[00:00:00] Speaker 00: Soundview Innovations versus Hulu. [00:00:02] Speaker 00: Mr. Ding, whenever you're ready. [00:00:04] Speaker 00: Good morning. [00:00:06] Speaker 02: Good morning, Your Honors, and may it please the court. [00:00:08] Speaker 02: Frederick Ding for Appellant Soundview Innovations. [00:00:11] Speaker 02: The district court's decision rests on three related and incorrect rulings on claimed scope. [00:00:16] Speaker 02: None of those issues is independent of the others. [00:00:18] Speaker 02: As the summary judgment order illustrates, those issues were all woven together explicitly by the district court. [00:00:25] Speaker 00: Yeah, but what is that? [00:00:26] Speaker 00: Just as a practical matter, I don't want to use you. [00:00:28] Speaker 00: We only have limited time. [00:00:30] Speaker 00: What if we were to question the sequence of reading the limitations in that order, and we were to disagree with the judge? [00:00:43] Speaker 00: Is that sufficient to dislodge the summary judgment? [00:00:47] Speaker 02: I think that this court should rule on all of the claim construction issues, because the district court said, I can't figure out if it's a buffer unless it's allocated in the way and in the sequence that I'm construing. [00:00:58] Speaker 02: So I think this court does need to address the other issues underlying. [00:01:02] Speaker 03: Let me try to clarify this. [00:01:04] Speaker 03: As I understood it, there were two separate grounds for granting summary judgment of non-infringement. [00:01:11] Speaker 03: Number one is that. [00:01:13] Speaker 03: The method claim 16 imposes an order on the arrangement of the steps, and they have to be done as they appear, or at least for the first two steps. [00:01:26] Speaker 03: And that's ground one. [00:01:28] Speaker 03: Then ground two is the defendant's products do not allocate a buffer as required by the claim, because the buffer has to be some kind of specialized buffer. [00:01:40] Speaker 03: So isn't it true that you have to win and overturn both of those grounds in order to undo the summary judgment decision? [00:01:50] Speaker 02: Yes, Your Honor. [00:01:51] Speaker 02: We do have to prevail on both this order of steps issue and on whether the accused product can satisfy the allocating a buffer. [00:01:58] Speaker 03: So then to put a finer point on it, if we were to agree with the district court on one of those two grounds, then that would be enough for an affirm of the appeal. [00:02:09] Speaker 03: Is that right? [00:02:12] Speaker 02: I think that is correct. [00:02:13] Speaker 02: If one of those issues, this court affirms the holding of the district court, that would be correct. [00:02:18] Speaker 03: So then therefore, wouldn't it be fair to say we as a court do not need to necessarily get to every single issue? [00:02:26] Speaker 02: The only reason I disagree with that, Your Honor, is that this patent is also asserted in another litigation. [00:02:31] Speaker 02: So beyond just whether this court's constructions will result the dispute in the Cooley case, it may also matter to the litigation against Dish and Sling in the Colorado district. [00:02:42] Speaker 02: So for that reason alone, it would make sense for this court to rule on the correctness of the district court's constructions, even if this court agrees with one and disagrees with another. [00:02:52] Speaker 03: I'm sorry to keep asking questions here. [00:02:54] Speaker 03: Did the Colorado district court adopt the constructions made here by CDCAL? [00:03:00] Speaker 02: The Colorado court has stayed that litigation and requested briefing on both parties from what the effect of this appeal will be. [00:03:08] Speaker 00: So both parties agree that- Wait, I'm sorry to interrupt, but is that the case where there's an ex parte re-exam going on that grew out of that? [00:03:15] Speaker 00: Because there's an ex parte re-exam going on with Dish, right? [00:03:19] Speaker 02: I don't believe. [00:03:21] Speaker 02: There may also be an ex parte re-exam with Dish. [00:03:23] Speaker 02: That's true. [00:03:24] Speaker 02: Involving this PAT? [00:03:25] Speaker 02: Involving different claims of this PAT, not Claim 16. [00:03:28] Speaker 03: There's no re-exam undergoing for claim 16. [00:03:33] Speaker 02: There was a previous re-exam I believe and I think that's concluded. [00:03:37] Speaker 02: So the claim survived? [00:03:39] Speaker 02: The claim survived. [00:03:42] Speaker 02: I'd like to address the order of steps issue, because Your Honor has signaled that we should turn to this issue. [00:03:49] Speaker 02: The applicants were very deliberate when they drafted this patent. [00:03:54] Speaker 02: They were very deliberate in how they worded Claim 16, just as they were deliberate in how they wrote orders of steps into other claims. [00:04:00] Speaker 02: And I think that this is not a pure matter of claim differentiation. [00:04:04] Speaker 02: It's a matter of the applicants knew how to require an explicit order, and they did so. [00:04:10] Speaker 02: And I think that claim 16, there's no doubt that it has to cover the embodiment of figure 7b. [00:04:18] Speaker 02: And that's why I go back to first principles, generally in claim construction, [00:04:22] Speaker 02: it would be incorrect to construe a claim in a manner that excludes the preferred embodiment, or even, in this case, the only embodiment of that claim. [00:04:29] Speaker 03: I think first principles would be to look at the text of the claim itself, right? [00:04:35] Speaker 03: Looking at the claim language. [00:04:37] Speaker 02: I agree, Your Honor. [00:04:37] Speaker 02: We should look at the claim language. [00:04:39] Speaker 02: We need to. [00:04:40] Speaker 02: We need to. [00:04:41] Speaker 02: But as the briefing has illustrated, it's not a clear-cut question on whether the grammar taken in isolation necessarily requires allocating after the request. [00:04:51] Speaker 03: And the reason I mean, the first step says receiving a request for an SM object, right? [00:04:59] Speaker 03: Yes. [00:05:01] Speaker 03: And then just to cut the second step, [00:05:06] Speaker 03: down, it says allocating a buffer to cache at least a portion of said requested SNLogject. [00:05:13] Speaker 03: Right? [00:05:14] Speaker 03: Yes. [00:05:16] Speaker 03: So obviously, there's an antecedent basis there. [00:05:20] Speaker 03: It refers back to the SNLogject in the first limitation. [00:05:24] Speaker 03: But it does more than that. [00:05:26] Speaker 03: It says requested SNLogject. [00:05:30] Speaker 03: And so that's a past tense usage of the word request. [00:05:35] Speaker 03: It's a past tense adjective for SM object. [00:05:41] Speaker 03: And past tense adjectives describe the state or condition of a noun [00:05:50] Speaker 03: resulting from a past action. [00:05:52] Speaker 03: Am I understanding that correctly? [00:05:55] Speaker 02: Past tense adjectives? [00:05:58] Speaker 02: I don't believe that a past participle necessarily requires that that be performed first. [00:06:02] Speaker 02: To give an analogy, if we say that we're allocating a warehouse to store goods ordered by customers, that doesn't mean that the customer has to order the goods before we build the warehouse or select the warehouse. [00:06:14] Speaker 02: It just means that by the time we are storing the goods, that's the second verb in that step. [00:06:20] Speaker 03: I'm not sure I can keep track of your shipment hypothetical. [00:06:27] Speaker 03: How about my hypothetical? [00:06:31] Speaker 03: There's a radio station. [00:06:32] Speaker 03: They play a million different songs. [00:06:35] Speaker 03: I'm a long time listener, but now I'm about to be a first time caller. [00:06:39] Speaker 03: And I'm going to ask for a particular song to be played, Hey Jude by the Beatles. [00:06:47] Speaker 03: You know that song, right? [00:06:49] Speaker 03: So now I'm going to ask for that song to be played. [00:06:54] Speaker 03: So that song has now become a requested song. [00:06:59] Speaker 03: And so if the radio station allocates a particular time period, 3.15 PM, to play that song, [00:07:09] Speaker 03: Um, has it, doesn't the fact that it's a requested song mean that it's something distinct from the other million songs at the radio station? [00:07:20] Speaker 02: There's no dispute that the requested song differentiates it from the other songs. [00:07:26] Speaker 02: And I think necessarily to be a requested song, there had to be a request. [00:07:32] Speaker 02: I think we have to go back to what the allocating step says. [00:07:35] Speaker 02: I'm sorry. [00:07:36] Speaker 02: I'm so fixated on my hypothetical. [00:07:39] Speaker 02: Yes. [00:07:39] Speaker 02: Let me address your hypothetical, Your Honor. [00:07:41] Speaker 02: If we're saying that we're allocating a time slot to play songs requested by a caller, that would be the pertinent hypothetical here. [00:07:49] Speaker 02: We can decide that we're putting in a time slot to play the specific songs that a customer or caller to the radio station is requesting before you actually get the call. [00:07:59] Speaker 02: In particular, if your next claim step is saying, [00:08:01] Speaker 02: now playing the song that was requested, which is exactly what's going on in Claim 16. [00:08:07] Speaker 02: The decision of when we do the allocating is a different point in time than when we actually play the song. [00:08:12] Speaker 02: That's the next action that Claim 16 recites. [00:08:16] Speaker 03: Another concern I have is that [00:08:20] Speaker 03: This claim uses the phrase said requested SM object multiple times. [00:08:25] Speaker 02: Yes, it does. [00:08:26] Speaker 03: And it seems to me this claim would have operated just fine if it had merely said said SM object. [00:08:35] Speaker 03: That is not the case. [00:08:36] Speaker 03: Why did the patent drafter choose to instead of saying said SM object over and over again, it kept saying said requested SM object? [00:08:45] Speaker 02: because the word requested disambiguates which SM object we're referring to. [00:08:50] Speaker 02: The first time SM object appears in claim 16 is in the preamble. [00:08:54] Speaker 02: It appears again with an indefinite article in the first claim step. [00:08:58] Speaker 02: And by using the word requested to refer back to that first step, it's making clear that all those three claim steps that use the said requested SM object are all pertaining to the same SM object. [00:09:09] Speaker 02: So it's only to identify which said SM object we're referring to, as opposed to that of the preamble. [00:09:15] Speaker 00: So wait a minute. [00:09:15] Speaker 00: So SED is doing all the work here? [00:09:17] Speaker 00: I mean, what if it said at least a portion of the requested SM object? [00:09:23] Speaker 00: Would that change the translation? [00:09:24] Speaker 02: If we simply wrote the requested SM object, I don't think that changes it. [00:09:29] Speaker 02: I don't think it's a said versus the issue. [00:09:31] Speaker 00: What work is said doing that? [00:09:34] Speaker 00: I mean, the is more natural a word than said. [00:09:37] Speaker 00: Said has some meaning, correct? [00:09:39] Speaker 02: Of course it does. [00:09:40] Speaker 02: Said means we're linking back up to the previously used antecedent. [00:09:46] Speaker 02: But there's two possible antecedents here. [00:09:48] Speaker 02: There's one in the preamble and there's one in the request step. [00:09:52] Speaker 02: So the two words together, said requested, makes clear. [00:09:54] Speaker 03: The preamble says having a content server which hosts streaming media objects. [00:10:00] Speaker 03: Yes. [00:10:02] Speaker 03: I guess to my concern, the first step says nsmobject and then the rest of the steps refer to requested smobject. [00:10:14] Speaker 03: It seems like it would have been just fine to say said smobject and we would know you'd be referring to nsmobject in the first step and not the content server which hosts streaming media objects in the preamble. [00:10:27] Speaker 02: that very well could have been a way to draft the claims. [00:10:29] Speaker 02: I want to bring this to the Alturas case, which I think is comparable. [00:10:32] Speaker 02: In Alturas, there was a first claim step that said reading a boot selection flag and a subsequent step that said setting said boot selection flag. [00:10:41] Speaker 02: And despite the fact that it used this antecedent reference, this court reversed the district court's construction requiring the order of steps, despite the fact that the only embodiment in the Alturas patent was one that did it in that order. [00:10:54] Speaker 02: Because an antecedent looking at it in that way alone [00:10:56] Speaker 02: is not the test. [00:10:58] Speaker 03: So you're saying Altura stands for the proposition that just because subsequent steps use the term said, that's not enough to lock in an ordering of the claimant dishes, is that right? [00:11:13] Speaker 02: I think that's right, because Altura says we have to then look at what the purpose of the invention is. [00:11:17] Speaker 02: And the same is true of the interactive gift express case. [00:11:20] Speaker 02: In interactive gift express, the court was especially moved by the fact that the specification has an embodiment that performs it in a different order. [00:11:29] Speaker 02: And I think that's what brings us to the figure 7B embodiment in this case. [00:11:33] Speaker 02: During prosecution, that's appendix 925, 926, the applicant said, support for the applicant's invention as claimed in claim 16, [00:11:43] Speaker 02: That's found in that embodiment. [00:11:44] Speaker 02: And so when the district court at Appendix 23 said, if the embodiment contradicts my construction, the embodiment is irrelevant. [00:11:53] Speaker 02: I have to focus only on the language of the claims. [00:11:56] Speaker 02: I think that's overstating the law. [00:11:58] Speaker 02: The law has never said that we need to disregard the embodiments of their intention with construction. [00:12:03] Speaker 02: Rather, the applicants clearly were thinking of how to claim [00:12:06] Speaker 02: that particular embodiment. [00:12:08] Speaker 02: They refer to it in prosecution, and when this court addressed the first appeal in Hulu 1, this court looked at that single buffer embodiment as well, the Figure 7B embodiment. [00:12:17] Speaker 02: And when we look at the description in column 8, it says that buffer's already there. [00:12:22] Speaker 02: It already holds the data for this particular requested SM object. [00:12:26] Speaker 03: It's strange language in the specification at column 8. [00:12:30] Speaker 03: It is assumed that there is data of K1 seconds in the buffer. [00:12:36] Speaker 03: I don't know. [00:12:37] Speaker 03: It didn't say we're allocating a buffer to have only this first portion of an SM object. [00:12:45] Speaker 02: I readily concede that that description does not use the verb allocate by itself. [00:12:50] Speaker 02: But there's no question that that very buffer which is used for the downloading and concurrently retrieving has to be the allocated buffer. [00:12:57] Speaker 02: Because we've been told by this court in the first construction in affirming the disclaimer [00:13:01] Speaker 02: that when we do the downloading and concurring and retrieving step, you must use the allocated buffer. [00:13:07] Speaker 02: And so, therefore, that links those two together. [00:13:09] Speaker 02: There's no doubt buffer B179 has to be the allocated buffer. [00:13:14] Speaker 02: I see I'm into my rebuttal time. [00:13:15] Speaker 00: You are, and I'll always show you rebuttal time. [00:13:17] Speaker 00: Let me ask you one more, just sort of housekeeping it. [00:13:20] Speaker 00: Certainly. [00:13:21] Speaker 00: What happens if we were to vacate the summary judgment? [00:13:25] Speaker 00: I mean, what's left of this case? [00:13:27] Speaker 00: I mean, validity comes back alive. [00:13:29] Speaker 00: There are live validity issues, which I assume have been mooted out by this, but would come back to life. [00:13:35] Speaker 00: What else is left in this case? [00:13:37] Speaker 02: It's just a matter of the trial. [00:13:40] Speaker 02: I actually don't know if there's any live validity issues remaining, but that was one of the counterclaims that would be subject to revival. [00:13:46] Speaker 02: Yes. [00:13:48] Speaker 02: Thank you. [00:13:48] Speaker 02: Thank you. [00:14:08] Speaker 01: Good morning, Your Honors, and may it please the Court. [00:14:10] Speaker 01: My name is Brad Berg, and I'm representing the Appellee Hulu LLC. [00:14:15] Speaker 00: Do you happen to know about your friend referred to the fact that there was an expert IV exam pending, and that's been completed involving Dish? [00:14:23] Speaker 00: Do you happen to know anything? [00:14:24] Speaker 01: No, I'm not involved with that, not much. [00:14:28] Speaker 01: I want to start today with a question that Judge Chen asked about the grammar of Claim 16 and whether the grammar of Claim 16 clearly indicates an order of steps. [00:14:37] Speaker 01: Because I think the grammar of Claim16 and the logic of Claim16 is clear that an order of steps is required. [00:14:44] Speaker 01: Claim16 has two, the first two limitations of Claim16 are site receiving a request for an SM object from one of said plurality of clients at one of said plurality of helper servers. [00:14:55] Speaker 01: And the second step is allocating a buffer at one of said plurality of helper servers to cache at least a portion of said requested SM object. [00:15:04] Speaker 04: The district court construed allocating a buffer as attaching a requested SM object to a buffer and filling the buffer with data for the SM object. [00:15:18] Speaker 04: Your brief doesn't address filling the buffer construction, and you have a footnote on that saying it's irrelevant. [00:15:27] Speaker 04: because the district court didn't find that your technology failed to meet that limitation. [00:15:34] Speaker 04: If we conclude that allocating a buffer doesn't require filling the buffer, wouldn't that also render attaching a requested SM object to a buffer incorrectly construed? [00:15:49] Speaker 01: So I think, as Your Honor noted, we didn't [00:15:56] Speaker 01: challenge this filling portion of a buffer. [00:15:58] Speaker 01: I do think that the filling portion of a buffer is supported by the specification that the judge, I think, has reasoning for why that's included. [00:16:05] Speaker 01: But the key part is the attaching part. [00:16:08] Speaker 01: And it's key in the sense that the judge was recognizing that the allocating a buffer is doing some particular work. [00:16:16] Speaker 01: It's the way that the buffer is associated with a particular SM object. [00:16:21] Speaker 01: And the judge used that attaching language that comes from column 10 as a way of illustrating that feature of allocating a buffer. [00:16:29] Speaker 01: I think that the filling is another thing that happens simultaneously with the attaching, but is a separate piece that the judge saw was happening with the allocating step. [00:16:43] Speaker 01: So I think these are two separate pieces. [00:16:47] Speaker 01: And I think the critical thing here is in [00:16:51] Speaker 01: In Hulu's summary judgment brief, we argued that the allocating step wasn't met by the configuration and the initialization that Soundview points to under the plain and ordinary meaning. [00:17:01] Speaker 01: We don't think a construction is actually required to affirm on the ground that there's no allocation, which we see as separate grounds from the ground that there's no buffer and from the order of steps. [00:17:15] Speaker 03: The district court didn't adopt that rationale though, right? [00:17:18] Speaker 03: Which rationale? [00:17:19] Speaker 03: The rationale, you just seemed to articulate a separate rationale for why you don't believe there's infringement. [00:17:25] Speaker 01: I think there's four independent rationales in the district court's order. [00:17:29] Speaker 01: I think the first one that the district court touches on is the construction of buffer and whether or not the accused caches in the edge servers can be buffers. [00:17:39] Speaker 01: And that comes in two limitations. [00:17:41] Speaker 01: It comes in the allocating a buffer limitation. [00:17:43] Speaker 01: It also comes in the downloading while concurrently retrieving limitation, which this court previously [00:17:50] Speaker 01: construed or affirm the district court's construction that it requires the same buffer to do both. [00:17:54] Speaker 01: So I think the district court's first ground is that buffer ground. [00:17:58] Speaker 01: And then I think there are three separate grounds in the allocating, and you can see that on Appendix 24, where the district court is talking about the allocating limitation. [00:18:10] Speaker 01: And there, the district court starts with a paragraph that says first and talks about the order of steps as one ground. [00:18:19] Speaker 01: Then the district court has a paragraph that says second and talks about the construction, applying the construction of allocating, that it does the attaching and filling. [00:18:27] Speaker 01: And then the district court has a paragraph that says furthermore, the configurations the plaintiff contends constitute allocating a buffer are not based on a request for a particular SM object. [00:18:37] Speaker 01: Rather, the evidence shows the configurations are based on a particular customer. [00:18:43] Speaker 01: I read that part of the district court's order as saying, even setting aside the order of steps, even setting aside the construction of allocating, there's separate language in the claim, the said requested SM object. [00:18:55] Speaker 01: The buffer is allocated for a particular purpose, to cash a portion of said requested SM object. [00:19:02] Speaker 01: And even setting aside those other things, this is an independent ground that the district court found. [00:19:08] Speaker 03: Well, my understanding of the district court's application of his claim construction was that he believed that the term buffer refers to some kind of specialized buffer that can only handle, receive, download, [00:19:34] Speaker 03: data from a particular SM object, is that right? [00:19:38] Speaker 01: Yeah, that's correct. [00:19:39] Speaker 03: And so then all this other analysis kind of launches off of that. [00:19:44] Speaker 03: I agree that... And so now, I mean, I know we started on ordering, but let's just stay here on buffer. [00:19:51] Speaker 03: Sure. [00:19:52] Speaker 03: Let's assume for the moment, I have no idea how he arrived at that conclusion. [00:19:59] Speaker 03: that a buffer as used in the claim must be this specialized unique buffer dedicated to only one function, and that is dealing with only a particularized SM object. [00:20:12] Speaker 03: I mean, that goes against the well-established meaning of what a buffer is. [00:20:18] Speaker 03: A buffer, as I'm sure you would agree, is a well-known term of art here in the world of networks and data transfers. [00:20:28] Speaker 01: So the district court looked at the language of the claim and saw that in the allocating step and in the downloading while concurrently retrieving step, the same buffer was associated with the SM object and indicated that was one reason that this was supported. [00:20:43] Speaker 01: Then the district court looked at the specification and attempted to determine what the plain meaning of buffer is in the context of the specification. [00:20:51] Speaker 01: And the specification is very clear that a buffer is only allocated in response to an SM object. [00:20:56] Speaker 03: Does the specification actually use the word only? [00:21:00] Speaker 03: That's the key word in this question. [00:21:03] Speaker 03: Can it only be storing information with respect to one SM object? [00:21:11] Speaker 03: I didn't see the word only in the specification. [00:21:14] Speaker 04: Could it be other SM objects as well stored on the same buffer? [00:21:19] Speaker 01: I agree with Your Honor that the word only is not used. [00:21:22] Speaker 01: As to your question, Judge Wallach, as to whether it could be other SM objects, the answer is no, and that's because it would undermine the entire purpose of the patent to associate a single buffer with multiple SM objects. [00:21:36] Speaker 01: and the district court notes this, is all about the addressing what the patent calls the heterogeneity of requests, meaning requests for the same SM object from multiple clients at different times. [00:21:51] Speaker 01: And so in that context, the invention of the patent is how do we use the existing storage, helper servers existed, helper servers had memory, to most efficiently [00:22:06] Speaker 01: get these SM objects that are being requested, the same SM object to the client. [00:22:11] Speaker 04: So supposing the user is a news organization, and the SM object is a portion of a speech by the president. [00:22:22] Speaker 04: And they're pretty confident that lots of people are going to, say, read about it and then want to see it. [00:22:31] Speaker 04: And so they in advance store that SM object in the buffer waiting for the rush of all the customers. [00:22:46] Speaker 04: Why isn't that a logical approach? [00:22:50] Speaker 01: So unless I misunderstand your Honor's question, I think we're talking about the order of steps now and whether or not the buffer can be pre-allocated. [00:22:57] Speaker 01: In your example, there's not multiple SM objects stored in the same buffer, correct? [00:23:02] Speaker 01: There's just one. [00:23:03] Speaker 01: So I think what that goes back to is the plain language of Claim 16 and the said request it. [00:23:10] Speaker 01: Is it technologically feasible or plausible to pre-allocate a buffer? [00:23:16] Speaker 01: I think that is a technologically plausible explanation. [00:23:22] Speaker 01: But it's not supported by the plain language of the claims with the said requested SM object explicitly referring back to... So now we're at the sequencing again. [00:23:32] Speaker 00: What do you say to figure 7b, which figures prominently in your friend's argument? [00:23:38] Speaker 01: Yeah, so that's what I was just going to get to. [00:23:40] Speaker 01: Once we see that there's a plain meaning in the claim language, then we look to the specification to see if anything would have this court deviate from what the plain language of the order of steps is. [00:23:53] Speaker 01: And my friend Mr. Ding has pointed to Figure 7B and talked about Figure 7B. [00:24:01] Speaker 01: And the important thing about Figure 7B, as Mr. Ding acknowledged, is that it simply doesn't talk about allocating. [00:24:08] Speaker 01: There's a buffer, and it's talking about a particular aspect of the invention, data transfer rate control, which is how you manipulate the data once it's in the buffer. [00:24:18] Speaker 01: And as a result, it focuses specifically on that issue of how you manipulate the data, how you send it out, how you get it back in, how you get it to the client as quickly as possible. [00:24:28] Speaker 01: But it never talks about allocating. [00:24:29] Speaker 01: It never uses the word allocating. [00:24:31] Speaker 01: And so someone looking at this embodiment and trying to understand claim 16 [00:24:36] Speaker 01: would know that you've got to go somewhere else for allocating. [00:24:39] Speaker 01: It doesn't make any sense to say the entirety of the written description for claim 16 must be in the data transfer rate control embodiment, where a step of claim 16 is specifically allocating, and the data transfer rate control never talks about allocating. [00:24:54] Speaker 01: And so the person with skill in the art would look at the other examples in the specification, which do explain allocating, and in particular, I think, would go to column six, where there's a paragraph that says, [00:25:06] Speaker 01: This is in line 36 of column 6. [00:25:11] Speaker 01: The method of allocating a ring buffer in the memory of an HS is now described with reference to figure 4 and goes through in that paragraph in specific detail how a buffer is allocated. [00:25:22] Speaker 03: So it's a different embodiment, I guess, is the concern. [00:25:28] Speaker 03: There's a few different embodiments you would expect. [00:25:30] Speaker 03: And then we try to look at all of the claims of this patent and see which claims most neatly track with the different disclosed embodiments. [00:25:39] Speaker 03: And it really feels like claim 16 [00:25:43] Speaker 03: in terms of best fit would be this column 8 on data transfer rate control, whereas the claims that come before claim 16 really track this ring buffer with the client request aggregation scheme, and then the remaining claims really go after this RTSP embodiment. [00:26:03] Speaker 01: Yeah, so absolutely the [00:26:06] Speaker 01: The third and fourth limitations of claim 16, the downloading while concurrently retrieving and the adjusting a data transfer rate limitations come from the embodiment in column 8, where figure 7B is. [00:26:20] Speaker 01: But the earlier steps, I don't think, come from that embodiment and can't come from that embodiment because it doesn't describe allocating. [00:26:27] Speaker 01: And I think I want to push back on the idea that these three embodiments are three separate embodiments. [00:26:32] Speaker 01: The patentee is pretty clear in column four that these embodiments are techniques meant to be implemented by a helper server together. [00:26:41] Speaker 01: If we go to column four, starting at line 53, it says, [00:26:49] Speaker 01: To overcome these drawbacks and further enhance the ability of existing caching systems to properly scale in terms of an object size, number of supported streams, illustrative embodiments of the present invention advantageously combine three methods which are implemented via novel system architecture to enhance existing caching systems. [00:27:09] Speaker 01: And then it lists proxy caching, client request aggregation, which is the ring buffer, and data transfer rate control, which is the column eight disclosure. [00:27:17] Speaker 01: And so I think looking at this [00:27:20] Speaker 01: combination of embodiments language, it's perfectly natural to look at Claim 16 and say, I get the written description for concurrently download or downloading while concurrently retrieving and adjusting a data transfer rate from column 8 from the 7B embodiment, and I get the written description for [00:27:41] Speaker 01: under what circumstances and when an allocation of a buffer happens from looking at the other embodiments because these are all to be implemented together and in fact every time the buffer talks about or the patent talks about allocating a buffer whether it's in the ring buffer embodiment or in the RTP RTSP embodiment which are [00:27:58] Speaker 01: the two places where it talks about allocating a buffer, it's always in response to a request for an SM module. [00:28:05] Speaker 00: I need to cut you off, just because your time's almost up. [00:28:07] Speaker 00: I have one further question, which is that the argument made in the briefs, I think, is that there's a, I don't know if we call it claim differentiation, but in statutory construction, as you know, we look to other things that are said and how they are said. [00:28:24] Speaker 00: And you've got claim 16 where the, [00:28:26] Speaker 00: It's very fastidious in terms of outlining what the steps are. [00:28:30] Speaker 00: It's very specific. [00:28:31] Speaker 00: Claim 13. [00:28:32] Speaker 00: Claim 13, I'm sorry. [00:28:33] Speaker 00: So why shouldn't we look at that in connection with 16 and what they meant, whether they meant to have sequencing in the order? [00:28:44] Speaker 01: Yeah, I think that the work done by said requested SM object is the same work that's done in Claim 13. [00:28:51] Speaker 01: by, I think it's upon... Upon receiving said first request. [00:28:55] Speaker 01: Upon receiving said first request. [00:28:58] Speaker 01: I think the work done by those two is equivalent. [00:29:00] Speaker 01: I don't see any reason why said requested SM object is less clear than upon receiving said request. [00:29:09] Speaker 01: I think the panning... It's different, certainly, but I don't see any reason that that would make us say, [00:29:19] Speaker 01: or should compel this court to say that the language in Claim 16 itself isn't clear. [00:29:25] Speaker 03: What about the prosecution history statement where for support for Claim 16, they pointed right at Figure 7B? [00:29:33] Speaker 01: Yeah, that's a great question, Your Honor. [00:29:37] Speaker 01: In the prosecution history part that's cited in the appendix, this is after the claim 16 had been rejected over the money reference that this court talked about at length in the previous appeal. [00:29:49] Speaker 01: And in response to that rejection, the applicant added the downloading while concurrently retrieving step and explained that that downloading while concurrently retrieving was the invention. [00:30:00] Speaker 01: And then points to column 8 as providing support for [00:30:06] Speaker 01: the invention, which was the downloading and concurrently retrieving step that was added. [00:30:09] Speaker 01: I don't read the prosecution history anywhere as saying that all the written description for claim 16 needs to come from that column, but rather the added portion that was added in response to the rejection comes from that column. [00:30:21] Speaker 01: And that's where the written support for that limitation comes from. [00:30:23] Speaker 03: And final question. [00:30:25] Speaker 03: If we were to agree with the district court that the claim imposes an order, a sequence of the claim limitations, but then also say we reject that [00:30:36] Speaker 03: the construction of buffer and the court's application of the buffer to the accused products. [00:30:44] Speaker 03: Would you be fine with that? [00:30:45] Speaker 03: Because we would ultimately be affirming. [00:30:47] Speaker 01: I mean, I think that the buffer construction should be affirmed. [00:30:51] Speaker 01: But obviously, I think any one of these foregrounds is independent. [00:30:56] Speaker 01: And affirming on any one of them is sufficient for summary judgment here. [00:31:00] Speaker 01: Thank you. [00:31:01] Speaker 01: Thank you. [00:31:13] Speaker 02: Thank you, Your Honors. [00:31:16] Speaker 02: The common premise underlying all of the district court's constructions is this idea that you must use a dedicated buffer that can hold only one SM object. [00:31:25] Speaker 02: And that is on its face not supported by the intrinsic evidence. [00:31:30] Speaker 02: It's not the type of limitation that we can read into a claim. [00:31:33] Speaker 02: And on this point, I think the case GE Lighting versus Agilite is very instructive. [00:31:37] Speaker 02: There, too, the patent claim used a well-understood term [00:31:41] Speaker 02: from which the inventors did not depart from the customary meaning as used in the art. [00:31:46] Speaker 02: And here, as the Hulu 1 opinion recognized, there is already a well-understood meaning of the word buffer. [00:31:52] Speaker 02: And nothing in the specification teaches that the buffer needs to be limited in the manner that the district court and Hulu have proposed. [00:31:58] Speaker 02: But more than that, once the district court began with this incorrect premise that a buffer must be so dedicated, [00:32:05] Speaker 02: That forced its error in understanding what the allocating a buffer limitation means. [00:32:11] Speaker 02: And it forced the error on the order of steps. [00:32:14] Speaker 02: And I think that Judge Wallach's popular speech example illustrates exactly what's going on. [00:32:19] Speaker 02: As the 796 patent, it's another one of the inventor's inventions, shows you can pre-allocate buffers for popular content. [00:32:26] Speaker 02: That's a very logical thing that was known at the time. [00:32:29] Speaker 02: And it's exactly what you would do with help reserves. [00:32:31] Speaker 02: The 796 patent has a very similar description. [00:32:33] Speaker 02: uses two of the same figures as the 2 and 3 patent. [00:32:36] Speaker 02: And that shows us how a person of ordinary skill understood the allocation of buffers. [00:32:41] Speaker 02: I do think this case goes back inevitably to the figure 7b embodiment. [00:32:47] Speaker 02: Mr. Byrd's argument assumes that it's OK for a claim like claim 16, that the applicants were clearly thinking of how to cover that embodiment, to nevertheless, that embodiment would not infringe the claim. [00:33:00] Speaker 02: And that's something that we don't usually see in claim construction. [00:33:03] Speaker 02: Cases like Interactive Gift Express tell us that if there's an embodiment that performs it in a different order, we should try to read the claims in order to reach that embodiment. [00:33:14] Speaker 02: And Mr. Berg agreed that in that embodiment, data for that particular SM object was already being held in the buffer of B179, which means, therefore, it must have already been allocated to cache that particular SM object. [00:33:29] Speaker 02: And I think from those premises, we get to the conclusion that [00:33:33] Speaker 02: Claim 16 as distinct from claim 13. [00:33:36] Speaker 02: Claim 13 recites an explicit order of steps. [00:33:39] Speaker 02: And claim 16 as distinct from claim 17, where claim 17 says each buffer is associated with a single SM object. [00:33:45] Speaker 02: Claim 16 is different. [00:33:47] Speaker 02: The applicants meant to write claim 16 differently. [00:33:50] Speaker 02: They used very particular language. [00:33:52] Speaker 02: And it's a very simple claim. [00:33:53] Speaker 02: The language on its own doesn't require a dedicated buffer. [00:33:57] Speaker 02: And the language on its own doesn't specify an order of steps. [00:34:01] Speaker 02: From those issues, the three separate constructions that the district court imposed on the meaning of the buffer, on the order of the steps, and on the requirement that allocating requires filling with data, all three of them are unsupported by the record. [00:34:14] Speaker 02: And so this court should reverse. [00:34:16] Speaker 02: Thank you. [00:34:17] Speaker 02: Thank you. [00:34:17] Speaker 00: We thank both sides.