[00:00:03] Speaker 02:
Our first case for argument today is 2018-2308, intellectual ventures versus unified patents.

[00:00:12] Speaker 02:
Mr. Stewart, please proceed.

[00:00:20] Speaker 00:
Thank you very much, Your Honor.

[00:00:24] Speaker 00:
With the court's permission, I would like to start with claim six of the 745 patent.

[00:00:30] Speaker 02:
This is the scanning limitation?

[00:00:31] Speaker 02:
This is the scanning limitation.

[00:00:33] Speaker 00:
Okay, go ahead.

[00:00:33] Speaker 00:
Claim 6 is the dependent claim with the scanning limitation, and it requires scanning from the least recently used file to the most recently used file, or from A to Z. And the board, unfortunately, from our perspective, construed scanning from A to Z to encompass scanning from A to B, or scanning from C to D. And this simply flies in the face of the plain language of the claim.

[00:00:57] Speaker 00:
It also has no support and specification.

[00:00:59] Speaker 01:
Really, the plain language?

[00:01:04] Speaker 01:
I mean, I understand an argument that if you put aside the self-tuning column, that the other mentions of scanning are A to Z, whether Z is included or not.

[00:01:18] Speaker 01:
But the plain language, it's a perfectly normal alternative to say,

[00:01:25] Speaker 01:
search from or something that doesn't require you to continue to the end.

[00:01:31] Speaker 01:
That is, if you find the thing you're looking for along the way, you stop, which is what Robinson is about, right?

[00:01:38] Speaker 00:
Yes, but I would disagree that that's consistent with the plain language of the plain.

[00:01:46] Speaker 00:
From A to Z until you find something,

[00:01:49] Speaker 00:
is a different set of language than just simply from A to Z. What the board essentially did is construe from A to Z to mean from A towards Z, which I think is a very different language.

[00:02:07] Speaker 00:
Judge Toronto mentioned the specification.

[00:02:11] Speaker 00:
I will get to the tuning issue in just one second.

[00:02:15] Speaker 00:
But all of the examples in the specification do show going from the LRU to the MRU without stopping short.

[00:02:23] Speaker 00:
And when the inventor wanted to stop short of the MRU, he knew how to phrase it.

[00:02:27] Speaker 00:
He simply, in Plane 15, for example, states that the scan excludes the MRU.

[00:02:34] Speaker 00:
And the board originally construed claim six to require scanning all the way from the LRU to the MRU inclusive.

[00:02:41] Speaker 00:
The board's reversal, of course, came in light of the tuning discussion in the patent specification.

[00:02:48] Speaker 00:
And I think that the board simply misunderstood what the specification discloses there with respect to the tuning function.

[00:02:55] Speaker 00:
That function relates to files that were frequently used in the past, but have not been used in quite a while, such as segments from a movie that were watched quite a while ago.

[00:03:07] Speaker 00:
While you're watching a movie, the files are being used repeatedly, but once you stop watching the movie, they're not used at all.

[00:03:15] Speaker 00:
These files have the potential to get stuck in the cache for a very long time, and the

[00:03:20] Speaker 00:
deducts from the frequency factors near the least recently used file to ensure that files that were frequently used long ago are eventually evicted from the cache.

[00:03:31] Speaker 00:
The specification language that the board relied on at column seven, line three, determines how close to the LRU a file must be for the deductions from the frequency factors to occur.

[00:03:44] Speaker 00:
It's not talking about how far from the LRU you scan

[00:03:48] Speaker 00:
to determine which files to evict.

[00:03:51] Speaker 00:
It's simply how far you move forward or backward to decide whether or not to start deducting from the LRU or from the frequency factors near the LRU to implement this tuning feature.

[00:04:08] Speaker 00:
Even if the tuning and scanning were the same thing, the plain language of the claims... Can I ask you this question?

[00:04:18] Speaker 01:
microphones are very strange.

[00:04:20] Speaker 01:
I'm going to get away from microphone, otherwise I can't hear myself think.

[00:04:28] Speaker 01:
The claim language, put aside the claim six, just this language about the least recent and least frequent.

[00:04:40] Speaker 01:
I assume you would agree that taking those terms as a matter of plain language, there may not in fact be such a individual file.

[00:04:50] Speaker 00:
That's correct, Your Honor.

[00:04:51] Speaker 01:
So if that's right, then the patent somehow has to provide

[00:04:58] Speaker 01:
in the spec, presumably, some information about how it is that you take these two properties, which need not coincide, and give meaning to them in such a way that it's possible to identify a file that would fall within this oddly put together, double characteristic claim element.

[00:05:24] Speaker 01:
Why doesn't the tuning, the self-tuning column, serve an essential function of giving meaning to language that cannot possibly have its plain meaning?

[00:05:37] Speaker 01:
That is, you have to decide... If you say frequency, frequency means a number of times over a unit of time.

[00:05:46] Speaker 01:
Define the unit of time.

[00:05:47] Speaker 01:
It can be lots of different things.

[00:05:49] Speaker 01:
since the computer first came out of the box, or since yesterday or in the last hour, why isn't that a part of the process of defining the key term of the claim?

[00:06:02] Speaker 00:
I don't think the tuning function relates at all to defining the phrase least frequently used and least recently used.

[00:06:12] Speaker 01:
Well, you want to say, you know, if the movie has not been viewed in the last 24 hours, get rid of it because we don't really think that that's a meaningful, most frequent thing anymore.

[00:06:25] Speaker 01:
Why isn't that one way of defining this odd double characteristic claim term?

[00:06:32] Speaker 00:
I just don't see it as being definitional at all or helpful in understanding what least frequently and least recently used file means.

[00:06:43] Speaker 00:
I mean, there are certainly embodiments disclosed throughout the specification which don't have tuning at all.

[00:06:49] Speaker 00:
It's an option.

[00:06:51] Speaker 00:
And in those embodiments, there would still be a least frequently and least recently used file that needs to be identified and evicted.

[00:07:01] Speaker 00:
Thank you, Your Honor.

[00:07:04] Speaker 00:
Moving from the claim construction, well, I think I was in the middle of saying that even if one viewed the tuning function and the scanning function as the same thing or interrelated, we believe the plain language of the claims requiring that

[00:07:25] Speaker 00:
one scan all the way from A to Z would exclude this interpretation of the tuning embodiment.

[00:07:31] Speaker 01:
What do you do about the board's alternative finding that even under your A all the way to up to or including Z, Burton teaches the element?

[00:07:44] Speaker 00:
That's correct.

[00:07:45] Speaker 00:
That's where I was moving next, Your Honor.

[00:07:49] Speaker 00:
The argument there is that Burton would scan every file if Burton had fewer than 1,024 entries in its cache.

[00:07:57] Speaker 00:
And Burton simply does not disclose such a small cache.

[00:08:00] Speaker 00:
Burton is explicit that its caches will have, quote, likely thousands of cache entries.

[00:08:09] Speaker 00:
That's column three, line four.

[00:08:11] Speaker 00:
And there was no argument presented below or presented in this court that it would have been obvious to modify Burton to use such a small cache.

[00:08:20] Speaker 00:
The argument made by EMC is that only that small caches were known

[00:08:25] Speaker 00:
which we agree with, small caches were known, but EMC makes no effort to explain why it would have been obvious to use a small cache in Burton, and the board's findings did not make any such finding either.

[00:08:37] Speaker 01:
Does Burton require a larger than 1024 entry cache, or just say, indicate that in general, caches are going to be bigger than that?

[00:08:51] Speaker 00:
He states that for his

[00:08:54] Speaker 00:
disclosure for his embodiments that the cashes are going to be thousands of cash entries.

[00:09:02] Speaker 00:
He doesn't specifically come out and state it would be wrong or inappropriate to use a small cash, but the absence of disclosure is not disclosure.

[00:09:21] Speaker 00:
I also think, we also think, Your Honors, that Burton does not even disclose directionally scanning from the LRU toward the MRU.

[00:09:30] Speaker 00:
Burton makes no mention of direction, and the text of Burton must trump the unsupported assertions of EMC's expert.

[00:09:37] Speaker 00:
That is what the board originally found when it initially denied institutional claim six, and the evidence on this point didn't change after institution.

[00:09:47] Speaker 00:
I'd like to move on briefly, I've got about two and a half minutes remaining here, to claims four and five, which include the limitation of frequency factors.

[00:09:58] Speaker 00:
And this limitation is relevant to both the Burton and Coretla grounds.

[00:10:02] Speaker 00:
The claims expressly require frequency factors indicating how often each of the corresponding files are requested by the operating system.

[00:10:12] Speaker 00:
The board construed the frequency factors limitation so broadly that these factors do not indicate how often a file has been requested.

[00:10:19] Speaker 00:
They bear no discernible relationship to the frequency of file requests.

[00:10:25] Speaker 00:
We showed this in the example on page 21 of our opening brief.

[00:10:28] Speaker 00:
And in that example, the so-called frequency factor is based in part on the number of times a file has been accessed so that it's consistent with the board's construction.

[00:10:38] Speaker 00:
But at the same time, the frequency factor is also based upon file size.

[00:10:42] Speaker 00:
You add the file size in bytes or bits or megabytes, whatever you like.

[00:10:47] Speaker 00:
You add the file size to the number of times the file has been accessed to get your resulting number.

[00:10:54] Speaker 00:
That resulting number simply doesn't bear any relationship to the frequency of access.

[00:10:59] Speaker 00:
The frequency number is lost in the overwhelming size of the file size.

[00:11:05] Speaker 00:
So we don't believe that the board's construction can be reconciled with the language of the claims.

[00:11:10] Speaker 00:
The petitioners argue that the specification mandates a broader construction than the plain meaning.

[00:11:16] Speaker 00:
But the petitioners treat plain construction as a game in which every thought or idea that the inventor had in the specification, the claims must be stretched to cover that.

[00:11:26] Speaker 00:
And that's simply not the law.

[00:11:28] Speaker 00:
We've cited cases to that effect in our reply brief.

[00:11:32] Speaker 00:
The petitioner's argument regarding the filed history, we believe, is also meritless.

[00:11:38] Speaker 00:
The petitioners argued that the patentee acquiesced in the examiner's unstated claim construction during the original prosecution, and they rely on the Lemelson versus General Mills case for that.

[00:11:48] Speaker 00:
But this court distinguished Lemelson in both Torr Farm versus Rambaxi and again in Woods versus D'Angelo.

[00:11:55] Speaker 00:
And in these cases, this court explained that Lemelson merely holds that limitations added to overcome a rejection cannot be immaterial to infringement.

[00:12:04] Speaker 00:
These cases leave no room for any argument that a claim amendment constitutes acquiescence in the examiner's claim interpretation, particularly unstated interpretation.

[00:12:15] Speaker 00:
I see that I'm running up on my opening time.

[00:12:19] Speaker 00:
OK.

[00:12:19] Speaker 02:
We'll save the rest for your rebuttal.

[00:12:21] Speaker 00:
All right.

[00:12:21] Speaker 00:
Thank you very much, Chair.

[00:12:22] Speaker 02:
Let's hear from Mr. Fleming.

[00:12:34] Speaker 03:
Good morning, Your Honor.

[00:12:35] Speaker 03:
May it please the court, Mark Fleming together with David Yen on behalf of the Appellees.

[00:12:39] Speaker 03:
I take my cue from Mr. Stewart and begin with scanning.

[00:12:41] Speaker 03:
But of course, I'm happy to address any issues in whatever order the court likes.

[00:12:45] Speaker 03:
I think with respect to scanning, the easiest way to resolve it is per Judge Toronto's question, that the board found as a fact that when the cash is less than 1024 entries, Burton scans all entries from the LRU all the way to the MRU.

[00:12:59] Speaker 03:
There's nothing in Burton constricting itself either in the specification or in Burton's claims to cashes that are greater than 1024.

[00:13:06] Speaker 03:
Mr. Stewart just admitted cashes smaller than 1024 entries were well known.

[00:13:10] Speaker 03:
The main appellate objection in the briefing

[00:13:12] Speaker 03:
from Intellectual Ventures is a procedural one.

[00:13:16] Speaker 03:
We would submit that that challenge is waived.

[00:13:18] Speaker 03:
The reply brief at page 20 concedes that they didn't object to this testimony before the board.

[00:13:24] Speaker 03:
They could have done so in their motion for observations.

[00:13:26] Speaker 03:
They could have done so at the oral hearing.

[00:13:27] Speaker 03:
this court does not reverse the board on grounds that were not put to the board in the first place.

[00:13:31] Speaker 02:
Well, even if we agreed with you on the board's alternative finding that Burton discloses scanning all of the items in the cash, because it doesn't actually specify a number that the cash has to have, and there could be cashes with less than $10.24, even if we agreed with you, is there any necessity in our reviewing the scanning limitation?

[00:13:55] Speaker 02:
And I ask because

[00:13:56] Speaker 02:
There are a number of claims that contain the scanning limitation that were dropped from the IPR.

[00:14:03] Speaker 02:
I don't know if you remember that, which means they're still out there.

[00:14:06] Speaker 02:
And what if it's our view that the narrower scanning limitation proposed by the appellant is in fact the correct one?

[00:14:14] Speaker 02:
Are they asserting these claims in existing lawsuits?

[00:14:19] Speaker 02:
Are you aware of?

[00:14:20] Speaker 03:
Not to my understanding, Judge Moore.

[00:14:23] Speaker 03:
I can check on that and follow up with the court.

[00:14:25] Speaker 03:
But to answer the question, I don't think the court needs to decide more than it should decide for purposes of this appeal.

[00:14:32] Speaker 03:
And for this appeal, deciding the substantial evidence question under the scanning limitation is sufficient to affirm on that limitation as to all claims against which is ministered here.

[00:14:41] Speaker 02:
But you do understand that the construction of scanning, that the same term appears in other claims which were not subject to the IPR.

[00:14:47] Speaker 02:
So they remain valid for this patent.

[00:14:49] Speaker 03:
Well, they haven't been challenged.

[00:14:51] Speaker 03:
There would be no IPR estoppel.

[00:14:52] Speaker 03:
And so if there were something in a future case and it became relevant, then either my clients or someone else could bring an IPR on that.

[00:14:59] Speaker 03:
And perhaps this court would have to address the issue then.

[00:15:01] Speaker 03:
But for purposes of this case, I don't think it's necessary.

[00:15:04] Speaker 01:
That said, if the court would like me to talk about it, I'm happy to.

[00:15:06] Speaker 03:
It might even be litigated in district court.

[00:15:08] Speaker 03:
It might well be, Judge Toronto.

[00:15:10] Speaker 03:
Sure.

[00:15:10] Speaker 03:
But I'm happy to address it if the court wishes.

[00:15:12] Speaker 03:
But I think for purposes of today's exercise,

[00:15:15] Speaker 03:
It is sufficient unto the day simply to affirm on the substantial evidence that the court had in order to find that Burton disclosed this limitation even under intellectual ventures construction.

[00:15:24] Speaker 01:
I think Mr. Stewart made two points about

[00:15:28] Speaker 01:
The insufficiency of Burton, one, the size of the cache and the other, the sequence in which the frequency factors are surveyed there.

[00:15:39] Speaker 01:
Correct.

[00:15:39] Speaker 03:
The direction of the scanning.

[00:15:41] Speaker 03:
That point, we believe, is also waived.

[00:15:43] Speaker 03:
We point that out on page 53 of our red brief.

[00:15:46] Speaker 03:
Intellectual ventures did not dispute the point below.

[00:15:48] Speaker 03:
The only argument that the board heard was that they thought Burton was insufficient because it didn't go all the way to the MRU.

[00:15:54] Speaker 03:
There was no suggestion that the direction was wrong.

[00:15:57] Speaker 03:
They point in their reply brief simply to a couple of demonstratives, but those were not discussed at the oral hearing.

[00:16:03] Speaker 01:
Is there anything you can point to in Burton putting aside the procedural objection that makes clear that it is directional as required?

[00:16:15] Speaker 03:
So the evidence the board relied on is the evidence of our expert, Dr. Kubiatois, who said that a skilled artisan would understand that in doing the scanning in Burton, you would begin at the least recently used unit and go upward from there.

[00:16:29] Speaker 03:
There's nothing more than that because this issue was not joined any further than that.

[00:16:33] Speaker 03:
Had he been asked, he would have said that the reason you do that is you can't identify where unit 1024 is going to be.

[00:16:41] Speaker 03:
It's not as though they're labeled on the side when you just start at 1024 and go down.

[00:16:44] Speaker 03:
You have to start at the bottom, go all the way up.

[00:16:46] Speaker 03:
And the only way you know you've gotten to 1024 is by counting them as you go.

[00:16:49] Speaker 03:
That's not in the record.

[00:16:50] Speaker 03:
He didn't say it.

[00:16:51] Speaker 03:
But that's because this wasn't joint.

[00:16:53] Speaker 03:
This could have been, but it wasn't.

[00:16:55] Speaker 03:
With respect to the least frequently, least recently used, which Mr. Stewart mentioned only briefly, and I will too, I think the same point applies, which is the board found, supported by substantial evidence, that the Lee reference discloses that limitation

[00:17:09] Speaker 03:
Even under the proposed construction that Mr. Stewart advances, which is that the system can't ever evict the most recently used unit, the board clearly relied on the heap and linked list embodiment of Lee, where the most recently used unit winds up in a heap, but the evicted unit only comes from the linked list.

[00:17:29] Speaker 03:
There's plenty of substantial evidence supporting that.

[00:17:31] Speaker 03:
I didn't hear Mr. Stewart challenge it this morning, so I think that's sufficient for affirmance of that ground, which has to do with claims four, five, and six.

[00:17:39] Speaker 03:
With respect to frequency factors, finally, indicates does not have a special technical meaning.

[00:17:45] Speaker 03:
Intellectual ventures has always relied on its ordinary and customary meaning.

[00:17:48] Speaker 03:
Burton itself uses that term repeatedly throughout its specification, saying that its variable, the LRU rank, indicates both the time of last access, which is recency, and the number of accesses, which is frequency.

[00:18:02] Speaker 03:
It says that repeatedly.

[00:18:02] Speaker 03:
We cite it in our brief.

[00:18:04] Speaker 03:
If whatever indicates means in this patent, it must also mean when Burton, speaking to the same skilled artisan, uses the exact same word.

[00:18:13] Speaker 03:
And that's reinforced by Dr. Kubiotowicz's testimony, which the board was entitled to credit and did credit on pages 82 to 83.

[00:18:19] Speaker 03:
So we think there's, again, substantial evidence as to all of those.

[00:18:24] Speaker 03:
The only claim as to which the claim construction issue is even relevant is claim 14.

[00:18:29] Speaker 03:
And that's just because of the curious procedural posture of this, that the board found that Burton disclosed frequency factors even under IV's construction, did not make such a finding as to Coretla, didn't opine on that.

[00:18:43] Speaker 03:
But we think the construction is correct in any event, not least because the same word is used by both Burton and the patent, and also because, as the colloquy with Judge Toronto indicated,

[00:18:53] Speaker 03:
The claims plainly cover embodiments under the heading of preferred embodiments that adjust frequency factors based on criteria other than frequency.

[00:19:00] Speaker 03:
It can't be that the frequency factor always is a raw count of the number of times a file was accessed.

[00:19:06] Speaker 03:
If that were the case, it would be impossible to get rid of the movie because the files would have been used a million times and it would stay in the cache polluting it.

[00:19:13] Speaker 03:
The key point of this patent.

[00:19:16] Speaker 03:
is that frequency factors get decremented over time.

[00:19:19] Speaker 03:
That's how you deal with the aging out of files that are no longer needed.

[00:19:22] Speaker 03:
And at that point, obviously, the frequency factor cannot be a raw count of the number of times the file was accessed.

[00:19:28] Speaker 02:
Your ability to talk without breaths or breaks is really quite remarkable.

[00:19:33] Speaker 03:
It's a family skill, Your Honor.

[00:19:34] Speaker 03:
I hope I'm not going too fast and that I haven't blown past things that are useful to you.

[00:19:38] Speaker 02:
No, but I feel like you've covered everything.

[00:19:40] Speaker 02:
Anything else?

[00:19:40] Speaker 03:
That's it, Your Honor.

[00:19:42] Speaker 03:
Thank you very much.

[00:19:42] Speaker 03:
I appreciate the court's attention.

[00:19:51] Speaker 00:
Yeah, I may be forced to race through as well.

[00:19:54] Speaker 00:
There are several points I would like to raise.

[00:19:57] Speaker 00:
First, I believe counsel's principal argument on Burton is that Burton does not teach away from using a small number of locks in a cache.

[00:20:06] Speaker 00:
But our point is that Burton doesn't teach it either.

[00:20:09] Speaker 00:
And this is obviously a case where the prior art needs to teach what is being asserted in order for there to be an anticipation or obviousness.

[00:20:20] Speaker 00:
Also, with respect to whether it would have been obvious to modify Burton to include a small cache, Burton, we think the proposed modification really would violate the fundamental principles of operation of Burton.

[00:20:36] Speaker 00:
Burton considers only files at the very bottom of the cache to ensure elimination of files that are at the LRU end of the cache while avoiding elimination of the MRU files.

[00:20:46] Speaker 00:
But EMC is effectively proposing to have Burton's algorithm examine files at the MRUN, which is contrary to the teachings of Burton.

[00:20:55] Speaker 00:
I did also want to address the waiver argument that was raised with respect to the directionality of the scanning.

[00:21:03] Speaker 00:
This court's case law holds that an issue is not waived unless it was neither pressed nor passed upon below.

[00:21:09] Speaker 00:
Harris Corp versus Erickson is an example of that.

[00:21:13] Speaker 00:
And here the board passed upon the issue of directionality twice, reaching different decisions each time, first in its institution decision and then in its final written decision.

[00:21:22] Speaker 00:
In addition, as counsel mentioned, we did present the argument in our slides at oral argument below.

[00:21:31] Speaker 00:
We did not have an opportunity to get to that slide.

[00:21:35] Speaker 00:
I don't have an opportunity to get to everything that I want here.

[00:21:38] Speaker 00:
But we do believe that that was a timely raising of the issue, and those slides were submitted as part of the record.

[00:21:43] Speaker 00:
The reason it was timely is that in the opening... What was the citation for then?

[00:21:47] Speaker 00:
For Erickson?

[00:21:48] Speaker 00:
For your slides or your argument?

[00:21:50] Speaker 00:
For the slides.

[00:21:51] Speaker 00:
It's the very last two pages of the appendix, pages 4705

[00:22:03] Speaker 00:
47.06, Your Honor.

[00:22:08] Speaker 00:
And as I was saying, we do believe it was timely to raise that in our oral argument, because in the opening brief, the petitioners conceded that Burton does not scan all the way to the MRU, that only the bottom section was scanned.

[00:22:22] Speaker 00:
So in our response, we argued that this concession was dispositive.

[00:22:25] Speaker 00:
The petitioners then reversed course in reply and argued for the first time that Burton's does scan all the way to the MRU if the cash is small.

[00:22:33] Speaker 00:
So we feel we're entitled to argue in response to that new argument that Burton does not in fact disclose the claim direction

[00:22:40] Speaker 00:
And we would say Chamberlain versus One World, a very recent case from this court in support of that.

[00:22:48] Speaker 00:
Counsel mentioned that had he been asked, the expert would have stated certain things.

[00:22:55] Speaker 00:
That should have been in the petition.

[00:22:58] Speaker 00:
There's no reason why that should have been delayed until now.

[00:23:08] Speaker 00:
If the court will indulge me, I can briefly address the Lee grounds.

[00:23:16] Speaker 00:
The board held that Lee never eliminates the MRU file, but the board was relying solely on petitioners' presentations and excluded us from presenting our evidence, as we noted in the brief.

[00:23:28] Speaker 00:
And so we feel that we are entitled to a remand to have our evidence considered.

[00:23:33] Speaker 02:
All right.

[00:23:33] Speaker 02:
Thank you, counsel.

[00:23:35] Speaker 02:
The case is taken under submission.

[00:23:36] Speaker 00:
Thank you, Your Honor.