[00:00:06] Speaker 01:
Our final case for argument today is 2018-1599, Personal Web Technologies versus Apple.

[00:00:46] Speaker 01:
Mr. Hadley, please proceed.

[00:00:50] Speaker 04:
Thank you, Your Honor, and may it please the Court.

[00:00:53] Speaker 04:
In the prior decision in this case, this Court remanded this case back to the Board and specifically directed it to do two things.

[00:01:04] Speaker 04:
First, as to whether the prior art reference, Woodhill, compares a content-based identifier against a plurality.

[00:01:15] Speaker 04:
the arguments had centered on the restore procedure in Woodhill and specifically Column 17.

[00:01:22] Speaker 04:
And this Court said, go back, analyze Column 17, and provide an evidence-based explanation as to how Woodhill shows this comparison.

[00:01:34] Speaker 04:
Secondly, this Court said, in terms of motivation and reasonable expectation of success, said to provide a clear, evidence-supported account of the contemplated

[00:01:45] Speaker 04:
Woodhill-Stefik combination, how it would work, and why that combination would support a conclusion that a skilled artisan would have had an apparent reason to make the combination and reasonably expected it to work.

[00:02:01] Speaker 01:
We agree with you that Woodhill does not inherently disclose how backup files

[00:02:09] Speaker 01:
If we agree with you that Woodhill doesn't inherently disclose that, we don't have to reach the other points on motivation to combine, correct?

[00:02:17] Speaker 01:
Because the office relied on Woodhill's disclosure to provide the element.

[00:02:21] Speaker 04:
Correct, Your Honor.

[00:02:21] Speaker 01:
All right.

[00:02:22] Speaker 04:
So turning to that specific point, we know now that we were correct when we said that Woodhill does not disclose a comparison of a content-based identifier against the plurality of values.

[00:02:38] Speaker 04:
In fact, in the remand decision, the board expressly said, and Apple admits, that Woodhill, quote, doesn't mention a comparison.

[00:02:47] Speaker 04:
Instead, as Your Honor mentioned, the board relied on an inherency theory in the remand decision.

[00:02:55] Speaker 04:
But the board misapplied the law of inherency, both misapplied the law and the facts in reaching the inherency decision.

[00:03:04] Speaker 04:
First, in terms of inherency, the law of this

[00:03:08] Speaker 04:
court is clear.

[00:03:09] Speaker 01:
We definitely know the law.

[00:03:11] Speaker 01:
So just get to your precise factual point.

[00:03:15] Speaker 01:
Isn't your point that it's not clear?

[00:03:17] Speaker 01:
Woodhill doesn't explain how it locates the backup files relevant to a received update request, and it's possible, you have to admit it's possible these backup files could be located by comparing the binary object identifier and the update request to a plurality of binary object identifiers in the backup server.

[00:03:33] Speaker 01:
It's possible, but

[00:03:35] Speaker 01:
It's not necessary to do so, because the update request also includes a link to the file identification record, which contains file location.

[00:03:42] Speaker 01:
So while it's possible they do this, it's not required by any means, because there's this alternative way the very thing could be done based on the exact disclosure in Woodhoe.

[00:03:51] Speaker 01:
Do I have your argument?

[00:03:53] Speaker 04:
You have my argument, Your Honor.

[00:03:54] Speaker 04:
And we know that it's possible, because that is the invention of the 310 patent, the True Names patent.

[00:04:01] Speaker 04:
There was no invention of this prior to the true name patent.

[00:04:06] Speaker 04:
And if this was something that could be inherent, you would think Woodhill having invented this brand new thing would at least say something about it.

[00:04:15] Speaker 04:
It didn't.

[00:04:15] Speaker 02:
Can you remind me again what is the board's view of why this particular limitation is met?

[00:04:24] Speaker 02:
Is its view that it's zeroing in on the disclosed

[00:04:29] Speaker 02:
the discussed binary object identification record in column 17, and then you extract the binary object identifier from that record, and then you use that binary object identifier to look through some database of binary object identifiers to find a one-to-one correspondence?

[00:04:52] Speaker 04:
Your Honor, I can't answer that question because the board never explained it.

[00:04:56] Speaker 02:
Well, I need you to try in order to figure out whether or not

[00:05:00] Speaker 02:
it, you know, is a reasonable reading of the reference.

[00:05:03] Speaker 02:
It's not a reasonable reading of the reference, and the only thing that the board... But is that your understanding of the board's understanding of column 17?

[00:05:12] Speaker 02:
The board only in addressing column 17... Or is it using the binary object identifier extracted from the binary object identification record in order to locate a file as opposed to a specific binary object?

[00:05:27] Speaker 04:
Well, we know it can't be the latter.

[00:05:29] Speaker 04:
because a binary object identifier can only identify a binary object.

[00:05:35] Speaker 04:
There is no disclosure.

[00:05:36] Speaker 02:
I might not disagree with you, but I'm just trying to understand what the board's understanding.

[00:05:41] Speaker 04:
There's no disclosure in Woodhill of a file identifier, a content-based identifier for a file.

[00:05:47] Speaker 04:
So we know that that isn't part of Woodhill.

[00:05:49] Speaker 04:
The only evidence that the board cited was paragraph 72 of the Goldberg Declaration.

[00:05:56] Speaker 04:
That was Apple's expert.

[00:05:58] Speaker 04:
And what the board pointed to and what Dr. Goldberg said was only this.

[00:06:03] Speaker 04:
And I quote, quote, the remote backup file server maintains some sort of file system or other mapping, i.e.

[00:06:11] Speaker 04:
database, that allows the binary object identification record to serve as a lookup for the requisite file data.

[00:06:19] Speaker 04:
There is nothing in Dr. Goldberg's declaration about how that description of taking a binary object

[00:06:29] Speaker 04:
identification record and somehow using that in a database or file server, either A, expressly requires the comparison of a binary object identifier against the plurality of values to find a file, or B, why that's inherent.

[00:06:49] Speaker 04:
In fact, Dr. Goldberg never provided any testimony whatsoever regarding the issue of inherency.

[00:06:55] Speaker 04:
The issue of inherency never came up in this case,

[00:06:58] Speaker 04:
either in the decision to institute, in the final written decision, it didn't come up until the board's remand decision.

[00:07:05] Speaker 04:
And before the remand, the board never said, look, we think we're going to go on an inherency theory, so we want you to brief inherency.

[00:07:13] Speaker 04:
That was something that the board came up with for the very first time in the decision on remand.

[00:07:18] Speaker 02:
I guess the board is trying to use some logical thought process and trying to understand exactly how do you

[00:07:28] Speaker 02:
access all these granules, and you first get to the granules by first getting to the binary objects, and how do you get to those binary objects?

[00:07:36] Speaker 02:
Well, you must be using the binary object identifiers.

[00:07:39] Speaker 02:
And that's what is contained, or that's what the binary object identification record contains.

[00:07:47] Speaker 02:
And so now you can go hunt down the corresponding binary object.

[00:07:58] Speaker 04:
Apple's argument and the Board's decision make express that they think this inherent disclosure happened, or this inherent process of comparing against the plurality happens at the very beginning of the restore process, where Woodhill at step 442, and this is shown in the appendix at, let's see, figure 5I of Woodhill,

[00:08:27] Speaker 04:
Step 442 simply says, obtain identity of the file.

[00:08:32] Speaker 04:
And that's where they say this comparison of a binary object identifier against a plurality has to occur.

[00:08:37] Speaker 04:
Well, there's two problems with that.

[00:08:39] Speaker 04:
First, we know that binary objects are simply small pieces of a file, one megabyte pieces.

[00:08:45] Speaker 04:
And there could be many, many binary objects within a given file.

[00:08:51] Speaker 04:
There's no explanation of how you could take one little piece of a file

[00:08:55] Speaker 04:
and a content-based identifier for that piece of a file, and use that to identify the entire file.

[00:09:02] Speaker 04:
There's just no explanation anywhere, and I can't think of a way that you could do that.

[00:09:07] Speaker 04:
But even more important, if you go back to figure 5i of Woodhill, and you look at the next step after 442, step 443 says compile a list of binary objects in the previous version of the file.

[00:09:23] Speaker 04:
So if you haven't even compiled the list of binary objects until after you've obtained the identity of the file and you've located the file, how could it be that comparing a binary object identifier to a set of binary objects that haven't even been located yet or compiled could somehow let you find the file?

[00:09:46] Speaker 04:
is just it's completely illogical.

[00:09:49] Speaker 04:
It turns the flow chart of figure 5i of Woodhill on its head.

[00:09:54] Speaker 04:
There's simply no explanation of how that would happen.

[00:09:57] Speaker 04:
And therefore, under the law, it can't possibly be inherent in Woodhill.

[00:10:03] Speaker 04:
In fact, Woodhill really teaches away from using the binary object identifiers.

[00:10:09] Speaker 04:
And one more point is that Woodhill earlier actually talks about how you find a file.

[00:10:16] Speaker 04:
regardless of what you want to do with it.

[00:10:18] Speaker 04:
There's a section of Woodhill that talks about compressing files.

[00:10:22] Speaker 04:
And it says the first thing you have to do is find the file before you compress it.

[00:10:27] Speaker 04:
And at column 11, line 57 through 12, line 5, and that's at appendix 1679, it talks about using the file name and the file location to find the file and then to look at the binary object offsets

[00:10:45] Speaker 04:
to find where each binary object is and then compress it.

[00:10:48] Speaker 04:
Remember in Woodhill, the reason you have this binary object offset is that the binary objects are a fixed size, except for perhaps the last one for whatever is left over.

[00:11:00] Speaker 04:
But they're exactly one megabyte long, starting at byte zero.

[00:11:04] Speaker 04:
So if you know where the file is, by the file name, the location, everything conventional,

[00:11:12] Speaker 04:
you can find where each binary object is in that file by simple counting, starting with zero and going in one megabyte segments.

[00:11:20] Speaker 04:
There's absolutely no reason why you would need to use a content-based identifier to find any binary object within a file in Woodhill.

[00:11:28] Speaker 04:
It's like an address inside of the file?

[00:11:31] Speaker 04:
Essentially, the binary object identifier doesn't serve as an address, because all the binary object identi... No, no, but I'm talking about the reliance on the offset.

[00:11:41] Speaker 04:
Correct.

[00:11:41] Speaker 04:
The lines on the offset effectively act like an address.

[00:11:45] Speaker 04:
So if you say, find binary object three, you go in three megabytes.

[00:11:50] Speaker 04:
And from megabyte three to megabyte four, that's exactly where that binary object is.

[00:11:55] Speaker 04:
And so again, in column 11, when you're talking about compressing a file, Woodhill expressly says, here's how you find a file.

[00:12:04] Speaker 04:
And it has nothing to do with comparing binary objects with anything.

[00:12:08] Speaker 04:
In fact, Woodhill,

[00:12:10] Speaker 04:
uses, of course, the binary objects in one-to-one comparisons.

[00:12:15] Speaker 04:
I think this was all agreed during the last hearing before this board, and even by the board, to save bandwidth, to determine what has changed.

[00:12:23] Speaker 04:
So you're looking at binary objects individually and comparing one-to-one with the binary objects for the corresponding version on the user's computer.

[00:12:34] Speaker 04:
And whether you're backing up or you're restoring,

[00:12:38] Speaker 04:
The whole idea is to make that one-to-one comparison and only use the bandwidth if there's been a change.

[00:12:45] Speaker 04:
There's simply no disclosure whatsoever in Woodhill of using a binary object identifier for this broader process that frankly wasn't invented until the TrueNames patent in the 310 patent and its predecessors, of using a content-based identifier for this broader purpose.

[00:13:04] Speaker 04:
of determining whether a particular piece of data, whether it's defined by a fixed size or not, to locate that anywhere in a network or a system.

[00:13:16] Speaker 04:
And that's, again, another reason why it can't possibly be inherent in Woodhill to use this content-based identifier to find the file in the first place.

[00:13:31] Speaker 04:
and certainly not before you even compiled your list of binary object identifiers as shown in the next step in Figure 5i of Woodhill.

[00:13:38] Speaker 01:
You're into your rebuttal time.

[00:13:39] Speaker 01:
Would you like to save some?

[00:13:41] Speaker 04:
Not unless there is any questions regarding the motivation to combine.

[00:13:47] Speaker 04:
Thank you.

[00:13:54] Speaker 03:
May it please the court.

[00:13:56] Speaker 03:
The key distinction here between what Mr. Hadley is talking about

[00:13:59] Speaker 03:
and what we're talking about with the restore procedure is that what Mr. Hadley is talking about with respect to, say, the compress routine, for instance, that's all taking place on the local workstation, whereas what we're talking about here in terms of the restore procedure is taking place on the backup file server.

[00:14:17] Speaker 03:
It's clear for the very first time.

[00:14:19] Speaker 02:
So just to match up with the claim language, what's the data item in column 17 of Woodhill

[00:14:28] Speaker 02:
that is being compared to a plurality of values?

[00:14:31] Speaker 03:
It's the binary object.

[00:14:33] Speaker 02:
So I think column 17... The binary object identifier?

[00:14:36] Speaker 03:
Correct.

[00:14:36] Speaker 03:
The content-dependent identifier, content-dependent name of the claims in the 310 patent is the... Okay.

[00:14:42] Speaker 02:
And then what are the plurality of values in column 17 that is being... that the binary object identifier is being compared against?

[00:14:51] Speaker 03:
Sure.

[00:14:52] Speaker 03:
It's a plurality, sort of a database on the backup file server.

[00:14:56] Speaker 03:
of all the binary objects, and it's very clear in column 17.

[00:15:02] Speaker 02:
Of all the binary object identifiers?

[00:15:04] Speaker 03:
Correct.

[00:15:05] Speaker 02:
In a specific file?

[00:15:07] Speaker 03:
No, just the binary object identifiers on the backup file server altogether.

[00:15:12] Speaker 02:
So there could be like 10 million binary object identifiers among, I don't know, a million or 500,000 files?

[00:15:22] Speaker 03:
There could be, and there's no indication anywhere in Woodhill

[00:15:25] Speaker 03:
that there's any sort of segregation by.

[00:15:27] Speaker 02:
It's very interesting.

[00:15:27] Speaker 02:
There's a database of binary object identifiers necessarily part of Woodhill that contains the binary object identifier for every binary object in every single file stored on the backup server.

[00:15:47] Speaker 03:
Correct.

[00:15:47] Speaker 03:
And that's the only evidence that there is before the board.

[00:15:50] Speaker 03:
The evidence that.

[00:15:52] Speaker 02:
Why would a system be created like that?

[00:15:55] Speaker 02:
I guess I don't understand.

[00:15:56] Speaker 02:
I mean, the backup server is going to have file names, right?

[00:16:02] Speaker 02:
So you can locate every single file.

[00:16:05] Speaker 03:
There is no indication whatsoever in Woodhill that the file names or locations are on the backup file server.

[00:16:11] Speaker 03:
Mr. Hadley, our personal web for the very first time in this appeal points to in the binary object identification record that gets sent over as part of the request.

[00:16:20] Speaker 03:
Now personal web for the very first time points to this link to the backup instance record.

[00:16:25] Speaker 03:
And from there, they point to, in the backup instance record, there's a link to the file, which has file name 40 and location 38.

[00:16:33] Speaker 03:
This is the very first time that that argument has been made.

[00:16:38] Speaker 03:
Back before the board, in the original filings that PersonalWeb made, it had argued sort of a different theory of how the file was located.

[00:16:46] Speaker 03:
But this pointing to the backup instance record, you don't see that in the patent owner's response.

[00:16:51] Speaker 03:
Dr. Dewar's declaration, Dr. Dewar was PersonalWeb's expert.

[00:16:54] Speaker 03:
You didn't see it in the motion for rehearing.

[00:16:55] Speaker 03:
And you didn't see it in their brief on remand.

[00:17:00] Speaker 03:
So they take issue with the board sort of not going through various alternatives and shooting them down.

[00:17:05] Speaker 03:
Well, this alternative was not before the board.

[00:17:07] Speaker 02:
I guess I'm just trying to understand why would there be such a database?

[00:17:11] Speaker 02:
No such database is mentioned in the Woodhill Patent.

[00:17:14] Speaker 02:
It talks about a file database.

[00:17:17] Speaker 02:
It talks about a backup queue database.

[00:17:20] Speaker 02:
And now you're talking about this very interesting other database.

[00:17:24] Speaker 03:
Even if there were some database, even if the database on the backup file server did segregate things by files, it still would perform this comparison because every file, unless it's just one meg, every file is gonna have more than one binary object.

[00:17:40] Speaker 03:
So the Woodhill is very clear in column 17 that the request that goes over, the update request that goes to the backup file server contains the binary object identification record

[00:17:51] Speaker 03:
for each of the binary objects of the previous version of the file that needs to be restored.

[00:17:59] Speaker 03:
So even if there is some segregation by user or file on the backup file server, you still have to locate the binary object to see whether or not the binary object is actually on the backup file server.

[00:18:11] Speaker 03:
So it's necessarily the case that there would be a comparison between the binary object identifier and

[00:18:19] Speaker 03:
you know, multiple or plurality, more than one binary object identifier.

[00:18:22] Speaker 02:
So I guess you're saying for a given file, maybe it's a 40 megabyte file, so therefore there's 40 binary objects in there.

[00:18:30] Speaker 03:
Correct.

[00:18:30] Speaker 02:
And then you have a binary object identification record, you extract the binary object identifier, and then you do what?

[00:18:38] Speaker 02:
You somehow compare that identifier against the 40 binary object identifiers in the previous version of the file?

[00:18:47] Speaker 03:
that is correct, to see whether or not that binary object identifier exists on the backup file server and where it's located so that the next step in the procedure, the granule restore procedure, can commence.

[00:18:58] Speaker 02:
But why does it need to be done that way, I guess, is what I'm trying to figure out.

[00:19:02] Speaker 02:
There's the offset field that kind of gives you an understanding of the location of all the given binary objects inside of the file.

[00:19:12] Speaker 03:
Well, I think going back to the substantial evidence standard,

[00:19:15] Speaker 03:
This is the board considered the evidence.

[00:19:16] Speaker 03:
The board looked at Woodhill.

[00:19:17] Speaker 00:
But it's a substantial evidence of necessity.

[00:19:21] Speaker 00:
And I thought Mr. Hadley described a scenario in which both on the backup server and the thing that you're on the local, you know that the thing you're looking for is in place number 17 in the file.

[00:19:40] Speaker 00:
And as long as you know that on the backup server,

[00:19:44] Speaker 00:
You've got, in Judge Chen's example, the 40 separate binary objects.

[00:19:52] Speaker 00:
You're looking at number 17, and all you need to do is compare this number 17 to this number 17.

[00:19:58] Speaker 00:
One-to-one comparison, not a plurality.

[00:20:01] Speaker 00:
As long as that's possible, then there's no inherence.

[00:20:04] Speaker 00:
I gather something like that is his theory.

[00:20:07] Speaker 03:
Correct, but there is nothing, there's no evidence whatsoever that on the backup file server,

[00:20:13] Speaker 03:
In order to do that, you would need sort of file name, location, things of that sort to kind of find where you're even going to start to do this lookup of the binary object identifier.

[00:20:23] Speaker 03:
And they haven't pointed to anything on the backup file server that shows that there's file name and file location.

[00:20:28] Speaker 03:
In fact, the binary object identification record 58 that gets sent is part of this update request, does not contain the file name and the file location.

[00:20:37] Speaker 02:
At line 30 of column 17, it says the manager program 24 obtains from the user

[00:20:44] Speaker 02:
the identities of the current and previous versions of the file, which needs to be restored.

[00:20:50] Speaker 02:
So it sounds like, you know, there's no need to go hunting around for what are the identities of the files.

[00:20:58] Speaker 02:
The user is providing that to the program.

[00:21:00] Speaker 03:
But the identities referenced there cannot be a conventional file name, as Mr. Hadley suggests, because the current version of the file and the previous versions of the file are all going to have the same conventional file name.

[00:21:12] Speaker 03:
If, say you have five versions of a, say a Word document, a legal brief, you have five versions and you have on the user's computer version five and you have maybe version four also, the previous version, because Whittle says you have the current version, one previous version, everything else is stored on the backup file server.

[00:21:31] Speaker 03:
So say you want to go and you want to restore to the current version to, you know, previous version two that's on the backup file server, you can't use the conventional name because you send back the conventional name

[00:21:42] Speaker 03:
to the backup file server, that's gonna give you versions one, two, and three.

[00:21:46] Speaker 03:
You're not gonna know what version you need to go to.

[00:21:49] Speaker 03:
And again, there's nothing indicating that the update request includes a file name or a location.

[00:21:57] Speaker 03:
That's not what's sent.

[00:21:58] Speaker 03:
What's sent over is the binary object identification record.

[00:22:01] Speaker 03:
Within the binary object identification record, there is no file name, there is no file location.

[00:22:11] Speaker 03:
You can sort of reach that through various links, but even there, as column three makes clear, the file name and location are the file name and location on the user's workstation.

[00:22:22] Speaker 03:
There's nothing indicating that any of that information is on the back of file server.

[00:22:26] Speaker 03:
I think column three, towards the bottom, is very clear on that.

[00:22:29] Speaker 02:
Are you saying that you're using the binary object identifier not only to find the prior version of the binary object,

[00:22:39] Speaker 02:
but also the file itself?

[00:22:43] Speaker 03:
You need the binary object identifier to locate the binary object on the backup file server.

[00:22:53] Speaker 02:
Right, but then is that the way you get access to the file on the backup server?

[00:22:58] Speaker 03:
Well, I don't think there's anything where you ever get access to a file.

[00:23:01] Speaker 02:
I'm trying to figure out chicken or egg first.

[00:23:03] Speaker 02:
Do you find the file first on the backup server, or do you find

[00:23:08] Speaker 02:
a specific binary object of a particular file on the backup server and then you kind of work your way to getting access to the backup file.

[00:23:16] Speaker 03:
I think it could work either way and I think under either instance, there would still be a comparison to a plurality of values.

[00:23:33] Speaker 03:
But again, there's nothing that the user's workstation

[00:23:38] Speaker 03:
is sending to the binary, sorry, it's sending to the user's workstation, it would in any way include file location, file name, anything of that sort.

[00:23:50] Speaker 03:
I think column three of Woodhill is very clear that the file name 40 and the file location 38 are all about what's on the user's workstation.

[00:24:01] Speaker 03:
And again, for the reasons I mentioned earlier, you can't just send a file name because if you're looking for, say,

[00:24:06] Speaker 03:
version one, version two, version three, and so on, are all gonna have the same file name.

[00:24:12] Speaker 03:
So using file name doesn't really get you anywhere.

[00:24:16] Speaker 02:
They're all gonna have the same name?

[00:24:17] Speaker 02:
It's not gonna, I mean, when I save different versions of a file, I have to name one, name two, name three, or something like that to differentiate between the different versions of a given document that I'm working on.

[00:24:30] Speaker 03:
Well, again, if we look to column three of Woodhill, column three, when it's talking about

[00:24:35] Speaker 03:
The file name, file location 38 and file name 40, that's all part of the file identification record.

[00:24:43] Speaker 03:
We see starting on line 64 of column three of Woodhill, it says for each file identification record, 34 and file database 25, one or more backup instance records, 42, are created that contain the information about the file.

[00:24:59] Speaker 03:
So it's clear that file location 38 and file name 40

[00:25:05] Speaker 03:
are all about the current version of the file that's on the user's workstation.

[00:25:11] Speaker 03:
You can create numerous backups, and you're gonna have, say, you know, whatever, five backup instance records, 50 backup instance records, however many versions you have, you're gonna have different backup instance records, but you're only gonna have one file identification record, and within that, one file name and one file location, and that's all referring to the current version of the file, not the previous versions.

[00:25:33] Speaker 03:
Again, looking at the evidence that was before the board.

[00:25:36] Speaker 03:
I mean, the board did consider Woodhill.

[00:25:38] Speaker 03:
They considered the testimony of Dr. Goldberg.

[00:25:41] Speaker 03:
The theories that Personal Web is putting forth now were not something that was presented to the board in Dr. Dewey's declaration.

[00:25:48] Speaker 01:
Yes, but the problem with that is you can waive waiver.

[00:25:50] Speaker 01:
And I just sat here and paged through your red brief, and the word waiver doesn't appear in it.

[00:25:54] Speaker 01:
So you didn't make your argument in the red brief that they have waived this argument that they presented to us about the link

[00:26:02] Speaker 01:
namely in Figure 3 that you see.

[00:26:04] Speaker 01:
So you're standing up here today and making it, but that doesn't help you.

[00:26:07] Speaker 03:
So I'm not making an explicit waiver argument.

[00:26:09] Speaker 01:
What I'm saying is they're trying to fault the board for not considering it, for not going through sort of... Well, they're faulting the board for not considering what's in the disclosure of Woodhill, which includes Figure 3.

[00:26:19] Speaker 01:
And you're saying, well, they didn't make that argument to the board below.

[00:26:22] Speaker 01:
And I'm saying, well, they made it in their blue brief, and they made it in their gray brief, and you disputed it, and you're red.

[00:26:27] Speaker 01:
But what you didn't argue is you didn't make that argument to the board below.

[00:26:30] Speaker 01:
You waived the ability to complain that they didn't make this argument to the board below, because you had notice of it in blue, and you made the waiver argument for the first time orally.

[00:26:40] Speaker 03:
What I'm saying is not I'm not making a waiver argument.

[00:26:43] Speaker 03:
What I'm saying is in addressing whether.

[00:26:46] Speaker 01:
You used the word waiver earlier.

[00:26:47] Speaker 01:
So I don't.

[00:26:49] Speaker 01:
Maybe you're not making a waiver argument now.

[00:26:51] Speaker 01:
But you definitely did earlier in the oral argument.

[00:26:53] Speaker 03:
Perhaps I misspoke.

[00:26:55] Speaker 03:
What I was intending to say was that when considering what

[00:27:00] Speaker 03:
What personal web argues is that the board didn't go through the proper analysis, that it didn't say these magic words of this possibility doesn't work and that possibility doesn't work and that possibility doesn't work.

[00:27:10] Speaker 03:
And I haven't seen any authority that requires the board to go through every single potential possibility and explain why it wouldn't work.

[00:27:17] Speaker 01:
What it did say, it made the necessary... Well, I understood them to say the file identification record could be the way in which these things are located.

[00:27:25] Speaker 01:
I understood you in the red brief to say, well, identification record, that's not even submitted or sent.

[00:27:33] Speaker 01:
And then they came back in gray brief and said, actually, it's attached right here.

[00:27:36] Speaker 01:
It's demonstrated in the patent.

[00:27:37] Speaker 01:
It's attached.

[00:27:39] Speaker 01:
So that's the nature of the argument, as I understood it.

[00:27:42] Speaker 03:
I could be wrong.

[00:27:43] Speaker 03:
So it's not that it's actually attached.

[00:27:45] Speaker 03:
What it is is a link in the database.

[00:27:48] Speaker 01:
But you said it wasn't present.

[00:27:49] Speaker 01:
And they said, sure, it is.

[00:27:51] Speaker 01:
It's the link.

[00:27:52] Speaker 01:
It's the link is sent, which contains the very information.

[00:27:55] Speaker 03:
the link is not, so the link does not, in a database structure, a link being sent doesn't contain the information.

[00:28:01] Speaker 03:
And in any event, the binary object identification record contains the link to the backup instance record.

[00:28:07] Speaker 03:
That does not contain the file name and the file location.

[00:28:09] Speaker 03:
Within the backup instance record on the user's workstation, there's actually a link to the file identification record, which there contains file name,

[00:28:19] Speaker 03:
40 and file location 38.

[00:28:21] Speaker 03:
That information never goes to the user works, I'd be sorry, to the backup file server.

[00:28:26] Speaker 03:
So it's not something that the backup file server ever gets.

[00:28:28] Speaker 03:
So I think in our brief, that is what we were arguing.

[00:28:32] Speaker 03:
Perhaps it wasn't entirely clear, but the file identification record is not something that the backup file server ever has access to, and there's no indication it does.

[00:28:43] Speaker 03:
It's purely a link to the backup instance record, which then contains a link, but in any event,

[00:28:48] Speaker 03:
the file name and the file location that are in the file instance record are information about the file, the current version of the file, on the user's workstation.

[00:28:57] Speaker 03:
That is not information about one of the backup previous versions on the backup file server.

[00:29:06] Speaker 02:
On the motivation to combine, Stefik is talking about accessing or approving the access of a file, and here in Woodhill we're talking about

[00:29:19] Speaker 02:
I guess the approval of access to a binary object or a binary object identifier?

[00:29:25] Speaker 02:
Sure.

[00:29:26] Speaker 02:
And to me that seems like a peculiar way to, you know, whitelist or blacklist access to something on a backup server.

[00:29:35] Speaker 02:
I would think you would be, as in Stefik, trying to approve or disapprove access to a given file and not to a given little data item inside of the file.

[00:29:48] Speaker 03:
You could have a list, I mean, all we're talking about in terms of the combination is basically taking the existing system of Woodhill, and I think there are a few places in the board's decision on remand where they basically say, you know, Woodhill gets you just about all the way there, but you just need to add a sort of small authorization layer of Stefic.

[00:30:08] Speaker 03:
And all you would need to do is add a field, say, in the database that's on the backup file server that checks, you know, does user A

[00:30:18] Speaker 03:
have access to these various binary objects.

[00:30:22] Speaker 03:
It would be a pretty simple thing to implement in a database structure.

[00:30:26] Speaker 02:
To approve the access to given object inside of a file as opposed to approval to access to the file as a whole.

[00:30:38] Speaker 03:
Correct.

[00:30:39] Speaker 03:
Because you could have it, the binary object is just, as we've spoken about, just splitting up the file into various pieces.

[00:30:46] Speaker 03:
So you could have a field

[00:30:48] Speaker 03:
in the database that you had brought up an instance where say it's a 40 meg file so you have 40 binary objects and you can have in the database on the backup file server a field that says okay for these 40 different binary objects, yes, user A has access or maybe whatever user A has access to is not listed there so therefore user A does not have access and then is not authorized

[00:31:16] Speaker 03:
to access that file.

[00:31:17] Speaker 03:
So it would just be a simple modification of the existing system.

[00:31:20] Speaker 03:
You could certainly do it on a binary object basis because you could have all the binary objects for a particular file when they're backed up.

[00:31:27] Speaker 03:
You could give permissions or not permissions or those things can change over time.

[00:31:33] Speaker 01:
Okay, let's have Mr. Hadley and let's hear from him on his rebuttal.

[00:31:40] Speaker 01:
Mr. Jay went over, so you have two minutes.

[00:31:43] Speaker 01:
And if you need a little bit of extra time, we'll be generous and certainly give it to you.

[00:31:48] Speaker 04:
I want to go back to Woodhill column 17.

[00:31:52] Speaker 04:
It expressly says at lines 28 to 32, the distributed storage manager program obtains from the user the identities of the current and previous versions of the file, not the binary objects of the file, which needs to be restored.

[00:32:08] Speaker 04:
We go to column to figure 5i at step 442, obtain identity of the file.

[00:32:15] Speaker 04:
And then we go to figure 3, and there is expressly disclosed something called a file identification record.

[00:32:23] Speaker 04:
The problem where the board and Apple fail to satisfy the law of inherency is that nothing in their expert's declaration there is absolutely

[00:32:36] Speaker 04:
no evidence whatsoever and no explanation in the board's decision about how you would find a file using simply a binary object identifier for one little piece of the file, particularly when the patent says obtain from the user the file and you have something called a file identification record that has the file name and the file location.

[00:33:02] Speaker 04:
I mean, there is no suggestion anywhere that you can't find files conventionally.

[00:33:08] Speaker 02:
That's the way you found files since computers... But I'm hearing the other side saying you use it to find a binary object, to restore a given binary object in a file, not to locate the entire file.

[00:33:22] Speaker 04:
The argument that the board made, or what the board found, was that you use the binary object identification record to find

[00:33:29] Speaker 04:
and the binary object identifiers, the content-based identifiers, to find the file by comparing the binary object identifier against a plurality.

[00:33:40] Speaker 04:
There is nothing in the record whatsoever and nothing in the declaration to support the idea that that's how you find a file.

[00:33:49] Speaker 04:
And everything in Woodhill says just the opposite, that you have all the information that you need to find a file conventionally.

[00:33:58] Speaker 04:
And if that's the case,

[00:33:59] Speaker 04:
then you can't possibly have an inherent disclosure of something that the board and Apple admit Woodhill doesn't mention, and that's a comparison of a content-based identifier against a plurality.

[00:34:13] Speaker 04:
And for that reason, we've been through this once, and we would respectfully ask that this not be remanded again back to the board.

[00:34:23] Speaker 04:
We've already been through nearly four years of IPR

[00:34:27] Speaker 04:
proceedings on this particular patent, but we would instead respectfully ask that the court render the matter as being valid and find that Apple did not sustain its burden of proof.

[00:34:42] Speaker 01:
Okay, I thank both counsel for their argument.

[00:34:44] Speaker 01:
The case is taken under submission.