[00:00:31] Speaker 01:
Our next case is Glasswall Solutions Limited versus Clear Swift Limited, 2018-1407.

[00:00:40] Speaker 01:
Mr. Carlson.

[00:00:41] Speaker 02:
Thank you, Your Honor.

[00:00:46] Speaker 02:
May it please the Court.

[00:00:47] Speaker 02:
I'd like to reserve five minutes for rebuttal.

[00:00:50] Speaker 02:
We are here because the District Court, after we asserted two patents against Clear Swift, granted a motion under 12 v.

[00:00:57] Speaker 02:
6 declaring that the

[00:00:59] Speaker 02:
claims of both patents were directed patent ineligible subject matter.

[00:01:04] Speaker 02:
And in doing so, the district court erred both substantively and in its procedure.

[00:01:10] Speaker 02:
If we recognize that this is a de novo review, I'd like to turn to the substantive portion, first of all, and recognize that it's clear under this court's precedent, as recently emphasized in the data engine decision,

[00:01:28] Speaker 02:
Claims must be read in the context of the specification.

[00:01:32] Speaker 02:
Claims must be read as a whole with reference to the specification and under the understanding of a person having ordinary skill in the art as of the time of the patent.

[00:01:43] Speaker 02:
These patents date back to the 2005 priority date in the original PCT filing.

[00:01:52] Speaker 02:
This Court has also recently recognized in the ANCORA decision that

[00:01:56] Speaker 02:
Improving security can be a non-abstract computer function improvement if done by a technique that departs from earlier approaches to solve a specific computer problem.

[00:02:09] Speaker 01:
This is simply the reverse of prior ROT processes.

[00:02:14] Speaker 01:
This pulls the safe systems rather than prior ROT, which pulled the unsafe ones.

[00:02:25] Speaker 01:
And it's just transferring data back and forth.

[00:02:29] Speaker 02:
I respectfully disagree, Your Honor, and here's why.

[00:02:32] Speaker 02:
The prior art systems, and the specification is very clear about this, depend upon virus definition files.

[00:02:38] Speaker 02:
You have to consult against this database of known malware in order to be able to exclude the part that you don't want from incoming electronic files.

[00:02:51] Speaker 02:
What the specification also teaches here

[00:02:54] Speaker 02:
is a departure from that.

[00:02:56] Speaker 02:
It's not simply the reverse.

[00:02:58] Speaker 02:
The difference is, rather than trying to determine what virus definitions are and new viruses that come in day by day or hour by hour, we recognize that electronic files correspond to a set of rules, a set of file specifications that are relatively static and fixed.

[00:03:19] Speaker 02:
And this is necessary in order to ensure interoperability of these files.

[00:03:23] Speaker 01:
Rules are ideas, right?

[00:03:26] Speaker 02:
Rules are specifics as to each of these files.

[00:03:33] Speaker 01:
Specific and not, they are abstract ideas.

[00:03:38] Speaker 02:
I would say they are values, Your Honor.

[00:03:39] Speaker 02:
The rules that are the abstract ideas are simply an overarching

[00:03:49] Speaker 02:
I guess approach whereas the specifics in computer communications require the application of specific values, specific rules of acceptable content.

[00:04:07] Speaker 02:
So the difference is this patent specification says we take the approach of for each type of file that is exchanged we look at

[00:04:20] Speaker 02:
what is expected for known content.

[00:04:23] Speaker 02:
And we parse an incoming file and check it for behavior that conforms to accepted known content.

[00:04:34] Speaker 00:
In this contemplated process, does the whole incoming file have to be scanned?

[00:04:40] Speaker 02:
No.

[00:04:40] Speaker 02:
And in fact, if you look at the preamble of the 045 patent, claim one,

[00:04:48] Speaker 02:
It explicitly says without scanning for unwanted code.

[00:04:54] Speaker 00:
I'm sorry.

[00:04:58] Speaker 00:
Without scanning, full stop would mean one thing.

[00:05:01] Speaker 02:
Yes.

[00:05:01] Speaker 00:
Without scanning for unwanted code would mean a different thing.

[00:05:05] Speaker 00:
Basically what's going on here is the old world had a

[00:05:09] Speaker 00:
ever-changing, extremely long list of bad guys, and you don't want them to come in, and you check to see if somebody who's at your door is a bad guy, and that may take a long time, and you may not have an up-to-date list, and you say, people are coming through the door, we just have a guest list, we have good guys.

[00:05:24] Speaker 00:
Maybe by type, maybe by individual, we're just gonna let those guys in.

[00:05:27] Speaker 00:
And that's the sense I think in which Judge Laurie, I understood what he was saying to say, this is kind of just the opposite.

[00:05:34] Speaker 00:
But in your system, do you still have to scan everybody at your door?

[00:05:40] Speaker 02:
No, and what the claims provide is that the incoming file is parsed.

[00:05:45] Speaker 02:
Parsing and scanning are different arrangements, and they are talked about differently in the specification.

[00:05:51] Speaker 02:
Parsing is breaking down the file into components in order to be able to check the content for conformance with known standards.

[00:06:01] Speaker 04:
But that's a commonly used technique.

[00:06:03] Speaker 04:
Is that right?

[00:06:04] Speaker 04:
In other words, in this particular patent, there's nothing

[00:06:09] Speaker 04:
special about the parsing, correct?

[00:06:14] Speaker 02:
Parsing is a known technique.

[00:06:17] Speaker 04:
And all you're doing is basically you're parsing the incoming file and you're simply filtering out things that might be troublesome and then passing along everything else.

[00:06:32] Speaker 04:
It seems to me this is sort of directly parallel to

[00:06:37] Speaker 04:
As you mentioned, that was an issue in intellectual ventures, for example.

[00:06:42] Speaker 02:
The difference is, Your Honor, once the parsing is complete and the known file components of acceptable file content are known, there is a substitute regenerated electronic file.

[00:06:56] Speaker 02:
And so the specification and the claims specify

[00:06:59] Speaker 02:
that now we are regenerating a new type of file and the new file is all that is passed.

[00:07:05] Speaker 04:
The original... But the new file is nothing more than sort of a sterilized version of the original file.

[00:07:13] Speaker 04:
There's nothing different, nothing new.

[00:07:16] Speaker 04:
It's just you pulled out the bad things or the potentially bad things according to the rules and simply passed everything else along.

[00:07:29] Speaker 04:
electronic filter.

[00:07:32] Speaker 02:
I respectfully disagree again.

[00:07:34] Speaker 02:
The specification teaches that what we pull out are expected conforming portions of the file, expected content that conforms to the file type specifications.

[00:07:49] Speaker 02:
The difference in the prior art approach is, rather than having to scan for something that's bad, we pull out, we parse and pull out what is

[00:07:59] Speaker 02:
expected the type of behaviors that conform to this file type.

[00:08:05] Speaker 02:
And there is a new file that's regenerated.

[00:08:09] Speaker 02:
It's not simply a lighting or restricting or blacking out part of the original file.

[00:08:16] Speaker 02:
We create a substitute regenerated electronic file.

[00:08:20] Speaker 02:
And the advantage to that is the file regenerator cannot pass any unwanted data.

[00:08:28] Speaker 02:
So the advantage that comes, and the reason why this is patent eligible, is the advantage that comes out of this is the prior approach and the conventional approach at the time was not parsing in the way that the patent describes.

[00:08:41] Speaker 02:
It was take an incoming file and scan it, code scanning, to determine whether there are signatures or pieces of the file that correspond to this virus definition file.

[00:08:54] Speaker 02:
And of course, if the virus definition file is not constantly updated,

[00:08:57] Speaker 02:
then you don't find that.

[00:09:01] Speaker 04:
So I would say this is sort of the electronic version of a sensor reviewing mail and pulling out offensive photographs and then passing the letters along.

[00:09:14] Speaker 04:
Because the letters are unchanged.

[00:09:16] Speaker 04:
It's the same thing here.

[00:09:18] Speaker 04:
The components that are not potentially dangerous are passed along.

[00:09:27] Speaker 04:
You're not in this regenerated file.

[00:09:30] Speaker 04:
There's nothing.

[00:09:32] Speaker 04:
That's not a newly created document that differs in its essential character from the original document, correct?

[00:09:40] Speaker 02:
I'd say that's incorrect, Your Honor, and the specification makes that clear.

[00:09:44] Speaker 04:
When the district court... How is that incorrect?

[00:09:48] Speaker 02:
The district court made sort of the same arrangement, saying, well, isn't this just like redacting documents?

[00:09:54] Speaker 02:
Well, when you redact documents, you're essentially

[00:09:57] Speaker 02:
using a virus definition file in your head.

[00:10:00] Speaker 02:
You're saying, I know the things that I don't want, and I will block them out.

[00:10:04] Speaker 02:
This is a different approach.

[00:10:06] Speaker 02:
That was a conventional approach.

[00:10:08] Speaker 02:
But what is claimed in these patents is a fundamentally different approach as the patent specification provides.

[00:10:16] Speaker 02:
The file is parsed.

[00:10:18] Speaker 02:
The parsed content is checked to see if it conforms to expected standards.

[00:10:25] Speaker 02:
And then there is indeed a new file that's written.

[00:10:28] Speaker 02:
It is a fundamentally new file.

[00:10:30] Speaker 02:
And the advantage that comes out of parsing and regeneration is a very high level of security that doesn't depend on knowing that something's bad.

[00:10:40] Speaker 02:
It depends on have we checked to see if the file content conforms to the expected file specifications after we have parsed it.

[00:10:54] Speaker 02:
So it's not the same thing.

[00:10:57] Speaker 02:
It's not redacting of documents because that requires knowing what's bad.

[00:11:03] Speaker 02:
We take the approach that it is not efficient, it is not the right approach to keep updating virus files to know what's bad.

[00:11:17] Speaker 02:
Instead, go to the fundamentals of the file type specification to know

[00:11:24] Speaker 02:
what files are expected, parse the file, take the content that conforms to those expected portions, and regenerate into a new file.

[00:11:36] Speaker 02:
So it's different from the intellectual ventures.

[00:11:38] Speaker 02:
I would say it's much more close to the Fingen file type, because in Fingen versus Bluecoat, this court recognized that when we, and Fingen actually did do

[00:11:52] Speaker 02:
a behavioral-based scan to avoid conventional code scanning and then wrote a new file that attached information regarding potentially suspicious behaviors.

[00:12:14] Speaker 04:
that was not present before and that file was then attached to the downloadable and gave the computer the opportunity to take action it couldn't take before.

[00:12:26] Speaker 04:
That's a completely different arrangement.

[00:12:29] Speaker 02:
I would say it's exactly the same arrangement, Your Honor, because the substitute regenerated electronic file of these patents is a newly created file.

[00:12:39] Speaker 04:
It's a file that didn't exist before, and it enables the computer to use... It didn't exist before because before it had content that you extracted.

[00:12:47] Speaker 04:
But other than the extracted component, it's the same, correct?

[00:12:55] Speaker 02:
No, that's not correct, Your Honor.

[00:12:56] Speaker 02:
What is extracted is data that conforms to the expected standard.

[00:13:02] Speaker 02:
And then there is a file regeneration process, and the regenerator

[00:13:06] Speaker 02:
takes only the accepted content, the only the conforming content, and generates the new file.

[00:13:12] Speaker 02:
So the original file is not passed unless it has been pre-approved.

[00:13:16] Speaker 02:
The original file doesn't get passed at all.

[00:13:19] Speaker 04:
The only thing that gets... Are you saying the content of the regenerated file is different from the content that was contained in the original file?

[00:13:34] Speaker 04:
And I'm not talking about

[00:13:36] Speaker 04:
the part that was parsed out.

[00:13:40] Speaker 04:
Well, let me rephrase it this way.

[00:13:42] Speaker 04:
The regenerated file is regenerated from what?

[00:13:47] Speaker 02:
The regenerated file is regenerated as the claims and specification teach from content that is found to conform with the expected standards.

[00:13:56] Speaker 04:
And that content comes from the original file.

[00:14:00] Speaker 02:
The content

[00:14:02] Speaker 02:
in the original file that conforms to the standards, is regenerated.

[00:14:07] Speaker 02:
The content that it does not conform is not regenerated.

[00:14:11] Speaker 04:
Well, you've got to use the word repeated, correct?

[00:14:15] Speaker 02:
Well, the firm, the word in the patent is regenerated, Your Honor.

[00:14:18] Speaker 04:
Yeah, but I'm trying to get at what we're talking about here in terms of the actual data.

[00:14:24] Speaker 04:
Yeah, this, the data... It's repeated data that comes from, that was contained in the original file.

[00:14:30] Speaker 02:
It's conforming data.

[00:14:31] Speaker 02:
So if the original file had all data that conformed, it's still not the same file.

[00:14:37] Speaker 02:
It is regenerated into a new file.

[00:14:40] Speaker 02:
And so even if there's something sneaky in there, as long as it's within the scope of, you know, the only stuff that's written by the regenerator is stuff that conforms in the original file.

[00:14:55] Speaker 01:
Councilor, you wanted to save some rebuttal time.

[00:14:58] Speaker 01:
We'll give you two minutes for rebuttal.

[00:15:00] Speaker 01:
And we'll hear from opposing counsel.

[00:15:04] Speaker 02:
If I may, Your Honor, I believe I'd ask for five minutes.

[00:15:06] Speaker 02:
So do I have five minutes?

[00:15:08] Speaker 01:
Well, you've talked through it now.

[00:15:09] Speaker 02:
You have two minutes.

[00:15:10] Speaker 02:
I beg your pardon.

[00:15:11] Speaker 01:
Thank you.

[00:15:13] Speaker 01:
Mr. Al-Salam.

[00:15:14] Speaker 03:
Thank you, Your Honor.

[00:15:15] Speaker 03:
It's Ramsey Al-Salam of Perkins Couey, along with my colleague, Lane Polizola.

[00:15:21] Speaker 03:
These patent claims are clearly directed to an abstract concept, the abstract concept of

[00:15:28] Speaker 03:
filtering for what is accepted, what's allowable or conforming versus the opposite or the converse abstract concept, which is filtering for what's not acceptable.

[00:15:40] Speaker 03:
As the court mentioned, it's like determining what is good versus what is bad.

[00:15:45] Speaker 03:
Counsel didn't like the example of the sensor because it says in the sensor example, you are filtering for something that you know is bad.

[00:15:54] Speaker 03:
Well, perhaps this example works better.

[00:15:57] Speaker 03:
It's like

[00:15:58] Speaker 03:
me telling my assistant, I only want to see bills and personal letters.

[00:16:04] Speaker 03:
I don't want to see any other type of mail.

[00:16:07] Speaker 03:
She says, well, what if a bill has advertising in it?

[00:16:10] Speaker 03:
Remove the advertising.

[00:16:11] Speaker 03:
I only want to see bills and personal letters, unless it comes from the court.

[00:16:15] Speaker 03:
If it comes from the court, I want to see it no matter what it is.

[00:16:18] Speaker 03:
That's the other exception in the claims.

[00:16:20] Speaker 03:
But this goes to the abstract concept

[00:16:24] Speaker 03:
of simply deciding what's allowable versus what's not allowable.

[00:16:29] Speaker 03:
Even when this court makes rules, this court can make those rules based on what we can do or what we cannot do.

[00:16:36] Speaker 00:
Did the prior art, which looked for bad stuff, contemplate one or both of two possibilities of what you do when you find bad stuff in a file?

[00:16:52] Speaker 00:
One possibility is reject the whole file.

[00:16:54] Speaker 00:
The other is remove the bad stuff and then take the good.

[00:16:58] Speaker 03:
What I can say based on what the record we have, it says for... There's a background section in the patent.

[00:17:03] Speaker 00:
Does the background section itself, for example, say... I know it says the prior art looks for bad stuff.

[00:17:11] Speaker 00:
Does it say, and the prior art

[00:17:14] Speaker 00:
covers both things you can do with once you've identified the bad stuff, namely, delete it and take the good or just say anything that has bad stuff, we reject the whole thing.

[00:17:25] Speaker 03:
I think it says the former.

[00:17:28] Speaker 03:
If you look at the 283 patent, Appendix 21, at column 3, lines 17 to 18, it says,

[00:17:38] Speaker 03:
Quote, known anti-virus programs aim to detect viruses and pass everything which is not detected to be a virus.

[00:17:46] Speaker 03:
So in other words, it's saying we pull out the bad stuff and we pass everything else.

[00:17:52] Speaker 03:
And what this is saying is we pull out the good stuff and pass only it.

[00:17:57] Speaker 03:
We don't pass anything which we have not identified as something good.

[00:18:01] Speaker 00:
Has anything in this case made of this notion of regenerating a file as itself an advance?

[00:18:13] Speaker 03:
No.

[00:18:14] Speaker 00:
So is regeneration here nothing other than big, I guess, the good stuff that was originally in the thing that came in concatenated or something with

[00:18:29] Speaker 03:
I agree, there's nothing in the specification, to my knowledge, that suggests that the regeneration of the file is in any way novel or non-obvious or anything other than simply a filtered version of what came first, as Your Honor mentioned.

[00:18:45] Speaker 03:
For example, in the Fingin case, the file that was created is a completely new file.

[00:18:52] Speaker 03:
First of all, those claims are limited to executable applications.

[00:18:56] Speaker 03:
And what it does there is it analyzes based on behavior what those applications do and creates a completely new file called the security profile that it attaches.

[00:19:07] Speaker 03:
And it does in itself.

[00:19:08] Speaker 03:
The claims don't there even refer to filtering.

[00:19:11] Speaker 03:
They send it on and allow each recipient to decide based on their own security profile whether to filter those applications.

[00:19:20] Speaker 03:
Here, this regeneration of a file is simply the same data with

[00:19:26] Speaker 03:
removing anything that hasn't been identified as conforming or allowable or good.

[00:19:32] Speaker 03:
This is very much, like all the cases, the courts, I understand, have struggled with what's an abstract concept.

[00:19:40] Speaker 03:
And there's probably a spectrum of what are abstract concepts.

[00:19:45] Speaker 03:
But this is one of the purest abstract concepts, because it is simply taking what sometimes referred to as whitelisting versus blacklisting and applying it

[00:19:56] Speaker 03:
to a known task for a computer, which is to filter data.

[00:20:01] Speaker 03:
And it's not improving the underlying functioning of the computer, unlike Infish, say, or maybe Berkheimer, which improved the way memory was stored in a computer, regardless, independent of the task being performed by the computer.

[00:20:19] Speaker 03:
This is not doing that.

[00:20:20] Speaker 03:
This is supposedly a better way to perform a specific task.

[00:20:26] Speaker 03:
that of filtering content.

[00:20:28] Speaker 03:
But the advantages of this, the alleged advantages are all just alleged advantages over the other abstract concept, meaning blacklisting or filtering for unwanted content.

[00:20:42] Speaker 03:
What we hear is, and this is the only alleged advantages, are that this other way of doing it requires storing all of these virus signatures.

[00:20:52] Speaker 03:
and that that can be a lot, and you might not be up to date.

[00:20:55] Speaker 03:
So it's better to look for good content than to look for bad content.

[00:21:00] Speaker 03:
But what we're doing here is just comparing two abstract approaches to filtering content.

[00:21:07] Speaker 03:
In fact, the telltale sign might be that you don't even need to have skill in the art to have written these claims.

[00:21:14] Speaker 03:
I'm not a computer scientist, but even I

[00:21:16] Speaker 03:
could theorize that if you're trying to filter, you could either look for bad content and remove it, or look for good content and only allow it to pass.

[00:21:28] Speaker 03:
The court also properly decided this under Rule 12.

[00:21:35] Speaker 03:
There is no claim construction arguments made.

[00:21:39] Speaker 03:
There was no suggestion that this turned on claim construction.

[00:21:42] Speaker 03:
There was no issues of fact as to

[00:21:46] Speaker 03:
that would suggest that this might not be something other than an abstract concept or it might be limited somehow.

[00:21:53] Speaker 03:
There were additional elements that were not routine or conventional.

[00:21:57] Speaker 03:
All of the additional elements, whether in the independent claims or the dependent claims, simply recite known steps or known computer technology.

[00:22:07] Speaker 03:
There is a, there was nothing that

[00:22:09] Speaker 03:
would have required the court to put this off.

[00:22:13] Speaker 03:
There was nothing in the record.

[00:22:14] Speaker 03:
Again, the only advantages are advantages that are contrasted with the other approach.

[00:22:22] Speaker 03:
And they're just two basic approaches to filtering.

[00:22:25] Speaker 03:
Either, and I know I'm repeating myself, either removing bad content or allowing good content.

[00:22:31] Speaker 03:
Both are just opposite sides of the same coin.

[00:22:41] Speaker 03:
Well, what else can I address for the court?

[00:22:44] Speaker 03:
Does the court have any other questions?

[00:22:46] Speaker 03:
I want to use my time, but I feel we've addressed some of the major things.

[00:22:50] Speaker 01:
No one ever gets penalized for not using all their time if you have nothing further to say.

[00:22:56] Speaker 01:
Unless they've forgotten something very important.

[00:22:58] Speaker 03:
Yeah, I'm trying to think if I've forgotten anything very important.

[00:23:03] Speaker 03:
I do think this follows the Symantec case, and I also want to go back to

[00:23:08] Speaker 03:
When this court's trying to decide what is an abstract concept, the court has said that taking data and then just comparing it to other data and making a decision on that is an abstract concept.

[00:23:22] Speaker 03:
And perhaps if the rules relating to the comparison are so specific or unique or somehow novel, then maybe that wouldn't be an abstract concept.

[00:23:32] Speaker 03:
But here, all we're doing is taking data that's coming in, comparing it

[00:23:37] Speaker 03:
to an existing database to see if it matches allowable content or matches a known sender, a known approved sender, and then sending it on.

[00:23:48] Speaker 03:
And that in itself is simply claiming an abstract concept to filtering data no different than if I were asking my assistant to filter the mail that comes to me.

[00:24:00] Speaker 03:
Thank you.

[00:24:00] Speaker 01:
Thank you, counsel.

[00:24:02] Speaker 01:
Mr. Carlson has two minutes.

[00:24:06] Speaker 02:
Thank you.

[00:24:07] Speaker 02:
There is an improvement in computer function disclosed and claimed in this patent.

[00:24:11] Speaker 02:
And the improvement is in the file regeneration process, as described in column three, beginning about line 30.

[00:24:20] Speaker 02:
It says this is a fundamentally different way.

[00:24:22] Speaker 02:
And what's the difference?

[00:24:23] Speaker 02:
And what's the improvement in computer function?

[00:24:26] Speaker 02:
Instead of having to update virus definition files, we look for content that is static.

[00:24:33] Speaker 02:
The file specifications

[00:24:36] Speaker 02:
rules are static.

[00:24:39] Speaker 02:
They don't change often for these different file types.

[00:24:42] Speaker 02:
And so that way, because we can do that, we don't have to depend on knowing what is bad and looking for what is bad.

[00:24:49] Speaker 04:
The rules have to have some correlation to things that present a risk, correct?

[00:24:59] Speaker 02:
The rules have correlation to what is expected, commonly expected,

[00:25:03] Speaker 04:
Right.

[00:25:04] Speaker 04:
And if the incoming data doesn't correspond to what is commonly expected, that raises a red flag.

[00:25:13] Speaker 02:
It may or may not.

[00:25:14] Speaker 04:
You don't have to... A different version of, you know, on the one hand, you're looking for something you know is a virus.

[00:25:20] Speaker 04:
On the other hand, you're looking for something that doesn't conform to expected rules.

[00:25:27] Speaker 04:
Correct, but... In terms of patent eligibility, I don't see that there's a difference.

[00:25:32] Speaker 02:
The difference is that it does improve computer function because you don't have to have these bloated virus files, and you don't have to update regularly.

[00:25:40] Speaker 04:
You don't have a bloated file full of rules.

[00:25:42] Speaker 02:
Well, it's a static file.

[00:25:45] Speaker 02:
It's a static set of rules because these don't change often.

[00:25:49] Speaker 02:
They do change, though.

[00:25:51] Speaker 02:
They do change.

[00:25:51] Speaker 02:
They do require some update, and the patent specifies that, but not with the lightning speed that viruses can be... That may be true, but I'm not sure that makes a difference.

[00:26:02] Speaker 04:
in terms of eligibility.

[00:26:05] Speaker 02:
With respect to our rules ideas, the McCrow case I think is illustrative where the patent claims that were approved there referred to rules that are in the specification.

[00:26:16] Speaker 02:
The same thing is happening here.

[00:26:17] Speaker 02:
We have rules for expected content, we define those rules, and they remain static.

[00:26:25] Speaker 02:
We parse the file, we regenerate a substitute electronic file,

[00:26:30] Speaker 02:
That's a new file that didn't exist before that is known to be safe.

[00:26:35] Speaker 02:
And that is the improvement in computer function.

[00:26:37] Speaker 01:
Thank you, counsel.

[00:26:38] Speaker 01:
We'll take the case under advisement.