like to, if I could, just jump back briefly to the Microsoft issue in my remaining time. We argued consistently below that the second prong required a showing of contributory infringement. The district court rejected that and said basically, and ruled that all it needs is direct infringement of the accused combination. If you read the second prong of 2.4 double I, there's no way that can be properly construed that way. The only way to make sense of the second prong is to limit it to contributory infringement and to apply the requirement that if Microsoft is to satisfy that second prong, they need to show that it's a staple article of commerce with substantial non-infringing uses, and Windows clearly can't meet that. It is a staple article and it has substantial non-infringing use, and that's why they never made that argument. They were limited to trying to bootstrap an inducement which we opposed at every stage. And unless there's any other questions, I appreciate the court's indulgence and allowing me the time to respond. Thank you. Thank you, Mr. Bollinger. The case will be taken under advisement. Thank you very much. Thank you. Next case is Shanghai Aswell et al. versus the United States and the Sioux Honey Association et al., 2009, 1163, Mr. Moshe. Good morning. I appear on behalf of Chinese respondents to challenge the Department of Commerce's determination in its second annual review of the anti-dumping duty order on honey from China for the period December 1, 2002 through November 30, 2003. We raised three issues on appeal, all of which relate to the manner in which the department calculated surrogate values. We know our burden is great. As I am certain you are here today from the Department of Justice and Petitioner's attorneys, as they have argued in their briefs, the department is entitled to considerable discretion as to how they interpret the anti-dumping law. And they'll argue that this court cannot substitute its judgment for that of the agency or re-weigh the evidence. We accept our burden, since we know that this court also recognizes that DOC discretion only goes so far. The department is required to calculate margins as accurately as possible. That's the key to the dumping law. Also, the department cannot accept surrogates which lead to absurd results. To calculate accurate margins, the department should rely on as broad and representative sample as possible. And this is critical for our case. The fact that the sample size may have many price points does not matter if the sample is unrepresentative. And there we refer to the data that they use to calculate the surrogate value. You're talking about the EDA? The EDA data. Is it your position that the EDA, they used the EDA alone and didn't rely on the three articles you're posing? Correct. But is your position that what they should have done is not just rely on your three articles but relied on some combination of the EDA and the three articles? Yes, Your Honor. We accept the fact that the EDA data represents the price of raw honey in one region of India for a representative period. But it's a very limited period. It's the first half of the period of review and it's one region and it's high-priced honey. And we believe that the record shows that the price of honey is very volatile and we believe the record shows the price of honey actually declined in the second half of the period of review. So what the department did is, by accepting this price as the price for the entire POR for all of India, it ignored substantial evidence on the record as to what the price should be when you look at all of the evidence. Well, but they had reasons for discounting the Kerala article and the Jharkhand article and the Girijan article. Yes, Your Honor. And we believe that the reasons they used to disregard those articles are the same reasons they could have used to disregard the EDA data. For example, they say that the articles are only for one period. The articles are only for one region. Single beekeepers. Well... Single producers. Single producers. But if you're operating in a market economy, the single producer represents the market. The single producer is selling at a market price because India is a market economy. So the fact that it's a single producer is what we're saying is not a valid reason to reject this article. The EDA data... I mean, we admit it has a lot of price points. But the fact that it has a lot of price points doesn't really mean that those price points are representative as to what the price of raw honey is within India within all of India for the entire period. So the EDA data is part. But the reasons that they rejected our data... You could reject the EDA data for the same basic reasons. It's not representative of a price for the entire period of review. And it's critical in this case... But EDA was multiple producers for the largest area in India. I don't believe the record shows it's the largest area. It is a honey-producing region. But I believe the record is going to show that other regions in India are just as significant in honey production. If you look at the EDA data, again, it's one region. The other articles are data from three other regions throughout the POR. Was the EDA data compiled by a first-year business student? I don't know who the EDA data... The EDA data is on the Internet. The EDA has a wonderful website. That's the Kerala article, of course. Well, Your Honor, they're articles. I mean, this is tough. I mean, it's very tough because we're dealing with information that we get from India on the Internet. I'm not saying that the EDA data is not good data for a particular region for a particular period. What I'm saying is that you have all this other data that the department just says no, no, no, no. They reject all the other data we put in the record. And then they reject the other evidence that we put in the record where we show that the price is very volatile. You look at the EDA data itself. It shows from one year to the next prices jump. And then, very significantly, you have the Tribune article, which was published in December. Prices in the Kerala region generally are lower. The economy is not as strong there. Prices are going to be different in different regions of India throughout the period. And that's the exact point we're making. They've allowed, again, one period, one piece of data, even though it has multiple price points. But of critical importance to us, you look at the Tribune article, which the department used in another review, which the department has said is a reliable source of data. You look at that Tribune article, it says that the price is declining in the second half of the POR. And very significantly, it says even if the price was 40 rupees per kilogram, Indian beekeepers would still be making honey. So here you have the Tribune article saying that you can make money at a price of 40 rupees per kilogram in India. So it was happening in India. It was happening in EDA that you had this tremendous ability of the India beekeepers to make a tremendous profit. If you can make money selling honey at 40 rupees per kilogram, the EDA data shows 76. So basically you're taking, and what was the reason that they had this tremendous price increase in India, again, if you look at the Tribune article, it was because the EU was banning Chinese honey. So the Indians just kind of moved in, and there was a tremendous demand throughout the world because of a non-market force where the European Union puts a ban on Chinese honey, so the Indians could increase the price of honey. And then you're taking that same Indian price, which has this tremendous profit, and you're saying that's the surrogate value of Chinese honey. With the Chinese, if there's a ban on their export of honey, the Chinese price would be even lower. So we're getting hit, really, from all points. What they're using, and what they're rejecting, what we're saying is the department is not looking at the entire record. They're not looking at all the evidence we're putting in. And one more critical point. It's not that they're not looking at it. There's no question here that they're looking at it because they were required to go back and opine as to what was not reasonable or why they didn't use it. So they looked at it. You're just disagreeing with the bottom line as to the usefulness and utility of it. If I just say I'm disagreeing with the bottom line, I'm saying we're re-weighing the evidence. But I'm not doing that. What I'm saying is their whole methodology is flawed. What they're doing is they're looking at each piece of data that we put on the record in isolation. If you take one of our articles, if we had just one article and the EDA data, and you put those up there on a board and say, well, which is better? I'm going to have to say that the one article to the EDA data, is better. But that's not the point. What they should be doing is looking at the totality of the evidence and trying to make the correct decision to get the most accurate margins by looking at a broad and representative database as to what's going on in India throughout the whole POR. What they've done is they go one by one and knock us down. You say that the Commerce set the normal value based on a raw value of $1,800 per metric ton. Where did you get that from? The raw value of honey, that is... Where was that in the record? Well, I hope we have... I had my clerks searching. We couldn't find it. I hope we had a site as how it's derived. It would be from one of the confidential exhibits, and I think we didn't use a confidential number, but that would be the data of raw honey that is on the record. Can you give me a page site and I'll look at it? Would it be possible for me to send something in afterwards right now? Hopefully I can get it after I sit down, but I may not be able to. Thank you. But I absolutely can get that because that's there. So again, what we're saying is on this particular issue, what the department did is they take one piece of evidence that we put in and compare it to the E-data, and as constant one by one, they knock down every single piece that we have because the EDA data is not that bad. But what they didn't do is they didn't look at the totality of the record, and they especially ignored all the evidence in the Tribune article as to how Indian beekeepers can make a profit selling at 40 rupees per kilogram, and how the tremendous increase in Indian prices in the middle of the POR was because of the ban on EU honey. And if you look at everything, if you look at the totality as to what the DOC should have done, rather than go one by one, we think that the methodology was contrary to law and the decision was not supported by the evidence. I'd like to talk about, if you have any more questions, on the first issue. On the second, you know, really clinical issue, which is financial ratios. And here it's a little bit different. Here the Department of Commerce had two financial statements. They had MHPC and they had APIS. And again, the results are just incredibly different. They accepted MHPC, which is an uplift on the honey price of approximately 40 percent because of the MHPC ratios, and they rejected APIS, which is an increase of less than 6 percent. So because of that decision, again, our margins are much higher than we believed would be proper. And the reason they rejected APIS is because they said that the financial statement was not complete because it didn't have an auditor's report on it. Well, the financial statement didn't have an auditor's report on it, but we don't believe that's sufficient for the Department of Commerce to throw out this statement, which came from APIS and we had third-party evidence in the record that shows that APIS is one of the leading exporters of honey from India and had substantial domestic sales of honey in India. The APIS report also, and this is also very significant, if you look at the balance sheet in P&L, it has an auditor stamp on it, it has an auditor number, and it says that the auditors are chartered accountants. We really have a problem getting information from India, and everybody has a problem, but in this case, the Department of Commerce, they had an auditor number, they had chartered accountants, and they just said, well, it doesn't have a report, the statement's no good. And if the Department of Commerce had a problem... Well, do we have to assume that they said the statement's no good? What they concluded, as I understood it, was that the other statement, the MHPC, which I guess had auditor's notes and all of the schedules, was better, was more reliable. Well... Isn't that right? Well, they said it was more reliable because it didn't have the report. That was the reason. But if you look at the two statements and assuming that our statement, let's say, had an auditor's report, would they have said that the MHPC was better? By our statement, you mean the APIS statement. I'm sorry. Assuming the APIS statement had an auditor's report. I mean, the APIS statement is from a chartered accountant with a number that I assume is a certified number, and it has a full set of schedules, and I think that's a factual mistake. If you look at the APIS auditor's report, you have schedules listed A to P under balance sheet. So what do you want us to do here on appeal? To look at the two reports, the Commerce's judgment was this one is more reliable because it has all of these notes and it has all these schedules and whatever, and we're supposed to, on appeal, say, no, they were equally reliable notwithstanding more notes or more information on one than the other. How are we to do that on appeal? One thing you could do, which I assume, you know, I don't expect you to do, would say, you know, we're right and they're wrong. The APIS is more reliable. But one thing you can do on appeal is you can say, look, it doesn't have an auditor's report. Let's try to see if you could get an auditor's report. You have a charted account. Let the Department of Commerce try to call somebody in India to see if there's an auditor's report. We don't even know if one ever existed. But let's give it another try to get an auditor's report. During the course of this proceeding, the Department of Commerce, a question arose as to information in the EDA data. So what do they do? They write a letter to EDA. The question arose during the course as to the Tribune article. What do they do? They go to the Tribune article to the author of the Tribune article and they ask him questions to go on the record. I'm not saying we don't... I'm sorry. You want to save three minutes? You're well into that time. Yes. We'll give you a full rebuttal. Thank you, Your Honor. Mr. Cassini for the government. May it please the Court. Commerce relied on the best available information here. First, with respect to the EDA data for valuing raw honey really, the only issue that may be of any concern that plaintiffs raised is that somehow the export prices that were placed on the record for raw and processed honey were allegedly lower than the raw value raw honey prices. However, those export prices were only lower after Aswell adjusted those prices using its own adjustments. In fact, the domestic industry adjusted those prices, too, and came up with a number that was higher than the surrogate value used by Commerce. Why not just consider all of this information and factor it in or out according to its strength? Well, that's exactly what Commerce did in determining that the It rejects it outright. Why not account for it? Well, Commerce did, but it found You could account for Kerala as being a region with lower prices. It would require a little more economic analysis, but you don't. You just reject it. But those articles that say the Kerala article were based on much smaller sample sizes or the source of the data Market economy, they ought to be part of the equation, right? I mean, any price factors into the economy, right? Any price might factor into the economy, but a single small volume But why do you exclude the three lowest? I mean, that looks a little suspicious. Well, Commerce put it in and then point out that it's a single point on a larger grid. Well, if the Department of Commerce were to do that It might lower the price a little bit, but that may be appropriate. Well, in this case, it would have any lowering of price would have been infinitesimal. The EDA data was based upon 700 beekeepers, and the other articles provided involved either only one producer or was clearly unreliable because there was absolutely no statement concerning where those prices came from, how the research was done, etc. And, in fact, what Commerce did was that the 75 value was more in the range of values in the Tribune article were higher than the EDA data. So, in fact, the Department of Commerce did not pick the highest value. The value that it ended up with was in the middle of the range of values that were out there. The appendix at 1809 discusses plaintiff's apparent anomaly, as does appendix 714 and 717. Also, the plaintiff's concern about the fact that the EDA data only went for the first half of the period of review is not really relevant here. If you look at the appendix at page 1512, it demonstrates that 86 percent of the honey produced under the EDA data was produced in March through June. It was produced in the spring, and only 5 percent was produced in October. So even if the prices went down by October, that would have been a very, very small change in prices if there were an October EDA price. How do you reconcile the 25 percent higher normal value than the export price? How do they make a profit at all? Well, at page 32 of our brief, the first footnote, we explain the calculational issues that the plaintiffs raised. In fact, they took a simple average of the EDA data to come up with what they viewed the EDA normal value to be. But what the Department of Commerce did was took a weighted average because much more honey was produced during the spring, and we explained that fully in our brief at the first footnote in our briefing. we explained that the plaintiffs claimed that the Department of Commerce somehow skewed the data or came up with a patently unreasonable or absurd determination just doesn't hold up to the actual evidence before the agency. Well, in my they can't have to show that it was patently absurd, they can just show that it wasn't the best and you could have done better, right? Well, what the plaintiffs can show is that Commerce's determination that the EDA data was the best is unsupported by substantial evidence. What I think they're proposing is some combination. So even if EDA is much better clearly as a whole than looking at the other three, my question goes to why it seems to me your answer to why Commerce shouldn't have included or taken account of those articles is simply that, well, harmless error. Even if it had, it wouldn't have made any difference. Well, not at all, because Commerce fully explained why it found those three other data sources to be unreliable and there would be unreliable. But even with respect to the timing, I thought I understood you to just say well, yeah, there was a problem with the second half but it wasn't the entire portion of the second half, it was just one quarter of the second half. So different months have much larger percentages. And the primary type of... So the data, your argument is our data was only off by five percent. No, in fact, our data, if there was a spike in prices, the October 2002 data that Commerce used as the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the that the primary source of the data was only five percent. So the primary source of the data was only five percent. So primary source of was percent. So the primary source of the data was only five percent. So the primary source of the data only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. And the primary source of the was only five percent. So the source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only percent. So the primary source of the data was only five percent. So the primary source of the data was So primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. So the primary source of the data was only five percent. this is a rare slide that I brought with me today, which gave a nudge to someone this case when it happened. And so this was why we brought this slide with me. It just degrades the best information that you had. Why aren't all prices in a market economy reputable data points? If you had all prices across the whole economy, they might be, but what you had in this case was the second largest honey producing region in India. But it was many data points over a long period of time. These other points were single points in time in a single place. And so it would unfairly weight the overall data toward those points. And I see my time is about up. Thank you, Mr. LaBerta. And I think Mr. Moskvac has a couple minutes of rebuttal. Thank you. Thank you, Your Honor. I'd like to make one point on the financial ratios, on the use of LIFO rather than FIFO. The Department of Congress made a decision. They used LIFO, they didn't use FIFO. There was no reason for that decision. They just were doing one, were not doing the other. There was no reason why one was better than the other. But the reason we're making this point, we're not saying they should use FIFO. We're saying if you have to make a choice between LIFO and FIFO in a financial statement without having any reason to use one over the other, there's something wrong with that financial statement. To have dramatic difference, there's some flaw in that statement and that's something the Department didn't consider. They threw out a statement because it didn't have an auditor's report, it didn't have a piece of paper. Instead of the piece of paper we have a stamp and a certification and the possibility if the Department of Commerce would ask, they would be able to get that report if the report exists. Nobody knows. And they accepted a report where they arbitrarily selected LIFO rather than FIFO. It probably would produce a negative profit. The Department picked LIFO because they picked LIFO. There's nothing in the record that the Department made a determination. What's critical to the record was that the Department had to make a very important assumption. There was no pending inventory in that financial statement. There was no material cost in that financial statement. And because of that the Department had to go into the air and say we're going to use LIFO. Now whether it's right or wrong may not be more important as the fact that the Department had to do it. They took a financial statement that had an incredible flaw and they tried to rehabilitate that statement and we had a statement which the only thing that possibly is wrong with our statement is it's lacking one piece of paper saying the auditor is saying it's okay where our statement has a certification and a number. Which you acknowledge that piece of paper might not exist. It may not exist. In which case what do we need to do? Send it back so you can have another audit? No. If the paper doesn't exist that means our statement is complete. If the paper exists let's try to have the Department ask to see if it exists and it will be put on the floor. much. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A9552. Public record. Not legal advice.