[00:00:01] Speaker 03:
All right, the next case before the court is Visual Memory versus NVIDIA Corp.

[00:00:07] Speaker 03:
162254, an appeal from the District Court and the District of Delaware.

[00:00:37] Speaker 03:
Mr. Weimblat, you want three minutes for rebuttal?

[00:00:40] Speaker 03:
Yes, please.

[00:00:44] Speaker 03:
Okay.

[00:00:44] Speaker 03:
Counsel, you ready?

[00:00:46] Speaker 03:
Yes, Your Honor.

[00:00:47] Speaker 03:
You may begin.

[00:00:58] Speaker 00:
May it please the Court, Rich Weimblat of St.

[00:01:00] Speaker 00:
Melissa Weimblat on behalf of Appellant Visual Memory.

[00:01:04] Speaker 00:
The 740 patent claimed inventions

[00:01:07] Speaker 00:
improve computers as tools.

[00:01:09] Speaker 00:
They improve the functioning of computer memory systems, which improves computers' performance.

[00:01:14] Speaker 00:
And they do not touch using a generic computer as a tool to carry out a task.

[00:01:20] Speaker 00:
The district court overgeneralized and oversimplified the claims when deciding NVIDIA's Rule 12b6 motion.

[00:01:27] Speaker 03:
But my problem is that I look at claim one, and it's kind of hard to overgeneralize that claim.

[00:01:34] Speaker 03:
I mean, it's about as broad as you could possibly be.

[00:01:38] Speaker 03:
Now I know you point to a lot of things in the written description that you think inform that claim, but tell me what you're actually claiming there.

[00:01:47] Speaker 00:
Okay.

[00:01:47] Speaker 00:
So claim one is an improvement upon prior art memory systems because in claim one there is a programmable operational characteristic that's defined by the computer based on the type of processor that determines

[00:02:05] Speaker 00:
what type of data, whether that be code or non-code, gets stored in the cache.

[00:02:09] Speaker 00:
And when you look at the specification, which criticizes prior art in column two, beginning at line 39 and going through line 56, one of the problems with these prior memory systems is they were designed for a specific processor or a family of processors.

[00:02:28] Speaker 00:
And when that memory system was installed in a computer at the

[00:02:32] Speaker 00:
that did not have the ideal processor for which it was made, design trade-offs occurred, and memory access time was reduced.

[00:02:39] Speaker 00:
So what claim one allows is one memory system that can be installed with multiple processors where it uses the programmable operational characteristic to determine whether, for example, instructions, which would be code, or operands, non-code, get stored in the cache.

[00:02:56] Speaker 04:
So just two questions.

[00:02:58] Speaker 04:
One is, is your position the preamble?

[00:03:00] Speaker 04:
It's wheat.

[00:03:01] Speaker 04:
It has wheat.

[00:03:01] Speaker 04:
and adds to the claim and should be interpreted and understood as part of the elements to claim, right?

[00:03:07] Speaker 00:
Yes, Your Honor.

[00:03:07] Speaker 04:
Okay.

[00:03:08] Speaker 04:
And then do I understand you to be interpreting the language type of data to be code and non-code or is it broader than that?

[00:03:18] Speaker 00:
Well, it's defined in column three, line 38 to 42.

[00:03:28] Speaker 00:
as used here in the term data refers to information that includes both code, data instructions, and non-code data.

[00:03:33] Speaker 00:
And elsewhere throughout the patent, it also says the same thing about instructions being code.

[00:03:43] Speaker 04:
So it's your position that claim one should be understood when it says type of data be code and non-code, and that that's expressly defined in the specification?

[00:03:52] Speaker 00:
That is correct.

[00:03:54] Speaker 00:
Yes, Your Honor.

[00:03:55] Speaker 00:
So then if you look at claims six and eight, which are very different from claim one, claim six requires the memory system having a programmable operational characteristic defined through configuration by the computer based on the type of process to determine what type of data, again, for example, code or non-code, is stored in the main memory and also requires open page bias.

[00:04:17] Speaker 00:
The specification explains that this increased system performance at column five, lines 22 to 30.

[00:04:24] Speaker 00:
Claim 8 requires a memory system that has a plurality of caches, three different programmable operational characteristics based on characteristics of the processor, which defines the requirements about two caches and the main memory system stored data.

[00:04:37] Speaker 00:
This, again, improved system performance, and that improvement is also explained at column 5, line 6 to 8.

[00:04:46] Speaker 00:
This is the record evidence that's before the court.

[00:04:48] Speaker 00:
We have statements in the specification criticizing the prior art, and throughout its

[00:04:53] Speaker 00:
description of the solution to the problem to be solved, there are statements about how the claimed solution improves upon the prior art.

[00:05:02] Speaker 00:
All of this was ignored by the district court.

[00:05:04] Speaker 03:
Is it your position that these claims are directed to something that's not abstract and therefore this would fall within the type of analysis done in Enfish, or is it your claim that there's something new and novel

[00:05:23] Speaker 03:
that's added, and that we're talking about a step two.

[00:05:27] Speaker 00:
Well, we look at step one of Alice first, and Enfish tells us that if the claims improve the function of a computer system itself, then it can be patent eligible.

[00:05:37] Speaker 02:
But Enfish tells us how to do it.

[00:05:39] Speaker 02:
And I know you've been trying to explain to me how these claims, how to do it, but at least in claim one, all I see is that there's a programmable operational characteristic

[00:05:50] Speaker 02:
but it doesn't tell us how to do that programmable work.

[00:05:56] Speaker 02:
So your honor, both Enfish and... Look, Enfish has a specific algorithm in it that says use this for the database and it will improve operations.

[00:06:10] Speaker 02:
I can point to and see in the specification and it limits it because it's a means plus function claim.

[00:06:16] Speaker 02:
This to me, and tell me how I'm getting this wrong, says

[00:06:19] Speaker 02:
Here's this configuration of what I think is routine equipment, right?

[00:06:25] Speaker 02:
It's a computer, a bus, a cache, and stuff like that.

[00:06:27] Speaker 02:
You didn't invent any of that stuff.

[00:06:31] Speaker 02:
You program it in a certain way, and that makes it function better.

[00:06:35] Speaker 02:
If you told me how you programmed it, that seems to bring it into Infish.

[00:06:40] Speaker 02:
But you don't, I don't see how this tells me how you programmed it to come up with a better function, a better computer like Infish.

[00:06:50] Speaker 00:
Column four, line 30 to 43, explains how the internal cache can be coded.

[00:06:58] Speaker 00:
The write buffer cache programmable operational characteristic as to whether it buffers data solely from a bus master other than a system processor is at column five, lines 48 to 50.

[00:07:08] Speaker 00:
As for the main memory type of data, code or non-code, for which main memory 12 through controller 22 provides a bias,

[00:07:17] Speaker 00:
is also a programmable operational characteristic.

[00:07:19] Speaker 02:
And that's at... What does that mean?

[00:07:21] Speaker 02:
I mean, here's the problem.

[00:07:23] Speaker 02:
And I apologize in advance, but I read all of this.

[00:07:27] Speaker 02:
It's not a long patent or a specification.

[00:07:28] Speaker 02:
I read all of it.

[00:07:30] Speaker 02:
Those words are difficult.

[00:07:31] Speaker 02:
So you need to explain to me what that's telling me how the computer is operating differently.

[00:07:39] Speaker 00:
OK.

[00:07:39] Speaker 00:
So the way it's operating differently is the programmable operational characteristic, when you look at the specification and the description, which

[00:07:47] Speaker 00:
sets forth the different examples.

[00:07:49] Speaker 00:
It's a feature that can be changed in response to the type or a character.

[00:07:53] Speaker 02:
No, no, no.

[00:07:53] Speaker 02:
I get that.

[00:07:54] Speaker 02:
I get that.

[00:07:54] Speaker 02:
That seems to me the gist of claim one.

[00:07:57] Speaker 02:
I'm sorry.

[00:07:57] Speaker 02:
I used the word gist.

[00:07:58] Speaker 02:
I know that's controversial.

[00:08:02] Speaker 02:
The basis of claim one is you have the system that instead of the old art, which is programmed in only one way and can't be changed,

[00:08:10] Speaker 02:
This is a more flexible system and can be programmed in different ways, but it doesn't tell me how.

[00:08:16] Speaker 02:
I mean, the idea that you can just program something in different ways to me seems relatively abstract.

[00:08:24] Speaker 00:
The CDL code that's attached to the patent sets forth an exemplary environment for how to do it.

[00:08:31] Speaker 00:
And it even says in the specification that CDL is a high level hardware description language and figure two, this is in column six, it shows

[00:08:41] Speaker 00:
the structure and mode of operation of each of the modules defined through the CDL listing.

[00:08:47] Speaker 02:
But you're not agreeing to be bound by that code.

[00:08:51] Speaker 02:
I mean, that's why INFISH was successful.

[00:08:54] Speaker 02:
They limited their invention to a specific algorithm that was a concrete improvement over a database.

[00:09:02] Speaker 00:
McRoe tells us that DINA's claims are allowed.

[00:09:05] Speaker 00:
So if you look at the specification,

[00:09:08] Speaker 00:
and the CDL code, that's providing the preferred embodiment that would teach one of ordinary skill in the art how to do it.

[00:09:15] Speaker 00:
Now, whether you're saying that the claim could be overbroad to a point that it's invalid based on prior art, that's for another day that record's not before us.

[00:09:23] Speaker 00:
But as for whether it teaches one of ordinary skill in the art what a programmable operational characteristic is, we believe that it does.

[00:09:31] Speaker 00:
And there's no evidence suggesting otherwise.

[00:09:34] Speaker 02:
And where in the specification does it show that to me?

[00:09:38] Speaker 00:
Starting at line one, going to line 16.

[00:09:41] Speaker 00:
It's the first two paragraphs.

[00:09:58] Speaker 00:
Your Honor, NVIDIA does not challenge that there were problems with the prior memory systems, as I previously explained.

[00:10:05] Speaker 00:
And NVIDIA also does not challenge that the claims seek to address this technological problem.

[00:10:10] Speaker 00:
Instead, they just stood before the court and the district court, and even in the briefing now, they just says, there's nothing invented there.

[00:10:17] Speaker 04:
I understand that the district court found that the abstract idea that your claims are drawn to is categorical data storage.

[00:10:26] Speaker 04:
Do you think that your claim is directed to that abstract idea or is it your position that is directed to something where, what is the abstract idea?

[00:10:37] Speaker 04:
And even if that is the case, do you have a step two analysis where you can say, well, even if the claim is directed to pedagogical data storage, there's a technological improvement here?

[00:10:49] Speaker 00:
So first of all, we don't think that there's an abstract idea here.

[00:10:52] Speaker 00:
We don't understand how this is categorical data storage, because it's actually one step removed from that.

[00:10:57] Speaker 00:
This is what's actually setting up the memory system for how it works.

[00:11:00] Speaker 00:
Now, if you're then going to ignore that, like the district court did, and say, OK, what's a memory system do?

[00:11:06] Speaker 00:
It's determining where code and non-code get where it's sent.

[00:11:12] Speaker 00:
That is storing data and setting forth where.

[00:11:15] Speaker 00:
But you have to look at how is the memory system being programmed in order to do that?

[00:11:19] Speaker 00:
And that's what the court ignored when deciding the motion to dismiss.

[00:11:25] Speaker 00:
I see you have an eclipsical look.

[00:11:26] Speaker 00:
Basically, what I'm saying here is the district court ignored how the memory system is set up so that it could be used with multiple processors.

[00:11:38] Speaker 04:
So for that reason, we don't think it's every... It's not just categorical data storage.

[00:11:43] Speaker 04:
Your system instead is directed to the idea of having a processor

[00:11:48] Speaker 04:
based on the different type of processor, it will dictate how the memory operates, including whether it's going to store code or non-code.

[00:11:55] Speaker 00:
Exactly.

[00:11:56] Speaker 00:
So it's not as simplistic as what the district court says.

[00:11:59] Speaker 00:
Now turning to step two, yes, we also believe that there's an inventive step here, because even if you were to look at the district court's abstract idea of categorical data storage, and then you look at the claims, we don't think there's any way that these claims will

[00:12:16] Speaker 00:
encompass the idea of putting a book on a library shelf, as the district report said.

[00:12:20] Speaker 00:
We still don't understand that analogy when you look at all these different elements of the claims.

[00:12:24] Speaker 00:
Claim six and claim eight in particular have all of these requirements as to what the program operational characteristics are.

[00:12:30] Speaker 04:
Can you go into specifics on claim six and eight and how it differs?

[00:12:33] Speaker 00:
Sure.

[00:12:33] Speaker 00:
So claim six has a programmable operational characteristic that determines the type of data stored in the main memory and it also requires an open page bias.

[00:12:44] Speaker 00:
An open-page bias is essentially when the memory calls code and then next it calls non-code, it will automatically put code back in the place where it previously found the non-code.

[00:12:58] Speaker 00:
So it makes the computer run more efficient that way.

[00:13:02] Speaker 00:
We still don't know how a librarian putting a book on a library shelf does that.

[00:13:06] Speaker 00:
The library analogy is too broad.

[00:13:09] Speaker 00:
Same thing with NVIDIA's analogy of attorneys put books

[00:13:12] Speaker 00:
at a close proximity to where they're sitting at their desk.

[00:13:15] Speaker 00:
It doesn't do open page bias.

[00:13:16] Speaker 00:
It doesn't have a programmable operational characteristic.

[00:13:20] Speaker 00:
Claim 8 has a plurality of caches and the main memory having programmable operational characteristics.

[00:13:28] Speaker 00:
So that's even further removed and more involved than Claim 6, meaning further removed from the library analogy.

[00:13:36] Speaker 03:
You're into your rebuttal time.

[00:13:37] Speaker 03:
You want to save the rest of it?

[00:13:38] Speaker 00:
Yes, Your Honor.

[00:13:39] Speaker 00:
Thank you.

[00:13:53] Speaker 01:
Your Honor, very briefly, I threw my back out this weekend.

[00:13:55] Speaker 01:
So if you see me grimace, it's not the court's question, which we welcome.

[00:14:00] Speaker 01:
It might be.

[00:14:02] Speaker 01:
Let me briefly address the points that counsel made.

[00:14:05] Speaker 01:
He pointed to examples in column four and column five regarding the 386 and 486 processors.

[00:14:13] Speaker 01:
And in column six, it says, nor is the invention limited to applications employing the 386 or 486.

[00:14:19] Speaker 01:
So that helps the claims not at all.

[00:14:22] Speaker 01:
He pointed to the CDL code, and I would respectfully submit, if you can read it, and I have colleagues who can, that there are no algorithms in there and it merely describes data categories.

[00:14:33] Speaker 01:
But it also says that's a preferred model.

[00:14:35] Speaker 03:
Isn't that a written description or an enablement problem and not really a 101 question?

[00:14:40] Speaker 03:
It seems to me we're collapsing all of these cases into 101 when they don't really belong there.

[00:14:45] Speaker 01:
You know, your question is an excellent one, Your Honor, but the reason it isn't is because the focus should be on the claims, of course, read in light of the specification.

[00:14:54] Speaker 01:
And what it says in column six is the CDL listing defines a preferred embodiment.

[00:14:58] Speaker 01:
So a preferred embodiment details can't help us if there's not sufficient detail in the claims themselves.

[00:15:05] Speaker 03:
Well, the claim would cover the preferred embodiment.

[00:15:08] Speaker 01:
Just because the claim may cover a preferred embodiment that lists details, that's very distinct from ENFISH, for example.

[00:15:15] Speaker 01:
The ENFISH gave us the how.

[00:15:17] Speaker 01:
What's missing from this patent is the how.

[00:15:19] Speaker 03:
The reason ENFISH provided... So you're saying that 101 now becomes a written description or enablement?

[00:15:24] Speaker 01:
No, Your Honor.

[00:15:25] Speaker 01:
No, Your Honor.

[00:15:26] Speaker 01:
What I'm saying is that because ENFISH interpreted the claim correctly as a 112-6 limitation, they then read in a four-step algorithm.

[00:15:36] Speaker 01:
if there was some basis here.

[00:15:37] Speaker 03:
But if our job is to say what is the claim directed to, then we're supposed to read the claim and we're supposed to read the claim in light of the specification, right?

[00:15:50] Speaker 01:
Absolutely.

[00:15:51] Speaker 02:
But that's different than... But if there's an embodiment in the specification, and I can't read this either, but if there's an embodiment in the specification, that actually is a new memory structure and it tells you how to create it.

[00:16:06] Speaker 02:
Why isn't that?

[00:16:08] Speaker 01:
The reason it's not, Your Honor, is because that's not what's being claimed.

[00:16:12] Speaker 01:
Instead, what we've got is a detailed CDL listing that, based on our understanding, doesn't have those, but regardless is disclaimed by the specification as providing any limitation.

[00:16:24] Speaker 03:
Well, we've got case after case after case that says we're not supposed to read a claim as excluding preferred embodiment.

[00:16:34] Speaker 01:
Well, that's true, Your Honor, but the analysis, the Alice analysis here, requires us to go through those two steps.

[00:16:41] Speaker 01:
And just because we have a claim that the computer operation is being improved, Alice itself says that that's no basis to grant patent eligibility.

[00:16:51] Speaker 03:
So the fact that... No, that's not what Alice says.

[00:16:54] Speaker 03:
Alice says if all you're doing is claiming a computer,

[00:16:58] Speaker 03:
that will be programmed to do a certain thing, if what it's doing is abstract, an abstract idea, then that's not enough.

[00:17:05] Speaker 03:
But that's not what they're claiming here.

[00:17:08] Speaker 01:
It is what they're claiming here, Your Honor.

[00:17:10] Speaker 01:
What I understood and read Alice to say is that it doesn't make patent eligible any patent which just purports to improve the functioning of a computer.

[00:17:18] Speaker 01:
And that's what we have here.

[00:17:19] Speaker 04:
Just because we have an elect... But Alice is slightly different if you think about it in terms of it's a situation where you're saying, hey,

[00:17:26] Speaker 04:
In the past, you had this particular business method, and this is the way you would operate it.

[00:17:30] Speaker 04:
And now I'm going to take that, and it doesn't have to be a business method, but a concept and a technique, a method, and I'm going to throw it on a computer now.

[00:17:38] Speaker 04:
It's going to be faster.

[00:17:40] Speaker 04:
It's going to be the work of better.

[00:17:42] Speaker 03:
So Alice wasn't looking at improving the computer.

[00:17:45] Speaker 03:
Alice was looking at using a computer to improve the activity.

[00:17:49] Speaker 01:
I'm just suggesting that that's part of what's in ALICE as it says, even though ALICE was about a business method that was computerized, that just because you claim it improves the computer doesn't make it patent eligible.

[00:17:59] Speaker 01:
Here, Your Honor, if you look at what's being claimed, what's being claimed is a prior art sequence of generic computer elements, no dispute about that.

[00:18:08] Speaker 01:
And then we have this programmable operational characteristic, not really defined.

[00:18:12] Speaker 01:
But if you look at column two,

[00:18:14] Speaker 01:
What it says is, in the past, most computer systems have been designed on the basis of a particular type of post-processor.

[00:18:20] Speaker 01:
And in fact, that's exactly what counsel says in their brief.

[00:18:23] Speaker 01:
In their brief, they say, in the past, at page 15, most memory systems were designed for use with a specific microprocessor or perhaps a family.

[00:18:32] Speaker 01:
So the, the alleged invention, this programmable operational characteristic, what it does is describe in the patent and in the brief as part of what was known.

[00:18:41] Speaker 01:
So then the question is,

[00:18:43] Speaker 01:
Where is the potential arguable inventive step here?

[00:18:46] Speaker 01:
It's in the programmable operational characteristic.

[00:18:49] Speaker 01:
What tells us nothing on how to go about doing that?

[00:18:52] Speaker 04:
Can I ask you some questions about that?

[00:18:54] Speaker 04:
Of course.

[00:18:54] Speaker 04:
OK.

[00:18:54] Speaker 04:
So the programmable operational characteristic, as I understand it, makes it so the memory works one way or another based on the type of processor.

[00:19:06] Speaker 04:
And as you read the specification, it's saying that

[00:19:10] Speaker 04:
You know, you're right.

[00:19:11] Speaker 04:
In the past, a particular memory might store particular data to work with a particular processor.

[00:19:16] Speaker 04:
But here, they're going to have a memory that can work with multiple different kinds of processors.

[00:19:20] Speaker 04:
And so the idea is that this programmable characteristic will indicate how the memory is supposed to work for a particular type of processor.

[00:19:29] Speaker 04:
Why isn't that improving either the memory or the computer system as a whole?

[00:19:34] Speaker 01:
Yeah, the reason, Your Honor, is because it's not telling us how to go about it.

[00:19:39] Speaker 01:
You know, again, it's like the Capital One case.

[00:19:41] Speaker 01:
We have an interactive interface.

[00:19:43] Speaker 01:
But there was no explanation of how to go about doing it.

[00:19:46] Speaker 04:
Do you think one of ordinary scaling art doesn't know how to make go about doing it?

[00:19:49] Speaker 01:
Yes, Your Honor.

[00:19:50] Speaker 04:
But why isn't that, as Judge O'Malley would say, a 112 problem?

[00:19:55] Speaker 01:
Because we have to look to the claim to determine whether there's something here that's inventive that's being added.

[00:20:01] Speaker 01:
And if all that's being added is a black box functional step that doesn't go about describing the how,

[00:20:08] Speaker 01:
Then we're in the same place as a variety of cases, including dealer track, which says you have to specify how the hardware is programmed to perform the steps.

[00:20:17] Speaker 04:
But those cases like dealer track are all about an idea.

[00:20:21] Speaker 04:
As I said before, you've got an abstract idea, and all of a sudden you're going to automate it on a computer.

[00:20:25] Speaker 04:
You've got an idea.

[00:20:26] Speaker 04:
And you're going to put it out and say, but it doesn't change the way the computer operates.

[00:20:30] Speaker 01:
Your Honor, there's two ways to look at this.

[00:20:33] Speaker 01:
One is that what we're dealing with is merely categorical data storage.

[00:20:37] Speaker 01:
That's a broad,

[00:20:38] Speaker 01:
That's what the district court found.

[00:20:40] Speaker 01:
We believe that's correct.

[00:20:41] Speaker 01:
Can you explain that?

[00:20:43] Speaker 03:
Sure.

[00:20:44] Speaker 03:
I do have to say that I was a little perplexed by this concept that it's like putting a book on the library shelf.

[00:20:51] Speaker 01:
So what it's like, Your Honor, is it's saying they provide two categories of data.

[00:20:58] Speaker 01:
Code, which is computer instructions, and both code and non-code, which means everything.

[00:21:03] Speaker 01:
And then there's the other description.

[00:21:04] Speaker 01:
I think you asked counsel whether the preamble was limited to a type of data.

[00:21:09] Speaker 01:
No argument about that below.

[00:21:10] Speaker 01:
That would be a claim construction issue.

[00:21:12] Speaker 01:
But regardless, the claims talk about also data that comes from the bus master and data that comes from everything.

[00:21:18] Speaker 01:
So we have two specific embodiments dealing with two types of data.

[00:21:22] Speaker 01:
And all it's saying, Your Honor, is we're going to put one of those two categories in a place that's more convenient and will operate more efficiently.

[00:21:30] Speaker 01:
So while the court, for example, may have

[00:21:32] Speaker 01:
a book of the court's rules, and it may have that on a bookshelf that's right behind its desk.

[00:21:37] Speaker 01:
It may have a variety of other cases a lot further away.

[00:21:39] Speaker 03:
But that seems to be what ultimately will occur later.

[00:21:42] Speaker 03:
That's like saying, I claim a more sophisticated school bus, but then I'm going to say that the abstract idea is to drive children from one location to another.

[00:21:54] Speaker 01:
No, Your Honor.

[00:21:54] Speaker 01:
This is saying on the front end that we're going to put data in place A or place B.

[00:22:00] Speaker 01:
What are we going to do that based on?

[00:22:02] Speaker 03:
It doesn't say we're going to put data in place.

[00:22:04] Speaker 03:
It says we're going to create a system that will allow data transfers through a precise type of mechanism and do it in a way that the processor was never capable of doing before.

[00:22:19] Speaker 01:
Well, Your Honor, I disagree.

[00:22:21] Speaker 01:
It's not telling us what the precise type of mechanism is.

[00:22:24] Speaker 01:
That's the distinction here with ENFISH.

[00:22:27] Speaker 01:
So what it's telling us is we're going to create a system

[00:22:30] Speaker 01:
that puts data in place A or place B. And that will be based on this undefined programmable operational characteristic.

[00:22:36] Speaker 01:
And I would like to briefly take issue with the discussion regarding claim eight and six.

[00:22:42] Speaker 01:
We're happy to discuss that.

[00:22:43] Speaker 01:
But the court was pretty clear in electric power whether a meaningful argument was presented for distinctive significance.

[00:22:50] Speaker 01:
And contract extraction talked about whether the patentee focused on defending the representative claim.

[00:22:57] Speaker 01:
Here we have 58 pages in two briefs.

[00:23:00] Speaker 01:
with one page in the main brief and two paragraphs in the reply brief focused on the dependent claims, there's no argument that any additional conventional elements impart inventiveness.

[00:23:10] Speaker 01:
And there's no arguments that the claims are different in any manner material to the 101 inquiry.

[00:23:14] Speaker 01:
There's just three sentences, for example, regarding claim eight that are repeated that merely say, well, claim eight has additional caches.

[00:23:21] Speaker 01:
Nothing new or different about that.

[00:23:23] Speaker 01:
And then it says claim eight has three programmable operational characteristics.

[00:23:28] Speaker 01:
Those are just three undefined functional boxes that tell us we're going to store memory in one of two places based on the specification says whether it's more efficient.

[00:23:39] Speaker 01:
The claims don't even tell us that.

[00:23:40] Speaker 01:
And I don't believe that that's clear to want to steal in the art.

[00:23:43] Speaker 01:
And I don't think that that provides any inventiveness.

[00:23:45] Speaker 04:
How was it argued below?

[00:23:47] Speaker 04:
Was there a representative claim or were the claims argued separately?

[00:23:50] Speaker 01:
Below, it was argued principally on a representative claim.

[00:23:53] Speaker 01:
That was the majority argument.

[00:23:54] Speaker 01:
But we also separately preserved and argued each of the claims separately.

[00:23:59] Speaker 01:
And if you look at the appendix at 16 to 17, you'll see the district court's description and going through those claims independently.

[00:24:06] Speaker 01:
What I would submit, Your Honor, is that hasn't been preserved on appeal, and it certainly hasn't been argued separately.

[00:24:12] Speaker 01:
Regardless, I'm happy to talk about those independent claims in the three sentences that are provided as the basis.

[00:24:19] Speaker 01:
As I said,

[00:24:20] Speaker 01:
They're more specific in that they claim different memory systems.

[00:24:23] Speaker 01:
That's just those conventional memories, main memory, plurality caches.

[00:24:28] Speaker 01:
There's a discussion of registers, which I would submit haven't become any more non-conventional since Benson evaluated registers in 1972.

[00:24:37] Speaker 01:
And then there's this fast page and first page technique, the council reference, which is nothing more than a pointer.

[00:24:43] Speaker 01:
Again, they're known techniques for pulling up the most recently accessed data.

[00:24:49] Speaker 01:
The whole

[00:24:50] Speaker 01:
Key here, according to this patent, is how do you do that?

[00:24:55] Speaker 01:
And the way that the patent says you do that is based on the programmable operational characteristic.

[00:25:00] Speaker 01:
And as the Court recognized, there's no description of how.

[00:25:03] Speaker 01:
There's no description, for example, of the rules that were set forth in the claims in McRoe, which went through these detailed descriptions of the mathematical basis on which you're going to provide facial animations as opposed to the subjective basis that were used by individuals beforehand.

[00:25:20] Speaker 01:
And there's certainly no description of technology that's comparable to BASCOM.

[00:25:26] Speaker 01:
Here, the council doesn't even claim that we have non-conventional and non-generic arrangement of known conventional parts.

[00:25:33] Speaker 01:
All we have is a sequence of known conventional memory pieces.

[00:25:38] Speaker 01:
And then we're saying the system is going to allocate what type of data goes into which memory area based on a programmable operational characteristic.

[00:25:48] Speaker 01:
That characteristic

[00:25:50] Speaker 01:
wasn't construed, the patentee didn't propose or present any potential.

[00:25:56] Speaker 04:
What was this decided?

[00:25:58] Speaker 04:
It was decided early in the case, right?

[00:26:00] Speaker 04:
I mean, did the court engage in claim construction?

[00:26:04] Speaker 01:
What the court did, Your Honor, at the very start of the claim construction argument, Judge Andrews, the district court, asked three times, not once, not twice, three times whether there was any claim construction proposed for that term.

[00:26:19] Speaker 04:
You said at the beginning of claim construction proceedings.

[00:26:22] Speaker 01:
No, I apologize.

[00:26:22] Speaker 01:
The 101 argument.

[00:26:23] Speaker 01:
I apologize, Your Honor.

[00:26:24] Speaker 01:
I misspoke.

[00:26:25] Speaker 01:
Asked three times, and then again, at the end of the argument, he literally said, look, in the hour that you've had to think about it, do you have anything that you want to propose for construing this claim?

[00:26:35] Speaker 01:
And counsel said, no.

[00:26:37] Speaker 01:
We just are saying that we're not sure what it means, but it may mean something, and the CDA may have some reference to it.

[00:26:43] Speaker 01:
I would submit that to the extent claim construction is relevant,

[00:26:46] Speaker 01:
to a one-on-one analysis, it's the burden of the patentee to at least identify what the claim term is and to propose a construction that may make a difference in the eligibility analysis.

[00:26:57] Speaker 01:
But here, that wasn't done, and I don't think it would make a difference.

[00:27:03] Speaker 01:
Respectfully, the district court correctly applied the two-step analysis here.

[00:27:08] Speaker 01:
They determined the focus of the invention was retrieving and receiving and storing data by the category of data.

[00:27:16] Speaker 01:
He then found that the additional limitations in the claims didn't add anything innovative.

[00:27:21] Speaker 01:
They were generic computer components, cache, main memory, bus, processor, and some registers described in less detail than those found ineligible in Benson.

[00:27:30] Speaker 04:
They then described this programmable op... Is it your position that any time you have conventional hardware in a computer system, if the improvement is in the nature of software, it's not eligible?

[00:27:40] Speaker 01:
No, it's not, Your Honor.

[00:27:42] Speaker 01:
That, I think, is much closer to the Enfish decision.

[00:27:47] Speaker 01:
My statement is that if your alleged improvement in the conventional use of conventional components is something software related, then the claim has to somehow be limited to what that software is.

[00:28:03] Speaker 01:
And here, all we have is the programmable operational characteristic.

[00:28:06] Speaker 04:
There's no... Why isn't that limited?

[00:28:07] Speaker 04:
It says that it's... I'm just going to push back a little bit and see how much you have to say.

[00:28:12] Speaker 04:
So the preamble says that there's a programmable operational characteristic that's going to depend on the type of processor and that it's going to impact the memory and the type of data that the memory stores.

[00:28:24] Speaker 04:
Why isn't that enough?

[00:28:25] Speaker 01:
Because what we would need after that are the sort of decisional rules, for example, that were set out in the claim in McRoe.

[00:28:33] Speaker 04:
So any time it's functional, you have to have very specific, narrow, even if it improves the computer,

[00:28:40] Speaker 04:
you have to have very specific narrow instructions on how it's performed.

[00:28:44] Speaker 01:
Well, I guess what I would say, Your Honor, is you need something more than just bare functional description.

[00:28:51] Speaker 01:
So you can't, as the Court has already decided, take a software brain and just say the software brain is going to make this analysis and going to come up with the result that's most efficient or best.

[00:29:02] Speaker 01:
You need something more than what we have here and what we have is nothing.

[00:29:06] Speaker 01:
I would respectfully submit that the District Court correctly

[00:29:09] Speaker 01:
applied two-step analysis and the court should have found.

[00:29:12] Speaker 03:
Thank you.

[00:29:14] Speaker 03:
All right, you have two minutes and 20 seconds left.

[00:29:18] Speaker 04:
Do you think one of ordinary skill in the art would know, with the knowledge of one of ordinary skill in the art, would understand how to implement the invention and claim one?

[00:29:27] Speaker 00:
Yes, Your Honor.

[00:29:27] Speaker 00:
This patent was invented by people that worked at NCR Corporation, which is a big computer software company.

[00:29:34] Speaker 00:
And these inventors were one of ordinary skill in the art.

[00:29:37] Speaker 00:
They wrote the code that's in the

[00:29:39] Speaker 00:
They were at the CDL that was submitted to the patent office.

[00:29:42] Speaker 00:
The answer is yes.

[00:29:43] Speaker 00:
And if that's a question as to whether an ordinary skill in the art would or would not, that's a fact question that's not right at the 12b6 stage when deciding motion to dismiss for immolidity under 101.

[00:29:54] Speaker 03:
What's your response to the argument that if you really want to fall into ENFISH, you have to put all the functional elements into the claim itself?

[00:30:04] Speaker 00:
ENFISH and AMDOTS both say you look at the written description when redoing the claims.

[00:30:08] Speaker 00:
And that's what you can do here to find that the claims are sufficiently described in specifications that one ordinary skill in the art could practice it.

[00:30:16] Speaker 02:
In other words, when you look at the claims... And is that that CDL language in column six you were referring to?

[00:30:23] Speaker 00:
It's the CDL language in column six.

[00:30:26] Speaker 00:
It's also the description in, I believe it's column four.

[00:30:32] Speaker 02:
Can you walk the column four through me again?

[00:30:35] Speaker 02:
Because I didn't understand that.

[00:30:37] Speaker 00:
So column four, line 21, beginning with an important feature of the present invention is the separation of the functionality of each of the caches and the selected definition of those functions based on the processor type.

[00:30:48] Speaker 00:
That explains how separating the functionality cache, in other words, where code and non-code gets stored, is a benefit.

[00:30:56] Speaker 02:
And then it- Why isn't that just saying, based on the processor type, we're going to do it differently?

[00:31:03] Speaker 00:
I don't fully understand your question.

[00:31:05] Speaker 02:
Well, throughout this, it seems to me that what you're saying is we have a flexible system, and based upon the processor used, we're going to program it differently to achieve the best results.

[00:31:16] Speaker 02:
It doesn't tell me what kind of things you're going to think about depending on the different type of system.

[00:31:22] Speaker 02:
And this sentence you just quoted me, this sounds exactly like the same thing.

[00:31:25] Speaker 00:
Later in the paragraph, it does.

[00:31:27] Speaker 00:
So for example, in column four, line 32,

[00:31:30] Speaker 00:
For a system employing a 386 or 386SX system processor, internal cache 16 holds only code data, whereas for a system employing a 46 processor, internal cache 16 holds both code and non-code data.

[00:31:42] Speaker 00:
So there's different statements like that throughout the specification based on the type of processor, whether it be a 386, 386SX, or 486.

[00:31:51] Speaker 00:
Because those are three different types of processors that require different items for the memory system in order to function efficiently.

[00:32:01] Speaker 00:
Now, my colleague said that visual memory did not make the statement that the components operate in an unconventional manner.

[00:32:11] Speaker 00:
That's false.

[00:32:13] Speaker 00:
The reply brief at page 14 specifically says, quote, when these limitations in the respective claims require that any, quote, generic components operate in an unconventional manner to achieve an improvement in computer functionality, end quote, citing Amdocs and Bascom.

[00:32:28] Speaker 00:
And that's referring to claims 1, 6, and 8.

[00:32:30] Speaker 03:
What's your response to the claim construction question?

[00:32:32] Speaker 03:
Why didn't you ask the court to construe claim one to necessarily include the information in column six?

[00:32:40] Speaker 00:
So we actually raised the same argument that we raised with your honors today.

[00:32:44] Speaker 00:
That's in the appendix at APPX 0451 is where it begins, where we pointed the district court to the elements of the specification.

[00:32:54] Speaker 00:
And then NVIDIA raised the CDL code.

[00:32:58] Speaker 00:
And we informed the court that that

[00:33:00] Speaker 00:
CDL code or the CDL listing was the best mode and was for enablement, but also would permit one of ordinary skill in the art to practice the invention.

[00:33:09] Speaker 04:
But you didn't say, you said it was more for enablement, not that the claims were limited to it.

[00:33:13] Speaker 04:
Or what did you say?

[00:33:14] Speaker 00:
We said that it was best mode.

[00:33:15] Speaker 00:
We said it was best mode.

[00:33:17] Speaker 03:
You didn't say it was preferred embodiment?

[00:33:23] Speaker 00:
Best mode can be considered a preferred embodiment.

[00:33:25] Speaker 03:
Essentially the same thing.

[00:33:26] Speaker 00:
Right.

[00:33:26] Speaker 00:
Just like if you look at what we said and what we said today,

[00:33:29] Speaker 00:
That would essentially be plain meaning to one of ordinary skill in the art.

[00:33:32] Speaker 03:
Okay.

[00:33:33] Speaker 00:
Not plain meaning based on what a non-ordinary person, non-ordinary person with a skill in the art thinks it means.

[00:33:40] Speaker 03:
Okay, you're out of time.

[00:33:41] Speaker 04:
What about the contention that you haven't raised?

[00:33:44] Speaker 04:
I'm sorry.

[00:33:45] Speaker 04:
I just said he was out of time, but you can go.

[00:33:46] Speaker 04:
Oh, sorry.

[00:33:48] Speaker 04:
But you're not out of time.

[00:33:50] Speaker 04:
What about the argument that you didn't argue the claims separately and that you were just treating them all as rising and falling?

[00:33:56] Speaker 00:
That's not true.

[00:33:57] Speaker 00:
Throughout the district court briefing, we did argue, just like we did here, that there are differences in claims.

[00:34:02] Speaker 03:
But is it true you only dedicated one or two sentences to it here?

[00:34:07] Speaker 00:
No, that's not true.

[00:34:08] Speaker 00:
We have sentences throughout the briefs.

[00:34:10] Speaker 00:
We also have different sections for each.

[00:34:15] Speaker 00:
In both the opening and reply, there's a section related to no representative claim, but also sprinkled throughout the arguments dealing with step one and step two are arguments that claims one, six, and eight are different.

[00:34:26] Speaker 00:
We run our time out of time, and thank you for the little brief attention, and we ask that you reverse the district court.

[00:34:33] Speaker 02:
Thank you.