Weather Underground, Incorporated v. Navigation Catalyst Systems, Incorporated et al

Filing 263

RESPONSE to 245 MOTION in Limine to exclude testeimony and expert reports of Defendants' Expert Richard Korf filed by Connexus Corporation, Firstlook, Incorporated, Navigation Catalyst Systems, Incorporated. (Delgado, William)

IN THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF MICHIGAN

THE WEATHER UNDERGROUND, INC.,
a Michigan corporation,
Plaintiff,
vs.

Case No. 2:09-CV-10756
Hon. Marianne O. Battani

NAVIGATION CATALYST SYSTEMS, INC.,
a Delaware corporation; CONNEXUS CORP.,
a Delaware corporation; FIRSTLOOK, INC.,
a Delaware corporation; and EPIC MEDIA
GROUP, INC., a Delaware corporation,
Defendants.
______________________________________________________________________
Enrico Schaefer (P43506)
Brian A. Hall (P70865)
TRAVERSE LEGAL, PLC
810 Cottageview Drive, Unit G-20
Traverse City, MI 49686
231-932-0411
[e-mail]
[e-mail]
Lead Attorneys for Plaintiff

William A. Delgado
WILLENKEN WILSON LOH & LIEB LLP
707 Wilshire Boulevard, Suite 3850
Los Angeles, CA 90017
[phone]
[e-mail]
Lead Counsel for Defendants

Nicholas J. Stasevich (P41896)
Benjamin K. Steffans (P69712)
Anthony P. Patti (P43729)
BUTZEL LONG, P.C.
HOOPER HATHAWAY, PC
150 West Jefferson, Suite 100
126 South Main Street
Detroit, MI 48226
Ann Arbor, MI 48104
[phone]
734-662-4426
[e-mail]
[e-mail]
[e-mail]
Attorneys for Plaintiff
Local Counsel for Defendants
______________________________________________________________________
CONNEXUS, FIRSTLOOK, AND NCS’S EVIDENTIARY OBJECTIONS AND
RESPONSE TO PLAINTIFF’S MOTION IN LIMINE REGARDING RICHARD KORF

MEMORANDUM OF POINTS AND AUTHORITIES
I.

INTRODUCTION.
Plaintiff’s motion in limine to preclude the testimony of Dr. Rich Korf is nothing more

than an eleventh hour attempt to exclude relevant, admissible evidence. Plaintiff has had well
over a year to challenge Dr. Korf’s testimony and certainly could have challenged it when it was
submitted into evidence in July-August 2011 as part of Defendants’ Motion for Partial Summary
Judgment. See Docket No. 196 (Declaration of Richard Korf). It did not do so, choosing instead
to bring this motion on the eve of the pretrial conference, in an effort to deprive Defendants of a
full hearing (with comprehensive briefing) on the issue.
In addition to its procedural deficiencies, Plaintiff’s motion is substantively deficient. In
support of its motion, Plaintiff predominantly argues that Dr. Korf should be prohibited from
stating opinions he has not made and does not intend to make. On the other hand, the opinions
Dr. Korf has provided would greatly assist the jury in this case. Dr. Korf, an expert in the field
of computer science and, specifically, artificial intelligence, can provide the jury with an
understanding of how Defendants’ automated registration system works. He can further provide
information as to the limitations of computer intelligence which will assist the jury in
understanding some of the challenges that Defendants faced in designing their system and why
some amount of human review is necessary.
All of Dr. Korf’s testimony is admissible and helpful; none of it is unfairly prejudicial.
For the reasons stated herein, Plaintiff’s motion in limine should be denied.
//
//

1
121061.1

II.

STATEMENT OF FACTS.
As this Court is aware, for years, Defendants engaged in the business practice of bulk

domain name registration. That practice necessarily involved, inter alia,: (i) inputting a large
amount of DNS error data; (ii) creating domain names from that data; (iii) testing those domain
names for profitability; and (iv) screening those domain names for similarity to trademarks. To
accomplish these tasks, Defendants utilized a system comprised of both software and human
elements. It is that system that is on trial.
Dr. Rich Korf is a computer science professor at UCLA. See Curriculum Vitae of
Richard Korf , Docket No. 196, Ex. B. One of his areas of research is artificial intelligence (or,
put differently, how computers can be built to think like human beings). Dr. Korf’s testimony in
this case is not only helpful to the jurors; it is critical. Dr. Korf can explain to the jurors what
Defendants’ software is designed to do. Or, put differently, did defendants write a piece of
software that affirmatively goes out and captures domain names that are similar to trademarks?
Or, to the contrary, did defendants write software that tries to screen out domain names that are
similar to trademarks? In addition, Dr. Korf—with his knowledge of artificial intelligence
systems—can help explain just how difficult it is (indeed, perhaps impossible at present) to write
software when the task at hand requires some contextual analysis as is the case with domain
names.
III.

ARGUMENT.
A.

Plaintiff’s Motion Is Nothing Short of an Eleventh Hour “Hail Mary.”

To borrow a term from the world of football, Plaintiff’s motion in limine is nothing short
of a last second “Hail Mary” pass. As Plaintiff readily concedes, Dr. Korf issued his first report

2
121061.1

in September 2010 and a Supplemental Report in November 2010. He was deposed on
December 8, 2010, at which time he discovered a few typographical errors in his report and
issued a corrected report on December 10, 2010. Plaintiff has had ample time since December
2010 to bring a Daubert challenge which could have been fully briefed and fully considered by
the Court. Nevertheless, at no point thereafter did Plaintiff challenge Dr. Korf’s qualifications or
opinions.
In fact, Dr, Korf’s opinions are already in the evidentiary record. Defendants relied on
Dr. Korf’s Declaration in Support of Its Partial Motion for Summary Judgment back in JulyAugust 2011. See Docket No. 196. Plaintiff brought no formal challenge (based in Daubert or
otherwise) at that point in time.
Instead, Plaintiff filed a motion on February 24, 2012 that is scheduled to be heard on
February 28, 2012, leaving Defendants with merely the intervening weekend to prepare this
response to the motion (along with responding to Plaintiff’s other motion in limine regarding Mr.
Berryhill and preparing a motion for protective order to quash an improperly issued subpoena)1.
Plaintiff’s tardy filing precludes any comprehensive briefing by Defendants (though, obviously,
this document represents Defendants’ best attempt with limited time), and it precludes full
consideration by the Court. That reason alone calls for the denial of the motion. Guy v. Ford
Motor Co., 1995 WL 92353 (E.D. La. 1995) (denying motion in limine to exclude expert and
noting “the motion was filed on the eve of trial precluding the Court’s ability to hold a full
Daubert hearing. It should be noted that this is not the first setting for this case and AlliedSignal
1

To make matters worse (and as Plaintiff was informed), the weekend of February 25-26 is
particularly problematic for Defendants’ counsel as his entire firm’s computer infrastructure
has been offline since the evening of February 24th for a major upgrade, leaving him without
access to his online case files and e-mails.
3

121061.1

had more than ample opportunity to file this motion in a more timely manner.”).
B.

Dr. Korf’s Opinions on the Design of Defendants’ System Are Admissible.

Plaintiff’s first argument—after making the necessary concession that an expert’s opinion
is admissible even if it embraces an ultimate issue to be decided by the jury—is that Dr. Korf has
provided “legal opinions” on defendants’ intent which should be excluded. Mot. at 2 citing Korf
Depo. Tr. 128: 3-4 (“The question was: What was the intent? What was it designed to do?”).
But, that line from Dr. Korf’s deposition only highlights that Plaintiff’s argument is
untrue. Dr. Korf was not asked to opine as to Defendant’s subjective intent. He was asked to
opine as to the design/intent of the automated registration system. The testimony, in full, is:
So, I wasn’t asked, you know, whether this system had ever registered—I
wasn’t asked to provide an opinion about whether the system had ever
registered a trade—domain name which may have been confusingly
similar. The question was: What was the intent? What was it designed to
do?
Korf Depo. Tr. 127:23-128:4 (Emphasis added). Obviously, that is why he posits the question
as: “What was it designed to do?” Dr. Korf was clearly referring to the system not to
Defendants. Defendants are not “designed.” A software system is “designed.”
Here, the design of the software system is of critical importance. As noted, above,
Defendant’s system accepted DNS error as input and used that error data to create domain
names. After creating these domain names, the system tested for matches in the USPTO
trademark database. But, the question necessarily arises: what would happen when there was a
match? Did the system automatically register the name (which would evidence bad faith)? Or,
did it flag the name for potential deletion (which would not evidence bad faith)? Or was it
something in between? Obviously, that difference in the software’s algorithm is crucial, and Dr.

4
121061.1

Korf’s testimony as to what the system was designed to do would be helpful to the jury. 2
Two of Plaintiff’s other complaints merit further response but are easily disposed. First,
Plaintiff complains that “Korf concludes that Defendants’ system is designed to avoid
cybersquatting, yet admits that he has not compared Defendants’ system to any other system.”
Mot. at 1, n. 1. That argument is a non-sequitur. Dr. Korf does not need to see anyone else’s
system to determine what Defendant’s system is designed to do. He can simply analyze
Defendants’ system in light of his years of experience, education, teaching and research.
Moreover and practically speaking, Plaintiff’s requirement of “comparison” creates an
impossible hurdle. Assuming other such systems exist, they are undoubtedly the proprietary
technology of their designers, including Defendants’ competitors. It is unlikely these
competitors would simply allow Dr. Korf access to their trade secrets and then permit him to
testify as to salient differences and similarities in open court.
Lastly, Plaintiff complains (several times) that “Korf…has no opinions regarding
anything beyond mere design, rather than actual performance…” Mot. at 11. That is true: Dr.
Korf’s focus is on design rather than performance. As it should be. As Defendants have
repeatedly stressed and as this Court previously noted in ruling on the cross motions for partial
summary judgment, the test under the ACPA is “intent” not “willful blindness,” or “gross
negligence,” or mere “negligence,” or “strict liability.” To examine software’s intent, as Dr.
Korf testified in his deposition, you have to look at the design, not the efficacy. In accordance
with the intent of its designers, Microsoft Windows was designed to work correctly all of the
2

Given that Mr. Korf is opining on the design of software (which can be objectively
ascertained) and not Defendants’ intent, Plaintiff’s citation to H.C. Smith Investments, L.L.C
v. Outboard Marine Corp., 181 F. Supp. 2d 746, 749 (W.D. Mich. 2002) (stating that only
jury can determine the intent of a party) is inapposite.
5

121061.1

time. Yet, anyone who has ever used a Windows-based computer has had to reboot it at some
point because Windows did not work as intended.
To bring the analogy one level closer to that of the world of law: it is Plaintiff’s intent to
preclude Mr. Korf’s testimony. Its lawyers have designed and drafted a motion in limine to carry
out that intent. But, if that motion is denied (as it should be), the motion’s failure says nothing
about the motion’s intent and design.
C.

Dr. Korf Is Not Providing Any Improper Testimony on the Element of
“Confusingly Similarity.”

In another attempt to strike Dr. Korf’s useful testimony, Plaintiff claims that Dr. Korf
intends to give testimony on whether the domain names in this case are confusingly similar to
Plaintiff’s trademarks and that Dr. Korf is attempting to educate the jury on this element of the
ACPA. Neither is true.
It is true that Dr. Korf refers to that language from the ACPA to explain an inherent
programming difficulty in the automation of string matching. But, one would surmise that it is
preferable to use the language from the ACPA than, say, another statute wholly unrelated to the
case at hand.
The issue, of course, is this: computers can do a great job of telling you how closely the
characters in a domain name match the characters in a trademark (i.e., “similarity”) but cannot do
a very good job at all in telling you whether that similarity is “confusing.” Take, as an example,
Plaintiff’s alleged common law mark WUND and its website <wund.com>. A computer can tell
you that <wind.com> is very similar to WUND because only one letter is replaced. However, it
cannot tell you whether the two character strings are confusingly similar. And, of course, they

6
121061.1

are not. “Wind” is a generic word for that force of air that comes “sweepin’ down the plains,”
and Plaintiff has no ACPA claim against the registrant of <wind.com>. On the other hand, a
computer would likely tell you that <wundweather.com> was not very similar to <wund.com>
because of the additional seven letters that are appended to the end of “wund”. Yet, Plaintiff
would almost certainly file suit against the registrant of such a domain name claiming “confusing
similarity.”
Dr. Korf’s testimony in this regard is helpful. It can explain the limitations of computer
intelligence and serves as a validation for Defendants’ decision to leave the final determination
on registration to human reviewers instead of creating a process that was entirely automated.
D.

Dr. Korf Is Not Providing an Opinion on “Semantics.” He Simply Used the
Word.

Plaintiff also takes issue with the fact that Dr. Korf uses the word “semantics” (which,
according to the dictionary simply means “the study of the meaning of words”) in explaining the
limitations of a computer’s ability to determine “confusing similarity.” Plaintiff protests that Dr.
Korf is not an English major or English professor. Mot. at 8. But, Dr. Korf is not providing an
opinion on “semantics” or even delving into the study of the meaning of words itself. Rather, he
is simply using the word to explain why a computer can determine whether a domain name and a
trademark are similar but not whether they are “confusingly similar.”
The example given by Dr. Korf in his Supplemental Report is particularly apropos. A
computer would tell you that the characters in <nikeshoes.com> and the characters in
<niceshoes.com> are nearly identical; their “edit distance” differs by only a single character.
But, a computer could not make a subjective judgment as to whether the similarity that existed in

7
121061.1

this example is “confusing” or not to a human being.3 A person can. Dr. Korf’s report speaks
for itself:
Ultimately, judgments about what is confusingly similar are complex
judgments based on the meanings of words and the contexts in which they
are used. While a computer can compare strings of characters such as
“niceshoes” and “nikeshoes,” it has no idea what these character strings
mean, and hence cannot make completely reliable judgments about what
would be confusingly similar to a person.
Korf Supplemental Report at ¶ 3. Clearly, Dr. Korf’s testimony is rooted in computer science
and artificial intelligence, topics which he has studied, teaches about, and publishes about. He is
imminently capable of discussing the limitations of artificial intelligence and such a discussion
would be helpful to the jury which should be allowed to consider the difficulties that Defendants
encountered in building their automated system.
E.

Dr. Korf’s Testimony About Defendant’s Present System Is Admissible.

Plaintiff also (half-heartedly) complains that Dr. Korf’s report only addresses the system
that Defendants had in place at the time but does not address previous iterations. That
presupposes that Defendants’ system was radically different earlier in time. It wasn’t, and
Plaintiff presents no evidence that it was. Defendants have always captured DNS error data for
purposes of constructing domain names. Defendants’ efforts to implement a trademark filter
commenced in 2004. Defendants have always had human review. While it is true that aspects of
the system have gone better (e.g., the trademark filtering became more sophisticated, additional
reviewers were added), that only speaks to the efficacy of the system, not the design.
Indeed, the improvements in the system and how the system worked at the proverbial
3

In fact, the domain name <nikeshoes.com> belongs to the well-known shoe manufacturer,
Nike, and <niceshoes.com> belongs to a media company in New York City (that does not
sell shoes at all apparently). Korf Supplemental Report at ¶ 3.
8

121061.1

“end of the day” (i.e, when Korf examined it) are critically important to a potential issue:
statutory damages. As Plaintiff argued over and over again in its first motion in limine, the
ACPA provides for statutory damages to provide for “deterrence,” if necessary. Assuming the
jury was inclined to award statutory damages, in order for them to make a determination as to
whether Defendants need to be “deterred,” they must have evidence of Defendants’
improvements over time and the end-state of the system. With that evidence in hand, a jury may
very well determine that Defendants already have been working, in good faith, to avoid
entanglements with trademark owners and that there is no need for “deterrence.” Without it, the
jury would be making a determination absent the requisite evidence.
Put simply, Dr. Korf’s examination of the current system does not render his testimony
inadmissible. To the extent Plaintiff wishes to cross-examine him about the time limitation of
his knowledge, it is free to do so. Rees v. Target Corp., 2008 WL 7440009 (E.D. Mich. March
31, 2008) (“Vigorous cross-examination, presentation of contrary evidence, and careful
instruction on the burden of proof are the traditional and appropriate means of attacking shaky
but admissible evidence.”) citing Daubert v. Merrill Dow Pharmaceuticals, Inc., 502 U.S. 579,
596 (1993).
F.

Plaintiff’s Fed. R. 403 Argument Lacks Merit.

Plaintiff’s last argument, based on Fed. R. Evid. 403, lacks any merit whatsoever. As
demonstrated, above, Dr. Korf’s testimony is highly relevant. To exclude it under Rule 403,
therefore, the Court would have to find that this relevance was substantially outweighed by the
danger of unfair prejudice, confusion, delay, etc. Plaintiff cannot carry this burden. There is no
“unfair” prejudice at all, much less substantial unfair prejudice. And, given that Dr. Korf’s

9
121061.1

testimony will significantly aid the jury in understanding Defendants’ systems and the challenges
that Defendants faced in building that system, his testimony could not possibly be deemed
dilatory or wasteful.
To support its Rule 403 argument, Plaintiff provides an automobile analogy that makes
no sense because, again, Plaintiff is focused on efficacy as opposed to intent. The design of
Defendants’ system did not materially change over time (i.e., the car always had brakes,
irrespective of whether they were anti-lock or not), and it is design that evidences intent.
Ironically, Plaintiff’s motion betrays its faulty premise when it argues that “[g]iven that Korf, as
noted above, has no opinions regarding anything beyond mere design, rather than actual
performance, of Defendants’ system, such testimony would only further delay trial.” Mot. at 11.
Rule 403 does not bar admission of Dr. Korf’s testimony.
IV.

CONCLUSION.
Dr. Korf is a well-regarded, published computer science professor with an impeccable

CV and a background in artificial intelligence. His opinions are firmly rooted in his background
and will be helpful to the jury. Plaintiff’s eleventh hour motion in limine to exclude his
testimony should be denied.
Dated: February 27, 2012

Respectfully Submitted,

/s/William A. Delgado
William A. Delgado
WILLENKEN WILSON LOH & LIEB LLP
707 Wilshire Boulevard, Suite 3850
Los Angeles, CA 90017
[phone]
[e-mail]
Lead Counsel for Defendants
10
121061.1

CERTIFICATE OF SERVICE

I hereby certify that on February 27, 2012, I electronically filed the foregoing
paper with the Court using the ECF system which will send notification of such filing to
the following:
Enrico Schaefer (P43506)
Brian A. Hall (P70865)
TRAVERSE LEGAL, PLC
810 Cottageview Drive, Unit G-20
Traverse City, MI 49686
231-932-0411
[e-mail]
[e-mail]
Lead Attorneys for Plaintiff

Nicholas J. Stasevich (P41896)
Benjamin K. Steffans (P69712)
BUTZEL LONG, P.C.
150 West Jefferson, Suite 100
Detroit, MI 48226
[phone]
[e-mail]
[e-mail]
Local Counsel for Defendants

Anthony P. Patti (P43729)
HOOPER HATHAWAY, PC
126 South Main Street
Ann Arbor, MI 48104
734-662-4426
[e-mail]
Attorneys for Plaintiff

William A. Delgado
WILLENKEN WILSON LOH & LIEB LLP
707 Wilshire Boulevard, Suite 3850
Los Angeles, CA 90017
[phone]
[e-mail]
Lead Counsel for Defendants

/s/William A. Delgado
William A. Delgado
WILLENKEN WILSON LOH & LIEB LLP
707 Wilshire Boulevard, Suite 3850
Los Angeles, CA 90017
[phone]
[e-mail]
Lead Counsel for Defendants

11
121061.1