[00:00:00] Speaker 05:
I mean, you might as well say where you are.

[00:00:03] Speaker 05:
We'll figure it out.

[00:00:07] Speaker 05:
162274, Netlist vs. Sandus.

[00:00:33] Speaker 06:
Sorry about that.

[00:00:33] Speaker 06:
Is that okay with you all?

[00:00:35] Speaker 06:
You'll figure it out.

[00:00:35] Speaker 06:
We'll keep it straight.

[00:00:36] Speaker 06:
Particularly with the cross appeals.

[00:00:41] Speaker 06:
We'll try to do our best.

[00:00:47] Speaker 01:
Are you sure you're supposed to be there?

[00:00:51] Speaker 00:
Yes.

[00:00:52] Speaker 00:
I have to stop and think, but yes, I can stop.

[00:01:16] Speaker 00:
Thank you.

[00:01:17] Speaker 00:
Good morning.

[00:01:17] Speaker 00:
May it please the court?

[00:01:19] Speaker 00:
This case demonstrates the need to protect the due process and procedural rights of patent owners where an IPR petitioner makes new arguments and factual representations regarding the prior for the first time at oral argument.

[00:01:32] Speaker 00:
And it also demonstrates the danger of allowing the board to fill in the gaps in the references.

[00:01:38] Speaker 00:
And time permitting and subject to your questions, I'd like to address four primary points

[00:01:43] Speaker 00:
two with respect to each reference and first with respect to the average reference.

[00:01:48] Speaker 00:
Netlist was denied its procedural rights because Sandus failed to tell me to disclose its theory that there's a corner case where the sequencers produce address signals that can pass through the address generation unit without transformation or modification.

[00:02:04] Speaker 01:
Let me ask you.

[00:02:05] Speaker 00:
Yes.

[00:02:06] Speaker 01:
Specifics.

[00:02:08] Speaker 01:
On page 53 of the blue brief,

[00:02:12] Speaker 01:
Netlist argues, SanDisk has failed to show an inherent anticipation by Avrabuj.

[00:02:19] Speaker 01:
And I'm putting ellipses in here.

[00:02:21] Speaker 01:
There has been no showing how Avrabuj's memory interface with its associated address generation unit can necessarily pass through untransformed, unmodified address signals, and yet still provide the peculiarities required by the alleged devices.

[00:02:39] Speaker 01:
And then you reference your prior

[00:02:42] Speaker 01:
argument above that.

[00:02:43] Speaker 01:
And I wrote a marginal note there saying, this is an interesting and good-looking argument, except I don't see any evidence above that actually supports the position.

[00:02:54] Speaker 01:
What's your evidence for that?

[00:02:56] Speaker 00:
Well, the position is essentially, Your Honor, that there's no disclosure of a pass-through.

[00:03:03] Speaker 00:
And to say that it would have to necessarily be there, they're pointing to the words as needed.

[00:03:10] Speaker 00:
to essentially mean that it must necessarily pass through.

[00:03:14] Speaker 00:
And the language doesn't support it.

[00:03:15] Speaker 00:
You'd have to interpret it.

[00:03:16] Speaker 00:
It's not must necessarily.

[00:03:17] Speaker 00:
It's can necessarily pass through.

[00:03:19] Speaker 00:
Can, yes.

[00:03:20] Speaker 00:
Very good, Your Honor.

[00:03:21] Speaker 00:
You're correct.

[00:03:22] Speaker 00:
Can pass through in some corner case.

[00:03:25] Speaker 00:
And the evidence doesn't show that at all.

[00:03:29] Speaker 00:
And just to highlight that, Avrabuj, expressly a paragraph six says that the memory may have different bit densities, access speeds,

[00:03:39] Speaker 00:
Addressing requirements access protocols and other particularities that's appendix 1036 Now how did those particularities of the memory interface get satisfied?

[00:03:51] Speaker 00:
Well, let's look at the term the words as needed in Context because I think we probably all agree that that's how it should be reviewed.

[00:03:59] Speaker 00:
We have a three-tier architecture in avro bush abyss and

[00:04:04] Speaker 00:
which at paragraph eight of 1036 says that the BIS provides only generalized commands.

[00:04:09] Speaker 00:
It's a generic command that communicates with tier two, the sequencer.

[00:04:14] Speaker 00:
And the tier two sequencer is said to have very limited control, and it has specific reference to putting a common clock domain between the sequencer and one or more memory modules.

[00:04:27] Speaker 01:
Let me ask you this.

[00:04:28] Speaker 01:
Looking at figure six, is

[00:04:34] Speaker 01:
42, the address generation unit used to transform or modify address signals from the sequencer at 8A?

[00:04:44] Speaker 00:
I believe so.

[00:04:47] Speaker 00:
So we have the second tier, the sequencer, which is said to have only limited control of the application, and it talks about a common clock domain.

[00:04:58] Speaker 00:
And then we have an express teaching for tier three, the memory interfaces.

[00:05:02] Speaker 00:
You can find that at paragraph 11 at a

[00:05:04] Speaker 00:
NX1036, which says they're designed in accordance with the particular signal interface requirements and physical characteristics of the respective one of the memory modules.

[00:05:14] Speaker 00:
So we don't have to speculate that there might be a pass-through.

[00:05:19] Speaker 00:
The teaching is you have a generic signal.

[00:05:21] Speaker 00:
You have very limited additional control for the sequencer.

[00:05:25] Speaker 00:
Primarily, it says for common clock domain.

[00:05:28] Speaker 00:
And then you have all of this expressed teaching from paragraph 11 that says,

[00:05:33] Speaker 00:
The memory interface is where all of the particular requirements for the memory are handled and generated.

[00:05:39] Speaker 00:
It's the bottom tier.

[00:05:40] Speaker 00:
There's no teaching that the sequencer can have a signal that can pass through in a form that's based on the particular physical characteristics of the memory module.

[00:05:51] Speaker 00:
It's not there.

[00:05:52] Speaker 00:
And it's not, I just took you through the context of Avriboosh to show why it's not there.

[00:05:57] Speaker 00:
But it's also, Avriboosh doesn't use the words only if needed.

[00:06:03] Speaker 00:
He says, as needed.

[00:06:04] Speaker 00:
And the context of Average is that there's a transformation.

[00:06:08] Speaker 00:
It's made as needed, which is in a way necessary to achieve the goal.

[00:06:12] Speaker 02:
Is that transformation at 42?

[00:06:14] Speaker 02:
Forgive me.

[00:06:17] Speaker 02:
That's the generator.

[00:06:18] Speaker 00:
Yeah, the 42, if that's the AGU, the Address Generation Unit, yes, that's what the specification teaches.

[00:06:25] Speaker 00:
That's where the transformation is.

[00:06:27] Speaker 00:
That's where all the particular requirements are handled.

[00:06:30] Speaker 01:
Where does it say that?

[00:06:32] Speaker 01:
In what paragraph?

[00:06:38] Speaker 00:
I took you through the context of as needed, Your Honor, but 42 says that the address generation unit, I'm sorry, the specification teaches that the address generation unit is what generates the address and control signals to be used by the memory devices.

[00:06:58] Speaker 00:
The specification teaches that there are two different generation units, one for addresses and one for data.

[00:07:08] Speaker 00:
And that the specification never equates them.

[00:07:11] Speaker 00:
And I'll tell you that I wanted to point out, and we did in our brief, that the board relied upon paragraph 56 of Average, which concerns the data generation unit and not the address generation unit.

[00:07:23] Speaker 00:
But Average teaches that it has two distinct units, and they're separate and distinct.

[00:07:29] Speaker 00:
And there's no block diagram for the address generation unit, but there is a block diagram for the data generation unit.

[00:07:36] Speaker 00:
And you'll see that the block diagram has these XOR gates, which provide the logic for the data generation unit, which is very detailed and specific.

[00:07:47] Speaker 00:
But there is no, and that's figure seven, by the way.

[00:07:53] Speaker 00:
But there is no such block diagram for the address generation unit.

[00:07:59] Speaker 00:
And it has distinct purposes.

[00:08:01] Speaker 00:
It has different inputs.

[00:08:02] Speaker 00:
It has different outputs.

[00:08:03] Speaker 00:
There's just no basis whatsoever to assume that the address generation unit has the same structure, function, or operation as the data generation unit.

[00:08:12] Speaker 00:
They're two distinct blocks.

[00:08:14] Speaker 00:
And the board, frankly, completely overlooked that because they cited the data generation unit

[00:08:20] Speaker 00:
as evidence of how the address generation unit operates.

[00:08:24] Speaker 00:
But it's just not.

[00:08:24] Speaker 01:
Let me take you over to Wayne.

[00:08:26] Speaker 01:
Yes.

[00:08:28] Speaker 01:
With respect to the Group B arguments, Netlist accused the PTAB of improperly burden shifting.

[00:08:36] Speaker 01:
But the PTAB addressed both parties' arguments and evidence and concluded that SanDisk proved unpatentability by preponderance of the evidence.

[00:08:45] Speaker 01:
Are you still standing by that burden shifting?

[00:08:47] Speaker 00:
Absolutely, Your Honor.

[00:08:48] Speaker 00:
And I'll tell you, and I thought we made it clear in our brief, and forgive me if we didn't, but the board only looked at the data writing functionality of the Hwang memory interface.

[00:09:01] Speaker 00:
And it ignored, it just ignored the testimony and evidence regarding the reading and shifting functionality of Hwang's transparent interface.

[00:09:12] Speaker 00:
And it's clear, if Hwang is

[00:09:15] Speaker 00:
is black and white on this.

[00:09:16] Speaker 00:
At appendix 1078, the test interface can read each word and then shift the entire word bit by bit for testing.

[00:09:25] Speaker 00:
So it has a read operation and a shift operation.

[00:09:29] Speaker 00:
And we tried to take you through figure four of Wang, which is the redundant shift operation.

[00:09:35] Speaker 00:
And it might be a little bit complicated, but I'll try to revisit that.

[00:09:41] Speaker 00:
Figure 4 shows you have a memory A and a memory B, and then you have a 4-bit word width and memory A and a 6-bit word width and memory B, and those are read into the memory interface.

[00:09:55] Speaker 00:
And you see the word latch, and while memory B has 6 bits, they all are read into the memory B interface, but the memory A interface, because it operates on a common control line, has to shift two extra times

[00:10:11] Speaker 00:
with redundant bits.

[00:10:13] Speaker 00:
And so the memory B interface has a read and shift operation that is influenced by the word width and memory B by the operation of the memory interface for memory B. And so the operation of one transparent interface is influenced, i.e.

[00:10:34] Speaker 00:
forced to continue shifting by the operation of the other

[00:10:39] Speaker 00:
And you won't find this in the board's decision.

[00:10:41] Speaker 00:
We put it in our brief.

[00:10:42] Speaker 00:
We cited it out.

[00:10:43] Speaker 00:
We argued it.

[00:10:44] Speaker 00:
We put the declaration evidence from our Dr. Seachen, who walked through it.

[00:10:51] Speaker 00:
And the board, for its part, stopped with the read functionality and didn't look to the, I'm sorry, the board only considered the writing functionality and didn't consider the read and shift functionality of the memory interface.

[00:11:08] Speaker 00:
And they stopped there and then they expressly said that Netlist, you haven't shown how the memory interfaces are operable independently.

[00:11:18] Speaker 00:
And that was an improper burden shift.

[00:11:20] Speaker 04:
But that was right after they said Sandus met its burden.

[00:11:25] Speaker 04:
I mean, this sounds like opinion writing issues.

[00:11:28] Speaker 04:
The board is allowed to reject your arguments by saying they're insufficient without being accused of burden shifting.

[00:11:34] Speaker 00:
Well, if they did that, your honor,

[00:11:37] Speaker 00:
Then they ignored the evidence that directly refutes the position that conclusion they reached.

[00:11:43] Speaker 00:
Oh, they disagreed with you.

[00:11:45] Speaker 00:
Pardon me?

[00:11:46] Speaker 02:
Well, they disagreed with you.

[00:11:47] Speaker 00:
They disagreed.

[00:11:49] Speaker 00:
Because they didn't acknowledge it.

[00:11:51] Speaker 04:
I can only have to acknowledge it though.

[00:11:53] Speaker 04:
I mean, if you think that they made an error and they, their decision lacks substantial evidence because they didn't address this argument specifically, then you argue that.

[00:12:02] Speaker 04:
I mean, to argue that they're burn shifting because in one sentence they say they don't agree with your argument is a bit too much.

[00:12:09] Speaker 00:
Fair enough.

[00:12:10] Speaker 00:
Your honor.

[00:12:11] Speaker 00:
Then, then, uh, I think point taken, but it doesn't change the result.

[00:12:18] Speaker 00:
The result is that the board didn't consider the evidence and argument that showed that one memory interface is influenced by the operation of another.

[00:12:26] Speaker 00:
And so whether you call it a burden shift or you take that language out... Well, you can't say they didn't consider it.

[00:12:32] Speaker 04:
They may have just disagreed with you.

[00:12:34] Speaker 04:
You can say that they were wrong and that substantial evidence doesn't support their decision, but it's not that they didn't consider it.

[00:12:42] Speaker 00:
Okay.

[00:12:44] Speaker 00:
Maybe it's best said that the board's decision is not supported by substantial evidence.

[00:12:51] Speaker 04:
This is a big deal.

[00:12:54] Speaker 04:
This is not the first time I've heard arguments that the board is shifting burdens because they use a sentence saying that the patentee has not

[00:13:04] Speaker 04:
demonstrated the validity of its argument, but they've otherwise said the petitioner has met its burden.

[00:13:09] Speaker 04:
And it's attempting to transform a substantial evidence question into a procedural legal error.

[00:13:15] Speaker 04:
And that's improper.

[00:13:16] Speaker 04:
You can argue substantial evidence, but taking one sentence out of context when they rejected your argument is not a legal error.

[00:13:24] Speaker 00:
Our view was that we couldn't tell that they considered the issue and they didn't rule on it expressly.

[00:13:31] Speaker 00:
I'm into my rebuttal time, if I may reserve.

[00:13:33] Speaker 00:
Yes.

[00:13:34] Speaker 00:
Thank you.

[00:13:56] Speaker 03:
May it please the court, net lists,

[00:14:01] Speaker 03:
Manufactured procedural argument for the Avrabouj grounds is both illusory and irrelevant.

[00:14:09] Speaker 03:
Under the uncontested claim construction, there can be little debate that Avrabouj's sequencer generates an address and control signal for testing the memory.

[00:14:22] Speaker 03:
That's the best address and control signal.

[00:14:24] Speaker 03:
Clarify something for me.

[00:14:27] Speaker 01:
Netlist argues.

[00:14:31] Speaker 01:
at least at page 25 of its brief, but several times, that it was only an oral argument that Sandisk alleged that Avrabuj teaches that the Adventists could pass through.

[00:14:45] Speaker 01:
And in the RedBerry, Sandisk responds that we did it at least five separate times in sites to various points in the PTAB proceeding the petition.

[00:15:02] Speaker 01:
expert declarations and reply and so on, saying they're clear notice of the past scenario.

[00:15:09] Speaker 01:
How do the parties have such totally contradicting views?

[00:15:14] Speaker 01:
I mean, it's almost a black and white thing.

[00:15:18] Speaker 03:
I can't speak to why Netlist has such a contradictory view, but the issue as it crystallized post-institution

[00:15:32] Speaker 03:
really was whether the operation of Avrabuj's address generation unit somehow removed what the sequencer unambiguously does from the scope of the control module, which is simply produce that.

[00:15:50] Speaker 03:
So as that issue developed over the course of post-institution,

[00:16:00] Speaker 03:
I guess perhaps it's best to just go straight to the record.

[00:16:04] Speaker 03:
If you look at the reply at page, this is at Appendix 666.

[00:16:19] Speaker 03:
So actually, let me back up, because both the petition and the board's institution decision

[00:16:28] Speaker 03:
said that the sequencer acknowledged that the operation of the address generation unit operates as needed when it translates what the sequencer has generated, which is a BISC address and control signal, a control signal for testing a memory device.

[00:16:50] Speaker 03:
and laid the evidentiary foundation for this dispute over how the address generation unit, what it actually does.

[00:16:59] Speaker 03:
And in the reply, and this is at Appendix 666, we state up at the top that there's no dispute that the sequencer produces the best address control signal.

[00:17:14] Speaker 03:
This is the second and third line.

[00:17:16] Speaker 03:
And then down below in the next paragraph, we say that the actual addressing information produced by the sequencer and the address generation unit 42 merely applies the received addressing information to a given physical memory configuration, row, column configuration.

[00:17:37] Speaker 03:
We cite to Alpert's declaration, where he again cites to the as needed

[00:17:44] Speaker 03:
language for the address generation unit.

[00:17:46] Speaker 03:
That's at paragraph 49, and that's at appendix 1403.

[00:17:49] Speaker 01:
Mr. Wimbiskas?

[00:17:51] Speaker 01:
I'm sorry?

[00:17:52] Speaker 01:
Mr. Wimbiskas?

[00:17:54] Speaker 01:
No, right.

[00:17:55] Speaker 01:
Oh, I'm sorry.

[00:17:56] Speaker 01:
Me?

[00:17:57] Speaker 01:
Sorry, yeah.

[00:17:57] Speaker 01:
I'm getting you.

[00:17:59] Speaker 01:
So right, OK.

[00:18:00] Speaker 01:
You switched sides.

[00:18:01] Speaker 01:
Yeah, I'm sorry.

[00:18:02] Speaker 01:
Yeah, I knew it was going to get me.

[00:18:05] Speaker 01:
So you just demonstrated the importance of the comma in the English language when you

[00:18:11] Speaker 01:
read that and you missed the comma.

[00:18:13] Speaker 01:
And of course, the comma changed the entire meaning of that paragraph.

[00:18:17] Speaker 01:
I'm just pointing that out for our grammarians in the audience.

[00:18:20] Speaker 03:
OK.

[00:18:22] Speaker 03:
Thank you, Your Honor.

[00:18:22] Speaker 03:
And for my clerks.

[00:18:30] Speaker 03:
So lest there be any doubt that Netlist understood the argument it was making, if you go to the deposition where they deposed

[00:18:39] Speaker 03:
Dr. Albert on this very issue.

[00:18:41] Speaker 03:
And this is Appendix 2241 through 2243.

[00:18:49] Speaker 03:
This is where the actual articulation of as needed shows up for the first time in the record was when they were deposing Dr. Albert.

[00:19:01] Speaker 03:
And in response to a series of six or seven questions, maybe more,

[00:19:07] Speaker 03:
about what is happening at the address generation unit.

[00:19:12] Speaker 03:
Dr. Alpert responds after a number of questions.

[00:19:15] Speaker 03:
He says, well, I was going to say it might not.

[00:19:18] Speaker 03:
It's according to the, this is at appendix 2242.

[00:19:22] Speaker 03:
It's according to the physical characteristics of the memory module.

[00:19:26] Speaker 03:
It's possible that the address can really just, address and control can just pass straight through depending on what might be required to map the patterns to the row and column organization.

[00:19:34] Speaker 03:
So it's only in response to their questioning on this particular issue that the pass-through comes out.

[00:19:43] Speaker 03:
It wasn't a new theory that was sprung upon Netlist, and certainly not at the oral hearing.

[00:19:49] Speaker 03:
When it came to the oral hearing, as the petitioner, Sandus, went first.

[00:19:55] Speaker 03:
Netlist goes second.

[00:19:56] Speaker 03:
Netlist stood up, and they made a lot of arguments in the oral hearing about they have new counsel, and they made a new argument on this, and a new argument on that, and a new argument here.

[00:20:05] Speaker 03:
They didn't call out this pass-through theory specifically as a new argument when they had the opportunity to do it at the oral hearing, and the board didn't view it as a new argument either.

[00:20:16] Speaker 03:
They didn't like the answer, I think, that they got when they deposed Dr. Alpert, but this was not a new argument that Sandus sprung on them right at the end.

[00:20:31] Speaker 03:
The issue, as it crystallized for the Avrabouj grounds anyway, was the impact of the data generation unit.

[00:20:37] Speaker 03:
And I want to address the inherent anticipation argument, Judge Wallach, that you raised earlier.

[00:20:44] Speaker 03:
There is substantial evidence in the record supporting the board's factual determination on how the address generation unit operates.

[00:20:53] Speaker 03:
If we could turn to Avrabouj, starting at paragraph 11,

[00:20:59] Speaker 01:
Do you have a page for us?

[00:21:00] Speaker 03:
I do.

[00:21:00] Speaker 03:
That's Appendix 1036.

[00:21:05] Speaker 03:
Mr. Wambiscus cited the paragraph 11, but in the middle of paragraph 11 at 1036, it says, each memory interface receives memory operations from a controlling sequencer and translates as needed, including associate address.

[00:21:22] Speaker 03:
In the next sentence, for example, a memory interface may... Wait a minute.

[00:21:25] Speaker 04:
You just read

[00:21:27] Speaker 04:
that sentence, each memory interference, where does it say as needed in that sentence?

[00:21:33] Speaker 03:
It does.

[00:21:34] Speaker 03:
It just translates memory operations, including associated address.

[00:21:38] Speaker 04:
Oh, you just left out the middle part.

[00:21:39] Speaker 03:
Yeah, as needed.

[00:21:40] Speaker 04:
Be careful.

[00:21:41] Speaker 04:
I'm sorry.

[00:21:41] Speaker 04:
It sounded like you.

[00:21:45] Speaker 04:
That's OK.

[00:21:45] Speaker 04:
Keep going.

[00:21:46] Speaker 01:
No, I'm sorry.

[00:21:46] Speaker 01:
That's why when I quote from briefs, I say there's an ellipsis in there.

[00:21:51] Speaker 03:
Right.

[00:21:52] Speaker 03:
So what I was driving at was the next sentence where it says, for example, a memory interface may

[00:21:57] Speaker 03:
translate addresses supplied by a controlling sequencer, and the memory interface may translate the data.

[00:22:05] Speaker 03:
Perhaps the most compelling substantial evidence, and the board cited to this, not just cited to it, but reproduced this section in both the institution decision and the final written decision, is if you look at paragraph 48 of ABRAVOUGE, about halfway down that paragraph it says,

[00:22:25] Speaker 03:
When electronic device two is operating in the BIST mode, however, the BIST enabled signal causes the multiplexers 45 and 46 to select the BIST address control signals and the test data provided by the respective higher level sequencer.

[00:22:44] Speaker 03:
If we look to figure six and we see what's happening there, which is an appendix 1030, it says that multiplexer 46 selects

[00:22:55] Speaker 03:
the BIST address control signal.

[00:22:58] Speaker 03:
And it's the multiplexer 46 that sits right in front of the line going out to the memory module.

[00:23:09] Speaker 03:
So the multiplexer selects the BIST address control signal.

[00:23:13] Speaker 03:
And then finally, with respect to the complaint about the data generation, the board relying on the data generation unit,

[00:23:23] Speaker 03:
the language as needed appears for both the address generation unit and the data generation unit.

[00:23:29] Speaker 03:
The board looked, and because the data generation unit provides an actual example of what as needed means, the board turned to that.

[00:23:37] Speaker 03:
And that's spelled out in paragraph 56 in Avrabuj at 1040, where it says there, and I think there's no dispute, that the data generation unit can pass the data through with no

[00:23:50] Speaker 03:
transformation or translation.

[00:23:53] Speaker 03:
So that's an example in the specification that the board relied on to show what as needed means.

[00:24:00] Speaker 05:
Can you, before your time runs out, can you turn to the Wang?

[00:24:04] Speaker 05:
Yes.

[00:24:04] Speaker 05:
Because it's a little confusing to me.

[00:24:08] Speaker 05:
I guess we've been looking at figure four for Wang and your friend talks about the memory interface

[00:24:17] Speaker 05:
And so can you, just using these diagrams, respond to what he said?

[00:24:21] Speaker 05:
I assume you disagree with his analysis.

[00:24:23] Speaker 03:
Yeah, I think to set the stage for that analysis and how the board viewed it, it's important to look at the 434 patent first at the top of column 8.

[00:24:37] Speaker 03:
This is at appendix 289, where it says that each data handler is operable

[00:24:46] Speaker 03:
independently.

[00:24:47] Speaker 03:
This is where the patentee describes what operable independently means.

[00:24:52] Speaker 03:
And it says, starting at the line three of column eight, it says, for example, each data handler is configured to write... Where are you?

[00:25:02] Speaker 03:
I'm at appendix 289 at column eight of the 434 patent, where the patentee explains what operable independently is.

[00:25:16] Speaker 03:
Right.

[00:25:17] Speaker 03:
So at column 8, line 3, it says, for example, and this is an example of operable independently, it says each data handler is configured to write to and or read from the corresponding plurality of data ports of one or more memory devices without being in communication, presumably with any of the other data handlers or other data ports of the memory devices.

[00:25:44] Speaker 03:
As such, each data handler can be used to generally independently test a portion of the memory space of the memory module.

[00:25:56] Speaker 03:
So that's the patentee's example of what operable independently means.

[00:25:59] Speaker 03:
If we turn back to appendix 1079, Judge Proce, where we were looking at Hwang, you can see figure two right there.

[00:26:10] Speaker 03:
There's no indication that any of the individual

[00:26:14] Speaker 03:
transparent memory interfaces are communicating with each other.

[00:26:19] Speaker 03:
And in the second, or actually the first full paragraph on Appendix 1089, it explains that, however, for transparent BIS, it is preferred

[00:26:35] Speaker 03:
to have the test interface generate test patterns locally.

[00:26:38] Speaker 03:
You're at 1079, not 1089.

[00:26:41] Speaker 03:
I'm sorry if I said 1089.

[00:26:42] Speaker 03:
I'm at 1079.

[00:26:43] Speaker 03:
Oh, OK.

[00:26:44] Speaker 03:
I'm sorry.

[00:26:45] Speaker 03:
So that's describing Wang's figure two.

[00:26:50] Speaker 03:
And in the first full paragraph, second sentence, it says, however, for transparent BIST, it is preferred to have the test interface generate test patterns locally to avoid test data routing from test controller to each memory array.

[00:27:05] Speaker 03:
So we have each transparent memory interface corresponds to a particular memory module, generates the data locally, and there's no evidence in playing that they communicate with each other while they are doing that.

[00:27:24] Speaker 03:
Another characteristic of operable independently from column 8 is that the data handlers are generally modular.

[00:27:32] Speaker 03:
in at appendix 1079 in figure 2 you see the transparent memory interfaces are generally modular.

[00:27:40] Speaker 03:
So the board really just took the characteristics of operable independently from the top of column 8 in the 434 patent and applied that to the Wang reference and made a determination that they operate independently.

[00:27:56] Speaker 03:
I see that I have about a minute

[00:27:59] Speaker 03:
left.

[00:27:59] Speaker 03:
I know that there is a cross appeal.

[00:28:01] Speaker 03:
I want to make sure I give the panel the opportunity to ask questions with respect to the cross appeal.

[00:28:08] Speaker 03:
Otherwise, I'm happy to cede the remaining of my time.

[00:28:15] Speaker 01:
Just on the cross appeal.

[00:28:18] Speaker 01:
Sandis argues that even though its petitions were unsupported with respect to weighing alone, such error was, quote, washed clean.

[00:28:26] Speaker 01:
by the institution decision and it should have been allowed to submit new evidence post institution as part of its reply.

[00:28:33] Speaker 01:
Assign me to a case that extends this exception to allowing a petitioner to submit new evidence post institution to supplement a ground raised in its petition.

[00:28:49] Speaker 03:
I don't have a case off of the top of my head for that.

[00:28:51] Speaker 03:
The Administrative Procedures Act I think stands for the

[00:28:55] Speaker 03:
for the proposition that if the board is going to institute a new ground that's not in the petition, that the petitioner is allowed to support that new ground.

[00:29:08] Speaker 01:
I am quite certain that if there was a case, you'd have it on the top of your head.

[00:29:14] Speaker 01:
So, okay.

[00:29:15] Speaker 01:
Thank you.

[00:29:18] Speaker 01:
Thank you, Your Honor.

[00:29:28] Speaker 00:
Your honors, a few quick points.

[00:29:30] Speaker 00:
First of all, in response to what counsel just said, the Alpert deposition testimony was never made part of the petition or the reply.

[00:29:40] Speaker 00:
The theory was never adopted.

[00:29:42] Speaker 00:
There was nothing to move to strike.

[00:29:43] Speaker 00:
There was nothing to respond to.

[00:29:46] Speaker 00:
We had just heard a lot of testimony that the board was about whether the Wang data handlers communicate with each other.

[00:29:52] Speaker 00:
But that's not the claim construction.

[00:29:54] Speaker 00:
The claim construction is without influence or control by another.

[00:29:58] Speaker 00:
and I walked it through why one influences the control of another.

[00:30:01] Speaker 00:
So turning then to the cross appeal, there is no case that allows counsel to sit back until the reply to bring in evidence for the first time that was necessary to establish a prima facie case.

[00:30:16] Speaker 00:
It's not there and the board basically said as much, we have rules, our rules are that the petition must have

[00:30:25] Speaker 00:
I detailed explanation of the significance of the evidence.

[00:30:28] Speaker 04:
Doesn't this depend on the board's conclusion that the ground it instituted on was raised in their petition?

[00:30:35] Speaker 04:
That they did raise an alternative?

[00:30:37] Speaker 00:
Yeah, and let me address that briefly.

[00:30:39] Speaker 04:
Well, just let me finish.

[00:30:40] Speaker 04:
I mean, if that's the case, then they don't have an argument because that's the thing they raised in their petition.

[00:30:46] Speaker 04:
They should support it then.

[00:30:47] Speaker 04:
If the board did raise an entirely new ground for rejection, certainly they're allowed to present evidence to respond to that new ground for rejection.

[00:30:57] Speaker 04:
They aren't required to anticipate the board's entirely new ground.

[00:31:02] Speaker 00:
Yeah, a couple of points to that, Your Honor.

[00:31:03] Speaker 00:
First, the board did make it clear that its decision to review on weighing alone was based solely on evidence expressly cited by the petitioner in its petition.

[00:31:12] Speaker 00:
And it did go on to say that there are procedural safeguards, and this court said,

[00:31:16] Speaker 00:
and Belden versus Birktek, you have to take advantage of your procedural safeguards, and those include not only moving to referee hearing or moving to withdraw the additional ground that you had to disagree with, but most importantly, moving to supplement, to provide supplemental information in support of the instituted ground.

[00:31:36] Speaker 00:
And the board made it clear, and that's what the rule expressly allows you to do.

[00:31:39] Speaker 00:
There's no rule that says I can sit back and wait to my reply, and for good reason, because that would really sandbag

[00:31:46] Speaker 00:
and disrupt the whole process.

[00:31:50] Speaker 00:
But the rules are clear.

[00:31:51] Speaker 04:
So your view is, if they had thought that it was a new ground, they should have taken steps before the reply to get their information in.

[00:32:00] Speaker 00:
Absolutely, and that was the reasoning in part from the board.

[00:32:03] Speaker 00:
The board said, we have these procedural safeguards.

[00:32:06] Speaker 00:
You have the right to put in supplemental evidence.

[00:32:08] Speaker 00:
And I thought at one last point, if I may jump to the end,

[00:32:13] Speaker 00:
With respect to HUANG and the reference to CERN, there is no explanation of any alleged common knowledge as a basis for CERN or in the petition.

[00:32:24] Speaker 00:
So the case how they cited about common knowledge doesn't apply.

[00:32:28] Speaker 00:
And the reference to a data handler, data handlers are to generate, modify, and transform.

[00:32:33] Speaker 00:
But CERN was a data buffer and a data buffer doesn't generate, modify, or transform.

[00:32:38] Speaker 00:
So there was no explanation as to any relevance they need to tie

[00:32:43] Speaker 00:
any alleged significance of the common knowledge, if they allege that it was there, they need to tie it to an adequate explanation.

[00:32:51] Speaker 00:
And you can find that in the Ariosa Diagnostics case that they cited in their brief.

[00:32:56] Speaker 00:
Thank you.

[00:32:56] Speaker 00:
Thank you, Your Honor.

[00:33:01] Speaker 05:
One minute.

[00:33:03] Speaker 05:
Just on the cross appeal.

[00:33:06] Speaker 03:
Thank you, Your Honor.

[00:33:07] Speaker 03:
On the cross appeal, the rules that the board has implemented, I don't

[00:33:12] Speaker 03:
think we're intended to cover this situation where the board institutes on a new ground, first of all.

[00:33:18] Speaker 03:
Second of all, we didn't disagree with the board that... Your argument depends on the fact that this is a new ground.

[00:33:24] Speaker 04:
If we agree with the board that this wasn't a new ground, you're done on the crossfield, right?

[00:33:31] Speaker 03:
That's right.

[00:33:32] Speaker 03:
I don't think there really can be any dispute that the petition

[00:33:35] Speaker 03:
did not raise Wang's loan.

[00:33:37] Speaker 04:
Our expert may have said... I think there can be a lot of dispute the way you worded your petition.

[00:33:41] Speaker 04:
But even if it is a new ground, why didn't you ask to supplement the record or the like before your reply?

[00:33:50] Speaker 03:
Well, so first, I don't think we felt that the record needed to be supplemented because this trivial

[00:34:04] Speaker 04:
implementation is something that... Well, if it didn't need to be supplemented, then why did you try to put in new evidence with your reply?

[00:34:12] Speaker 03:
Well, so the supplemental, the submission of supplemental information in the rules, the board, in our experience practicing before the board, routinely denies those requests.

[00:34:24] Speaker 03:
They're typically granted for instances where a patent owner makes a challenge to whether or not it references a

[00:34:33] Speaker 03:
printed publication or something like that.

[00:34:38] Speaker 03:
That's typically the scenario where the board allows the submission of supplemental evidence.

[00:34:46] Speaker 03:
But for these trivial implementation details, it wasn't a matter of putting evidence into the record.

[00:34:53] Speaker 03:
It was a matter of arguments.

[00:34:56] Speaker 03:
So what

[00:35:00] Speaker 03:
It was a matter of saying that this is a trivial implementation detail.

[00:35:07] Speaker 03:
We did not need to rely on CERN.

[00:35:10] Speaker 03:
CERN shows this unambiguously.

[00:35:14] Speaker 03:
The evidence was sufficient in the record to go forward.

[00:35:20] Speaker 03:
We didn't have an issue with the board's institution decision.

[00:35:27] Speaker 05:
Time's up.

[00:35:28] Speaker 05:
Thank you.

[00:35:28] Speaker 05:
We thank both sides.

[00:35:29] Speaker 05:
The case is submitted.