[00:00:00] Speaker 03:
We have four argued cases this morning.

[00:00:02] Speaker 03:
The first is number 23, 1622, Google versus MindBase HQ.

[00:00:08] Speaker 03:
Mr. Kelly.

[00:00:10] Speaker 00:
Good morning, Your Honors.

[00:00:11] Speaker 00:
May it please the board.

[00:00:12] Speaker 00:
The board was mostly right when it determined that most of MindBase's claims to databases of words were not patentable.

[00:00:21] Speaker 00:
But it erred in two significant respects that led it to the wrong conclusion for the dictionary routine claims and for the claims directed to normalization.

[00:00:30] Speaker 00:
For the dictionary routine claims, the flaw in the board's decision was that it failed to construe dictionary routine as requiring only the use of a dictionary as a lookup function as opposed to the creation of a dictionary.

[00:00:44] Speaker 00:
The patent, mind-basis patent, explains how its dictionary works.

[00:00:49] Speaker 00:
It compares an entered word with a word in the dictionary, just like we all do when we use a dictionary, as a lookup function.

[00:00:56] Speaker 00:
The board got to the wrong place, though, because the board construed the claim, or it didn't construe the claim, but it applied the claim as if it required dictionary creation, which it doesn't.

[00:01:06] Speaker 03:
And even mind-based doesn't.

[00:01:08] Speaker 03:
Does the specification describe both dictionary creation and lookup?

[00:01:14] Speaker 00:
It describes the use of the dictionary to look up the word that's entered, and it does have discussions about dictionary creation, but the claim

[00:01:23] Speaker 00:
limitation is about the lookup function.

[00:01:27] Speaker 00:
That's consistent with what MindBase put in its infringement contentions that it served us with.

[00:01:32] Speaker 00:
In fact, MindBase has been able to never take a position on whether or not its claims are broad enough to read simply on a lookup function.

[00:01:42] Speaker 04:
But the claim limitation is a dictionary routine for automatically classifying and storing words.

[00:01:51] Speaker 04:
In your construction, what is the automatic classifying and storing?

[00:02:00] Speaker 00:
It's the return of the definition that's in the Conlin database.

[00:02:05] Speaker 00:
So yes, it talks about automatically classifying.

[00:02:08] Speaker 00:
And the way the board and mind base sort of gets around Conlin is Conlin uses the phrase semi-automatic.

[00:02:15] Speaker 00:
And Conlin used that phrase in terms of how it creates its table three.

[00:02:20] Speaker 00:
Now what Conlin does is it starts with dictionary definitions from multiple databases.

[00:02:24] Speaker 00:
It pulls them in, that's in Conlin's table two.

[00:02:27] Speaker 00:
And then through methods it calls semi-automatic, it creates its table three.

[00:02:32] Speaker 00:
That's the table that's in the Conlin database that's used to provide the definitions of terms.

[00:02:38] Speaker 00:
And it's in that table three that we see the classifications that correspond to the types of words in claim 43 that have to do with, I'm sorry,

[00:02:49] Speaker 00:
in claim 14 that have to do with verbs, they have to do with tangible objects.

[00:02:54] Speaker 00:
Those are the classifications that need to be returned.

[00:02:57] Speaker 04:
And that is what's in there.

[00:02:58] Speaker 04:
I think there's just another way of asking the same question.

[00:03:01] Speaker 04:
Sure.

[00:03:01] Speaker 04:
But at page 43, the board says, in mapping this limitation 14.3, which I think is the one we're talking about, and onto Colin, petitioner has not shown how looking up a word, for example, the word chip, discloses limitation 14.3.

[00:03:18] Speaker 04:
In part, this petitioner and your expert have not explained how yielding the word's definition and lexical information and storing that query discloses automatically classifying and storing words entered into said database.

[00:03:33] Speaker 04:
I think you feel you've already answered that question, but can you try again and help me understand your answer to that?

[00:03:39] Speaker 00:
Absolutely, Your Honor.

[00:03:41] Speaker 00:
So in that part of the decision that the Court is just reading, that's the part where the Board purported to accept our construction and said, well, even under your construction, Conlon doesn't do it because of the automatically classifying part.

[00:03:53] Speaker 00:
The problem with that analysis is that if you agree with our construction, that getting the dictionary to return the definition is within the scope of our construction, then if that definition includes the classification of the word,

[00:04:08] Speaker 00:
In other words, if it's a verb or a tangible object, that gets returned with the definition that is the classification.

[00:04:14] Speaker 00:
So the word that got put in is automatically classified in the sense that the classification is returned by the dictionary.

[00:04:20] Speaker 04:
The classification is telling the user what type of word it is.

[00:04:26] Speaker 00:
Yes.

[00:04:26] Speaker 04:
It's a noun, it's a verb, it's a whatever.

[00:04:28] Speaker 00:
Right.

[00:04:28] Speaker 00:
And that's what's in Conlin's table three.

[00:04:31] Speaker 00:
And Your Honor, to go back, Judge Seig, to your question about how

[00:04:37] Speaker 00:
the mind-based system builds its dictionary.

[00:04:39] Speaker 00:
You'll see an example at the bottom of column 14, where mind-based discusses this return of the classification of the word.

[00:04:47] Speaker 00:
And mind-based gives an example involving a red car.

[00:04:50] Speaker 00:
And it says, well, look, if the person puts in red car as simply a tangible object, the mind-based dictionary will switch that to car.

[00:04:59] Speaker 00:
It'll say it's a tangible object, and it's just a car.

[00:05:03] Speaker 00:
And then it lets the user respond

[00:05:05] Speaker 00:
to confirm or to say that that's not correct.

[00:05:09] Speaker 00:
So in other words, if the definition with the classification that comes back is somehow flawed, that's when the user's interaction comes in and that can modify MindBase's dictionary.

[00:05:20] Speaker 03:
But that suggests that the board's claim construction was wrong in terms of what automatic means and maybe the semi-automatic feature of Conlin would come within the correct claim construction.

[00:05:34] Speaker 03:
No?

[00:05:35] Speaker 00:
No, Your Honor, because the semi-automatic function of Conlin has to do with the building of the database, which is, I think, what Your Honor is talking about with that example at the bottom of claim column 14.

[00:05:46] Speaker 00:
But that, again, is not what the claim is directed to, because the claim is directed to returning the classification, not sort of building the dictionary.

[00:05:54] Speaker 00:
And that is precisely how MindBase has characterized it.

[00:05:59] Speaker 03:
No, I think you misunderstood my question.

[00:06:00] Speaker 03:
My question is,

[00:06:02] Speaker 03:
Looking at the description of dictionary creation in the specification, it seems as though dictionary creation is perhaps not automatic, that it requires user input.

[00:06:13] Speaker 03:
So perhaps the board's construction of the word automatic in the claim is wrong, and that automatic ought to be allowing for some amount of user input.

[00:06:28] Speaker 03:
given the description of dictionary creation in the specification.

[00:06:35] Speaker 03:
And that would perhaps mean that under the construction that I'm suggesting, Conlon would disclose dictionary creation.

[00:06:49] Speaker 00:
Your Honor, that would be a third construction of the claim.

[00:06:52] Speaker 00:
It's frankly not a construction that

[00:06:56] Speaker 00:
we've advocated for so far, but I think it illuminates the problem here, is that whatever the board did with this claim was flawed.

[00:07:04] Speaker 00:
And what we need is clarity about what this claim term means.

[00:07:08] Speaker 00:
Now, Your Honor, I would push back a little bit on that construction, because what the final limitation claim 14 is talking about is simply the classification and the storage of the entered word.

[00:07:20] Speaker 00:
And to the extent that anything meets the scope of that claim, it wouldn't matter if it's also done semi-automatically.

[00:07:28] Speaker 00:
And what I mean by that is even if MindBase has two functions, that is one is just returning a definition with a classification in it, and the second is adding to the dictionary by doing the semi-automatic sort of correction of the classifications, that it wouldn't matter that that second step is also contemplated by MindBase as long as the claim covers

[00:07:50] Speaker 00:
the first step of just returning the classification, because that's what Conlon does also.

[00:07:55] Speaker 00:
But I do understand your Honor's point, and again, MindBase has been able to remain silent on this throughout this whole proceeding, even in this case.

[00:08:05] Speaker 00:
In its red brief, it doesn't take a position on whether or not the claim reads on simply a look-up function.

[00:08:12] Speaker 00:
And it's doing that to keep its options open down the line.

[00:08:15] Speaker 00:
And it's critical not just for the dispute before the court from the board, but also because of the ongoing district court action that this court, however it views the scope of the claim, provides some clarity to the parties.

[00:08:27] Speaker 02:
If we agree that

[00:08:30] Speaker 02:
that the board didn't actually apply your construction.

[00:08:34] Speaker 02:
Should you send it back and have the board itself review its construction decision, or are you saying that it should adopt your construction and the alternative?

[00:08:43] Speaker 00:
Well, Your Honor, as for our construction argument, I mean, that is, of course, subject to de novo review, and we would advocate for this Court adopting our construction, or at least definitively determining what the construction is.

[00:08:54] Speaker 00:
And then, yes, a remand would be in order for the Board to either apply that construction, and even if that construction can only come out one way, there are claims dependent

[00:09:04] Speaker 00:
from claim 14 that would require further analysis from the board.

[00:09:07] Speaker 04:
So in a remand... I apologize for your answer.

[00:09:09] Speaker 04:
I thought you were arguing that it was undisputed that Conlin discloses this limitation under your construction.

[00:09:18] Speaker 04:
Were we to adopt your construction?

[00:09:21] Speaker 00:
That's not what the board said.

[00:09:23] Speaker 00:
I mean, that's certainly what we think, and we think it's very easy to get there.

[00:09:26] Speaker 00:
There would still be a remand required, because there are other claims that the board didn't reach.

[00:09:30] Speaker 00:
And I guess maybe I didn't say my answer quite correctly.

[00:09:33] Speaker 00:
But yes, on claim 14, if our construction is correct, there's no other way to read Connell.

[00:09:39] Speaker 04:
Can I ask you on the normalization claims?

[00:09:42] Speaker 04:
I understand what the board was saying, I think, is,

[00:09:46] Speaker 04:
you made a new argument in your reply that wasn't in your petition.

[00:09:50] Speaker 04:
Can you show me where in your petition you think you made the argument that the board says was new?

[00:09:58] Speaker 00:
Yes.

[00:10:01] Speaker 00:
So I'm going to preface this by acknowledging that our petition's discussion of the final three limitations of the normalization claims was, say, terse.

[00:10:15] Speaker 00:
And we see this at page 2548 of the record.

[00:10:28] Speaker 00:
So the way claim 43 is written.

[00:10:36] Speaker 02:
What was that?

[00:10:36] Speaker 00:
2548.

[00:10:36] Speaker 00:
2548, Your Honor.

[00:10:44] Speaker 00:
So the way Claim 43 is written is that in its initial part of the claim, it has the normalization functions in it.

[00:10:53] Speaker 00:
So what we've called deduplicating, what we've called disambiguating, that all occurs in what I'll call the top half of Claim 43 when it compares within each of the databases that are inbound, whether those databases have duplicates or need disambiguating.

[00:11:10] Speaker 00:
But once that is done, that's when we get to the final three steps

[00:11:14] Speaker 00:
of the claim.

[00:11:15] Speaker 00:
And that's what's treated at page 2548, which is page 83 of our record.

[00:11:19] Speaker 00:
And yes, we say, okay, once that occurs, in step 43.3

[00:11:25] Speaker 00:
You compare the normalized databases.

[00:11:27] Speaker 00:
And then in step 43.4, you record the common data elements.

[00:11:31] Speaker 00:
And then finally, you just store one location of it.

[00:11:35] Speaker 00:
And we cite our expert on each one of those.

[00:11:39] Speaker 00:
And our expert cites Fong on each one.

[00:11:41] Speaker 00:
And it cites Conlon.

[00:11:43] Speaker 00:
And it cites other part.

[00:11:44] Speaker 04:
You don't say anything about inherency, correct?

[00:11:47] Speaker 00:
Oh, we do not use that word here, Your Honor, because the theory of our case, and this is what is discussed a little bit higher up

[00:11:55] Speaker 00:
is that it would have been obvious at least to take Conlon and when doing the normalization to follow the steps in Fong.

[00:12:00] Speaker 00:
So what our expert did is said, okay, if you follow the steps in Fong when doing that normalization, these will be the final three steps and when the board expressed some uncertainty about that,

[00:12:13] Speaker 00:
That's when we came back with the rebuttal declaration, where Dr. Jansen took great efforts to stay true to his original theory.

[00:12:21] Speaker 00:
Because in each of the three or four paragraphs where he discusses this issue, he begins by quoting the paragraph in his initial declaration.

[00:12:30] Speaker 00:
And then he says, OK, here's what I said.

[00:12:33] Speaker 00:
Here's what I pointed to.

[00:12:34] Speaker 00:
And then, yes, he builds on it by saying, I'm going to use a simple sort of tutorial device.

[00:12:39] Speaker 00:
I'm going to use this very simple explanation with apples and oranges and limes, just a few components.

[00:12:46] Speaker 00:
And as he goes through, he says, here is how you would do Fong's normalization steps.

[00:12:53] Speaker 03:
If we were to assume that the board got it wrong in saying the reply was improper, and we conclude that the reply was proper,

[00:13:02] Speaker 03:
What is the board saying about the deficiencies in the reply?

[00:13:07] Speaker 03:
I know with respect to one limitation, it says there wasn't any text.

[00:13:10] Speaker 03:
There was just a figure.

[00:13:13] Speaker 03:
But with respect to the others, what's their view as to what's the matter with the reply showing?

[00:13:20] Speaker 00:
Your Honor, I'm not sure I can answer that, because what the board said is it just wasn't going to accept the reply.

[00:13:26] Speaker 00:
It wasn't going to consider it.

[00:13:27] Speaker 03:
So it didn't go on and say, even if we accept the reply, it's insufficient?

[00:13:35] Speaker 00:
No, Your Honor.

[00:13:35] Speaker 00:
It said, and this is at page, I believe, 57 of the record, it says, we decline to consider these late arguments and evidence.

[00:13:45] Speaker 00:
And that's four lines from the bottom of page 57.

[00:13:49] Speaker 00:
So it outright declined to consider a rebuttal declaration that couldn't have been truer to the original declaration.

[00:13:54] Speaker 00:
This is exactly why we need rebuttal declarations.

[00:13:57] Speaker 00:
This is exactly what parties should be permitted to do.

[00:14:00] Speaker 00:
And MindBase could have cross-examined Dr. Jansen if they thought there was any problems with his rebuttal declaration, and they didn't do so.

[00:14:09] Speaker 00:
And unless the court has any further questions, I'll reserve the final minute.

[00:14:12] Speaker 03:
OK, we'll give you two minutes.

[00:14:19] Speaker 01:
I'll start with the reply evidence since that's what Google led off with.

[00:14:46] Speaker 01:
Now Google argues that their 119 page

[00:14:49] Speaker 03:
for bottle declaration should be admissible because... So we're just passing over the question of the propriety of the reply.

[00:14:57] Speaker 03:
Let's assume hypothetically that we said the reply is correct.

[00:15:01] Speaker 03:
What's the matter on the normalization claims with respect to the reply showing?

[00:15:10] Speaker 01:
The argument that Google makes with respect to normalization is that the word normalization

[00:15:18] Speaker 01:
exists prior to mind-based invention.

[00:15:22] Speaker 01:
That's not in dispute.

[00:15:24] Speaker 01:
We don't dispute that.

[00:15:25] Speaker 01:
As far as what they're normalizing, it's the detail of

[00:15:28] Speaker 01:
of how the board is approaching the normalization, as with some other terms, that they may see that, and that's enough for them.

[00:15:37] Speaker 01:
And if I could just make one point on the reply, there is the key to... I'm not understanding your argument.

[00:15:44] Speaker 03:
Are you saying that if the reply is accepted, it was sufficient to... No.

[00:15:52] Speaker 01:
No, the detail is there again, because the normalization

[00:15:54] Speaker 01:
I know for a lay audience, the normalization is inherent in any database management system.

[00:16:01] Speaker 01:
And so it's the details of that that make it specific to a particular invention or version of a DBMS.

[00:16:07] Speaker 01:
So no.

[00:16:08] Speaker 01:
So the reply argument testimony is not enough, because it nearly talks about fall, which is not anticipatory or rendering obvious of the patterns.

[00:16:21] Speaker 02:
Normalization, that's the elimination of redundancies, isn't it?

[00:16:25] Speaker 01:
It's one of the things.

[00:16:26] Speaker 01:
And so for FOMA, which is a textbook, entry-level textbook, all database management.

[00:16:31] Speaker 01:
So that would be different than just plain looking up.

[00:16:34] Speaker 01:
Correct.

[00:16:35] Speaker 01:
That's an internal process of the database management system.

[00:16:39] Speaker 01:
So depending on the structure of the DBMS, you have to have a different normalization process internally.

[00:16:45] Speaker 01:
So that's why MindBase doesn't claim to have invented normalization.

[00:16:50] Speaker 01:
They're just showing how they need to normalize their innovative DBMS.

[00:16:57] Speaker 04:
I do want to come back to whether the reply was proper or not.

[00:17:00] Speaker 04:
It sounded like you might want to get there, too.

[00:17:03] Speaker 04:
My first question is, whether it was new or not, is that a matter?

[00:17:07] Speaker 04:
I think that's a matter we reviewed de novo ourselves.

[00:17:10] Speaker 04:
Do you agree with that?

[00:17:11] Speaker 04:
Yes.

[00:17:12] Speaker 04:
OK.

[00:17:14] Speaker 04:
And also, it appears that you did not, or at least I can't find, where you argued to the board

[00:17:20] Speaker 04:
that the rebuttal was improperly new.

[00:17:23] Speaker 04:
Can you show me where you made that argument, if you did make that argument?

[00:17:28] Speaker 04:
I see it 4177, where you responded on the merits.

[00:17:32] Speaker 01:
I think we just took it to respond on the merits, because we felt that they were still off.

[00:17:40] Speaker 01:
You're right, we probably had that procedural avenue that we could have.

[00:17:43] Speaker 04:
So the board found it was improper and new, even though you didn't ask it to do that.

[00:17:48] Speaker 01:
I think we argued in the oral hearing at the board that whether they weighed it lightly or whether they ignored it altogether was up to them.

[00:18:05] Speaker 01:
But to your point, we didn't go that route very deeply.

[00:18:11] Speaker 04:
Mr. Kelly, tell us this is exactly the type of rebuttal or reply

[00:18:16] Speaker 04:
declaration that we should want the board to allow.

[00:18:20] Speaker 04:
What's wrong with his view on that?

[00:18:22] Speaker 01:
Well, so if a petitioner is allowed to reply with 119 papers, that's longer than almost any other document in this whole case.

[00:18:34] Speaker 01:
I mean, to me, it's facts of Google's trying to game the system, thinking that the respondent is not going to be able to rebut the rebuttal declaration.

[00:18:45] Speaker 03:
Well, they're not relying on any new art in the reply.

[00:18:48] Speaker 03:
It's just a question of elaborating on what Fong teaches, right?

[00:18:53] Speaker 01:
But this is all expert testimony, right?

[00:18:55] Speaker 01:
This is all just expert declarations.

[00:18:57] Speaker 03:
Wait, wait, wait.

[00:18:59] Speaker 03:
Is my statement correct?

[00:19:01] Speaker 03:
That they're just, in the reply, elaborating on what Fong teaches?

[00:19:05] Speaker 01:
Elaborating on a previously disclosed prior art.

[00:19:10] Speaker 01:
But the elaborating prior art, I mean, this is all new testimony.

[00:19:15] Speaker 01:
these are new charts and graphs.

[00:19:17] Speaker 01:
They even use it through oral hearing.

[00:19:18] Speaker 01:
They use it in their briefs here.

[00:19:20] Speaker 01:
And we argued that that volume, the 119 pages, was too much on the face of it.

[00:19:28] Speaker 01:
How could that be all respondent to just what they had when it's one of the largest arguments in the case?

[00:19:35] Speaker 01:
But the board did not reject this based on the length of it at all.

[00:19:40] Speaker 01:
Instead, the board described the new evidence

[00:19:46] Speaker 01:
as petitioners argued, were not made in the petition.

[00:19:49] Speaker 01:
And at the hearing, petitioners stated that the reply went deeper into the analysis, so the thrust of the analysis did not change.

[00:19:57] Speaker 01:
However, the board continues, inherently is not argued in the petition.

[00:20:03] Speaker 01:
And we note that the arguments in the reply and evidence relied upon, including a rebuttal declaration, is significantly more detailed.

[00:20:09] Speaker 01:
That's at page 55, 56.

[00:20:12] Speaker 01:
They did consider it.

[00:20:14] Speaker 01:
They did go through it and see what was offered.

[00:20:18] Speaker 01:
It wasn't a matter just of the timing.

[00:20:20] Speaker 01:
It wasn't a matter of the length, as we proposed.

[00:20:26] Speaker 01:
I'd like to move to another topic.

[00:20:29] Speaker 01:
Yeah, you ought to talk about the dictionary claims.

[00:20:34] Speaker 01:
So claims 43, the dictionary routines.

[00:20:39] Speaker 01:
I'm not sure if Google is trying to construe

[00:20:42] Speaker 01:
automatically classifying and soaring, or if they're trying to construe creating.

[00:20:48] Speaker 01:
But somehow they've come to the conclusion on their own that those two are not the same.

[00:20:53] Speaker 01:
I mean, from the beginning, from the petition and every step of the way, Google has said, we want to rely on plain and ordinary meaning.

[00:21:00] Speaker 01:
And that's what my base has said.

[00:21:02] Speaker 01:
But now they say, well, this construction provides that we'll be right.

[00:21:07] Speaker 01:
Where was this construction in the petition?

[00:21:10] Speaker 04:
So assume that we feel we have a claim construction dispute in front of us about the claim term automatically classifying and storing words.

[00:21:19] Speaker 04:
Do you oppose the construction that Google is presenting, at least to us?

[00:21:25] Speaker 01:
I don't think I understand how their reference that uses the word semi-automatically can be automatically.

[00:21:34] Speaker 04:
That's an application of the construction, and I'm happy to hear what you have to say about that.

[00:21:38] Speaker 04:
Do you accept their construction, or do you oppose their construction?

[00:21:41] Speaker 04:
We oppose it.

[00:21:42] Speaker 04:
So what's wrong with their construction, which says you don't need to create the database.

[00:21:48] Speaker 04:
You only have to use it as a lookup.

[00:21:50] Speaker 04:
That's essentially their construction of this term.

[00:21:54] Speaker 01:
Like I say, they're trying to construe the word create to not equal.

[00:21:57] Speaker 01:
I don't understand what is creating a dictionary.

[00:22:02] Speaker 04:
Is that the extent of your objection to their proposed construction?

[00:22:06] Speaker 01:
Their reference does not, their prior reference does not literally say it, so they now construe it and, you know, they're just, you know, I just don't feel, how was it creating it?

[00:22:19] Speaker 01:
One interpretation, I'm not paraphrasing the patents here, the claim terms, but creating a dictionary, isn't it classifying, you know, making a,

[00:22:31] Speaker 01:
a determination, a definition of word, and then storing it?

[00:22:34] Speaker 01:
Isn't that what creating a dictionary is?

[00:22:39] Speaker 01:
So I just don't understand how the A to B. So I disagree with their construction, yes.

[00:22:45] Speaker 01:
But my problem was with it.

[00:22:47] Speaker 04:
Then how about the application?

[00:22:48] Speaker 04:
If we were to disagree with you and adopt their construction of this automatically classifying and storing words term, do you have a contention

[00:22:59] Speaker 04:
that Conlin does not disclose or make obvious that limitation under their construction?

[00:23:12] Speaker 01:
To me, Conlin does not make a... It's a collection... I feel like we are getting... The word database here is being used differently.

[00:23:24] Speaker 01:
Conlin is using it as a collection of words.

[00:23:27] Speaker 01:
the collection of information, and the patents are disclosing a DBMS.

[00:23:34] Speaker 01:
But to your point, yes, I disagree.

[00:23:37] Speaker 01:
Could you repeat the question?

[00:23:42] Speaker 04:
Let me ask it this way.

[00:23:45] Speaker 04:
There are references I'll give you in Conlin that talk about we can't do certain things automatically.

[00:23:50] Speaker 04:
We can only do them semi-automatically.

[00:23:54] Speaker 04:
I understand Google would say all of those

[00:23:57] Speaker 04:
statements deal with creating a database, not with just merely using it for lookout.

[00:24:04] Speaker 04:
Do you agree?

[00:24:05] Speaker 04:
I don't mean do you agree that's what they argue, but do you agree that they're right about that?

[00:24:10] Speaker 01:
I don't believe that Collin uses it that way.

[00:24:12] Speaker 01:
My first question would be, if Collin does do something automatically,

[00:24:16] Speaker 01:
I don't recall anywhere in Collin that they use that term automatically.

[00:24:21] Speaker 01:
They always use the term semi-automatically.

[00:24:23] Speaker 01:
It's not like they separate.

[00:24:25] Speaker 04:
But do they ever use the term semi-automatically to refer to anything other than creation?

[00:24:31] Speaker 01:
As I recall, Collin uses semi-automatically several times throughout their disclosure.

[00:24:41] Speaker 03:
Is it correct?

[00:24:44] Speaker 03:
As I was suggesting earlier, the specifications description of dictionary creation isn't automatic in the sense that it requires user input.

[00:24:59] Speaker 01:
I think it describes several ways, and I think it does describe an automatic creation of the specification for both patterns.

[00:25:12] Speaker 03:
I thought the description of the dictionary creation and the specification contemplated use for input.

[00:25:28] Speaker 03:
Am I wrong about that?

[00:25:30] Speaker 01:
I think it describes several ways of doing it.

[00:25:35] Speaker 01:
Where does it describe fully automatic?

[00:25:47] Speaker 01:
Would I be able to find that for you on rebuttal time?

[00:26:04] Speaker 01:
Sure.

[00:26:05] Speaker 01:
Thank you.

[00:26:06] Speaker 03:
And I know you wanted to talk about your cross appeal, which we haven't gotten to yet, so we'll give you a few minutes to do that.

[00:26:14] Speaker 03:
Thank you.

[00:26:15] Speaker 01:
Well, first, with regard to claims 43 through 46, Google alleges that they, or rather, the board ruled that they had not shown accountability for the recording one location of each said common data element.

[00:26:37] Speaker 01:
But Google says, the reason is Google relies on, if we look at the experts, Google's experts

[00:26:44] Speaker 01:
declaration regarding claim nine.

[00:26:48] Speaker 01:
Claim nine is a claim merely that says each said data element is stored only once.

[00:26:54] Speaker 01:
And the expert describes that in a paragraph.

[00:26:56] Speaker 01:
And then for claim 43.5, the expert merely says, as explained above for claim nine, each data element is stored only once in the main table.

[00:27:07] Speaker 01:
But that's not what claim 43 says.

[00:27:10] Speaker 01:
So he's using the same

[00:27:12] Speaker 01:
paragraph for claim 9 for two different claims.

[00:27:15] Speaker 01:
Claim 9, as I just said, is each set data element stored only once.

[00:27:18] Speaker 01:
Claim 435 is recording one location of each set common data element, each of set databases.

[00:27:26] Speaker 01:
And his expert explanation, and the same is just as explained above for claim 9, each data element stored only once.

[00:27:34] Speaker 01:
And the same is true for 44.6.

[00:27:36] Speaker 01:
Again, this is a third different type of recording of information.

[00:27:41] Speaker 01:
It's not even

[00:27:42] Speaker 01:
It's not even a data element stored.

[00:27:44] Speaker 01:
And it's not only once.

[00:27:45] Speaker 01:
It's more detailed.

[00:27:48] Speaker 01:
It's more copy on it.

[00:27:50] Speaker 01:
But they just regurgitate.

[00:27:51] Speaker 01:
The expert just regurgitates the same basis for it.

[00:27:55] Speaker 01:
And the board pointed that out.

[00:27:57] Speaker 01:
The board says, this is the full extent of Google's analysis.

[00:28:02] Speaker 01:
This is page 58.

[00:28:05] Speaker 01:
Paragraph 2, the board says, this is the full extent of Google's analysis of this limitation.

[00:28:11] Speaker 01:
Titian did not explain or was it rather apparent how the analysis for claim 9 and limitation 10.2 taught the recited limitation in 44.6.

[00:28:22] Speaker 01:
And that applies to claims 43, 45, 43, and 44, of which have two dependent claims, 45 and 46.

[00:28:29] Speaker 01:
So for those claims, at least based on those elements,

[00:28:35] Speaker 01:
The board has pointed out exactly, and Google hasn't rebutted, how this one paragraph that their expert used for claim nine should apply when the claim terms are all different, 43 and 46.

[00:28:46] Speaker 01:
If I could move on to one more point.

[00:28:54] Speaker 03:
Well, are you going to discuss the cross-bill?

[00:28:58] Speaker 01:
I want to get on to this one point.

[00:29:02] Speaker 01:
I understand what you're saying.

[00:29:06] Speaker 01:
Uh, the, the, the level of ordinary skill in the art, you know, I have the, the citation of supplemental authority and the, um, um, you know, it's the same, but the PTAB, that decision came down a month after the decision in our case.

[00:29:25] Speaker 01:
And then the, the amount of support is the same for there.

[00:29:29] Speaker 01:
So it was a one TPAT PTAB panel.

[00:29:33] Speaker 01:
ruled that it was insufficient, conclusory, and the other, another PTAP.

[00:29:38] Speaker 01:
I also, I mean, if there's a disagreement in the PTAP about what is sufficient evidence for a level of ordinary spill in the arc.

[00:29:45] Speaker 01:
And I, like I said, I have to mention that the board adopted a level of ordinary spill in the arc that includes computational linguistics, computational lexicography, which is in terms which have never been defined.

[00:30:00] Speaker 01:
And then the board uses this

[00:30:02] Speaker 01:
yardstick for all of its unpackability analysis.

[00:30:06] Speaker 02:
The specification, however, is just full to references of words that pertain to lexicon.

[00:30:17] Speaker 02:
And it talks about verbs and nouns and the different usages.

[00:30:22] Speaker 02:
All of that is the stuff that lexicon consists of.

[00:30:27] Speaker 01:
So the invention is a database management system, and they have to describe it in some way that's not bits and bytes of computer information.

[00:30:36] Speaker 01:
And so they've chosen verbs, right?

[00:30:39] Speaker 01:
And they've chosen one or two other words.

[00:30:40] Speaker 01:
But it's not an entire grammatical endeavor that common is undertaking.

[00:30:46] Speaker 01:
It's a DBMS that they just have to describe some way, and they've chosen those words for describing it.

[00:30:56] Speaker 03:
I think we're out of time, but we'll give you a minute for rebuttal.

[00:31:00] Speaker 03:
Mr. Kelly.

[00:31:07] Speaker 00:
Thank you, Your Honors.

[00:31:08] Speaker 00:
Just three quick points.

[00:31:12] Speaker 00:
First, on the dictionary routine, we saw counsel once again try to avoid acknowledging whether or not they believe use of a dictionary is within the scope of the claims.

[00:31:24] Speaker 00:
They tried to avoid saying that through this whole litigation.

[00:31:27] Speaker 00:
I think Your Honor saw counsel try to avoid it again here.

[00:31:31] Speaker 00:
We need clarity on what the dictionary limitation means and specifically whether it reads on mere use of a dictionary as a lookup table.

[00:31:39] Speaker 00:
Second, counsel said repeatedly that the rebuttal declaration was 119 new pages.

[00:31:48] Speaker 00:
That's his theme, that it was this huge addition to the record.

[00:31:54] Speaker 00:
completely inconsistent with the board's decision.

[00:31:58] Speaker 00:
The board did refer to the rebuttal declaration, but the part it refused to consider was five pages, pages 41.10 to 41.14, just five pages, just a handful of paragraphs that elaborated in the very same direction on the very same points that Dr. Jansen said the first time around.

[00:32:18] Speaker 00:
This is exactly the type of rebuttal declaration that the board should have considered and erred by not considering it.

[00:32:24] Speaker 00:
The last thing I'll say on his cross appeal, which is the only thing he raised, was the level of ordinary skill in the art.

[00:32:32] Speaker 00:
And Judge Raina, as you recognized, and as the board recognized, and it goes through this at appendix page

[00:32:40] Speaker 00:
19, 18, 20, it explains that the specification had so many references to words, to things like lexicography, that it was, quote, too numerous to recount.

[00:32:54] Speaker 00:
There was abundant evidence, and certainly substantial evidence, to support the board's finding about the level of ordinary skill in the art.

[00:33:02] Speaker 00:
And unless the court has any questions, I'll sit down.

[00:33:05] Speaker 03:
OK.

[00:33:05] Speaker 03:
Thank you.

[00:33:06] Speaker 03:
Thank you.

[00:33:08] Speaker 03:
Mr. Iconis?

[00:33:22] Speaker 01:
I haven't located that.

[00:33:25] Speaker 01:
I'm confident that the patent says that, but I haven't located it for you.

[00:33:30] Speaker 01:
That if I could... Yes, one minute.

[00:33:40] Speaker 01:
And then one more point on the...

[00:33:46] Speaker 01:
the board's decision.

[00:33:48] Speaker 01:
In their decision, they construed, sua sponte, and for the first time, the term lexical database.

[00:33:55] Speaker 01:
Of course, that term does not exist in the past.

[00:33:58] Speaker 01:
It exists only in the prior art reference.

[00:34:01] Speaker 01:
So whatever definition Google or MindBase was using during the course of the PTAB trial,

[00:34:10] Speaker 01:
It is now in question because the construction is different from what the definition of what's provided for in Conway.

[00:34:18] Speaker 01:
So based on that context.

[00:34:24] Speaker 01:
OK.

[00:34:26] Speaker 03:
Thank you.

[00:34:28] Speaker 03:
Thank both counsel.

[00:34:29] Speaker 03:
The case is submitted.