Decision Insights, Inc. v. Sentia Group, Inc.

UNPUBLISHED

UNITED STATES COURT OF APPEALS
FOR THE FOURTH CIRCUIT

No. 07-1596

DECISION INSIGHTS, INCORPORATED,

Plaintiff - Appellant,

v.

SENTIA GROUP, INCORPORATED; THOMAS H. SCOTT; MARK
ABDOLLAHIAN; JACEK KUGLER; BRIAN EFIRD,

Defendants - Appellees.

Appeal from the United States District Court for the Eastern
District of Virginia, at Alexandria. Claude M. Hilton, Senior
District Judge. (1:06-cv-00766)

Argued: September 25, 2008 Decided: February 12, 2009

Before WILLIAMS, Chief Judge, WILKINSON, Circuit Judge, and
Richard L. VOORHEES, United States District Judge for the
Western District of North Carolina, sitting by designation.

Affirmed in part, reversed in part, and remanded by unpublished
per curiam opinion.

Nicholas Hantzes, HANTZES & REITER, McLean, Virginia, for
Appellant. Edward Francis O’Connor, O’CONNOR, CHRISTENSEN &
MCLAUGHLIN, Irvine, California, for Appellees.

Unpublished opinions are not binding precedent in this circuit.
PER CURIAM:

Decision Insights, Inc. (“DII”) appeals the district

court’s grant of summary judgment on all claims in favor of

Sentia Group, Inc. (“Sentia”), Mark Abdollahian (“Abdollahian”),

Brian Efird (“Efird”), Jacek Kugler (“Kugler”), and Thomas H.

Scott (“Scott”). DII also appeals an adverse sanctions ruling

for purported discovery violations.

The underlying civil action arises from a dispute between

DII and Sentia surrounding the latter’s development and use of a

competing software application that implements a decision-making

model using expected utility theory. 1 More specifically, DII

alleges that Abdollahian, Efird, Kugler, and Carol Alsharabati

(“Alsharabati”), who was not named as a defendant, disclosed

trade secrets to Sentia in violation of Virginia’s Trade Secret

Misappropriations Act, Va. Code Ann. § 59.1-336. DII also

asserts that Efird, Kugler, and Abdollahian breached their

respective contractual and fiduciary obligations by disclosing

other confidential and proprietary information protected by DII.

DII alleges that Scott conspired with Sentia to induce DII’s

former employees to breach their agreements with DII. Because

the district court did not consider DII’s software compilation

1
Expected utility theory, described below, encompasses
several disciplines, including mathematics, economics, political
science, and psychology.

2
claim as a separate and independent alleged trade secret, we

affirm in part, reverse in part, and remand with instructions.

I.

A.

DII first developed software called a “Dynamic Expected

Utility Model” (“EU Model”) in the nineteen-eighties. 2 The EU

Model, DII’s primary asset, is an analytical tool used in

preparing negotiating strategies by assessing risk, comparing

the impact of differing operating positions, and detailing

trade-offs among various alternatives. 3 DII has used the EU

2
DII’s “Dynamic Expected Utility Model” (“EU Model”) is
also known as DII’s “Political Analysis Information System”
(“PAIS”) software.
3
DII first defines the component issues and then implements
state-of-the-art data collecting procedures (including
utilization of a subject area expert), in order to identify the
key data relative to each of the following:

1) identification of the stakeholders (i.e., groups,
individuals, companies or governments) with potential
interest in issue;
2) identify and quantify the policy positions of each
stakeholder;
3) identify and quantify the resources that each
stakeholder may employ to influence their or its
preference on the issue; and
4) identify and quantify the actual importance each
stakeholder attaches to the policy outcome, thereby
deriving their salience toward the issue.

(JA at 888-89) After research and data collection, numerical
values are associated with the responses to these four elements.
(Continued)
3
Model in the operation of its business since its inception in

1989. DII owns the assets, copyright, and all proprietary rights

to the EU Model.

Dr. Bruce Bueno de Mesquita (“Dr. Bueno de Mesquita”), a

former employee of DII and leading published authority on

expected utility theory generally, created the original source

code for DII’s software in the mid-1980s. 4 Gary Slack (“Slack”),

a DII analyst and member of DII’s Board of Directors, modified

and updated DII’s software in the early 1990s. Slack testified

that he and Dr. Bueno de Mesquita essentially wrote DII’s

software program from scratch.

In 1998, DII hired Carol Alsharabati to make additional

modifications to the EU Model. Alsharabati, who then lacked any

formal or informal computer programming training, was provided a

copy of DII’s computer code and required to sign a

Based on this data, the EU Model computer software calculates
dynamic bargaining positions with respect to stakeholder
positions over bargaining rounds based on calculating and
predicting changes in stakeholder positions.
4
The terms “source code” or “code” refer to “a document
written in computer language which contains a set of
instructions designed to be used directly or indirectly in a
computer to bring about a certain result.” Trandes Corp. v. Guy
F. Atkinson Co., 996 F.2d 655, 655 (4th Cir. 1993).

4
confidentiality agreement. 5 Alsharabati’s work with DII was her

first experience writing code for an EU Model. Alsharabati had

a copy of the DII code on her computer but testified that she

erased it after her work for DII was complete. Alsharabati

concedes she gained valuable experience while working on the DII

project.

During their work on behalf of DII, Abdollahian, Efird, and

Kugler all worked with Alsharabati and had access to the DII

source code for the EU Model and the other alleged confidential

and proprietary materials DII now seeks to protect. 6 Both

Abdollahian and Efird entered into Trade Secret Nondisclosure

Agreements (“Agreements”) with DII. 7 Efird’s contract also

included a restrictive covenant not to compete. Kugler executed

an Agreement that was never signed by DII. Between October 2001

and December 2002, Abdollahian, Efird, and Kugler all left DII

to form Sentia Group, Inc.

5
Alsharabati, who describes herself as “self-taught” in the
field of computer programming, has masters and doctorate degrees
in political science.
6
Abdollahian was in an exclusive consulting role with DII
between 2000 and 2002. Efird was employed with DII during the
same time period. Kugler was a director and major owner of
stock in DII through December 2002.
7
Abdollahian refused to sign a Consultant Agreement
presented to him by DII in 1997.

5
 B.

On November 5, 2002, Abdollahian, Kugler, and Scott

incorporated Sentia Group, Inc. Scott was appointed Sentia’s

Chief Executive Officer, Abdollahian became Sentia’s Chief

Operating Officer, and Efird became an executive vice president.

Kugler, a major shareholder, performed consultant work.

According to Scott, Sentia was initially formed with the

idea that Sentia would obtain a software license from DII and

the two companies would divide responsibilities based upon

geographic territory. In late 2002, while acting on behalf of

Sentia, Kugler attempted to negotiate a nonexclusive worldwide

royalty license with DII but the parties did not reach

agreement.

Rather than continue to negotiate with DII, Sentia decided

to develop its own software application to perform the same

essential functions and analysis as DII’s software. Sentia

sought legal advice and was advised in December 2002 as follows:

“[W]e emphasize that preferably any individuals who had contact

with or access to the code of the prior company not be involved

in development of the new software program.” (JA at 790) Counsel

cautioned Sentia that if this were deemed unavoidable given the

requisite expertise, the “new code [should] bear no resemblance,

functionally, structurally, or otherwise, to the code of the

prior company.” (JA at 790) Sentia’s counsel also suggested that

6
Sentia carefully document aspects of the development process -

its intended function as well as design development and actual

development. 8 (JA at 791)

At the suggestion of Abdollahian, Kugler, or both, and

notwithstanding counsel’s advice, Sentia hired Alsharabati to

develop software for Sentia. Working with a group of software

students with no prior experience in EU Models, Alsharabati

wrote Sentia’s first code in what DII describes as “record

time,” or approximately six weeks. 9

According to DII, Sentia’s software is the same as its own

EU Model in terms of method. DII alleges that in running both

programs, its comparisons obtain “equal results.” 10 DII contends

that achieving equal results is not possible unless all of the

8
The same letter explained counsel’s understanding that
because Sentia sought to implement “vastly new and improved
theoretical models, and that any theoretical models used by the
prior company have largely been the subject of academic
publications and are well known in the field, it appears
possible to develop a new software program that has no
substantial similarity to the software program used by the prior
company.” (JA at 790-91)
9
Sentia’s software was initially named the “Machiavelli”
code, and later referred to as the “Senturion” application or
model. Although presently in different computer languages, the
first version of Sentia’s code was in the same computer language
as DII’s code – Visual Basic.
10
DII explains that when it compares the two programs,
Sentia’s output results are identical to DII’s model “to 2
decimal places of accuracy.”

7
parameters, variables and sequencing associated with the

expected utility are equal. DII also claims that the Machiavelli

code contains portions of DII’s source code which are “commented

out” and that this fact provides “proof positive” that Sentia

used the DII code as a reference in writing the Sentia software

program. 11

DII asserts that Sentia is conducting business in direct

competition with DII and using the EU Model in its business to

DII’s detriment.

C.

DII commenced its suit on June 30, 2006, and alleged causes

of action for breach of contract (Count I), conspiracy to commit

breach of contract (Count II), conspiracy to injure another in

trade, business or profession (Count III), misappropriation of

trade secrets (Count IV), breach of fiduciary duty (Count V),

and conversion (Count VI). 12

11
Here, the term “comment” refers to computer language in a
source code that is not part of the functional code, but
functions instead as a guide to future programmers working with
the source code and a mechanism for explaining changes in the
development of the code.
12
DII does not advance its claims for conspiracy to commit
breach of contract, conspiracy to injure another in trade,
business or profession, breach of fiduciary duty, or conversion
on appeal and is therefore deemed to have abandoned those
particular claims. Thus, only Counts I and IV remain.

8
 The parties grappled with the framing of the legal issues

before the district court. On February 13, 2007, the magistrate

judge granted partial relief on a motion to compel filed by

Sentia. The magistrate judge determined that DII’s discovery

responses were inadequate, in part, due to DII’s failure to

identify its trade secrets with specificity. The magistrate

judge held Sentia’s request for monetary sanctions in abeyance

and directed DII to produce:

“[A] clear and express verified statement containing
only those items which Plaintiff considers to be
actual trade secrets and which Plaintiff has
reasonable grounds to believe were misappropriated by
Defendant. Plaintiff shall clearly differentiate
between the material which is public knowledge from
that material which is allegedly Plaintiff’s trade
secret, proprietary, or confidential material.”

(JA at 281)(emphasis added).

On February 20, 2007, Sentia filed a second motion for

sanctions claiming that DII’s Fourth Supplemental Answer to

Interrogatories was still deficient. DII’s Fourth Supplemental

response separately identified each of the twelve components of

the code as processes implemented within the code. DII

attempted to identify each individual component by including the

lines of source code (i.e., mathematical equations) that

corresponded to each. DII’s response unequivocally identified as

a trade secret its entire DII source code as a total

compilation.

9
 On February 23, 2007, the magistrate judge conducted a

hearing to resolve several discovery disputes, including

Sentia’s motion for sanctions. Sentia argued that because DII

did not allege Sentia copied its code, the only thing left was

DII’s claim regarding the 12 processes. 13 The parties engaged in

a lengthy debate over whether lines of code were sufficient to

identify and define the alleged proprietary process claims.

Sentia argued that, as framed by DII, Sentia was unable to

defend on the trade secret claims.

On March 2, 2007, DII produced to Sentia an expert report

prepared by Gary Slack containing additional narrative and

detailed flow charts showing the structure of the code and

identifying each alleged proprietary process. The same report

contained the identity and description of the alleged

proprietary processes, the variables, the constants and the

parameters. Slack’s report also explained how the DII code

operates as a whole.

On March 5, 2007, the magistrate judge directed DII to

produce by March 9, 2007, “to the extent they exist, all

13
Sentia’s counsel made the following argument: “Remember
their specific representation isn’t that they [Appellees] copied
their code, although they say the entire code is trade secret,
because their own expert acknowledges we didn’t copy their code,
it’s what lies within the code.” (JA at 88, 91)

10
algorithms, block flow diagrams, narratives, and other documents

associated with the development of the twelve sections” of

software code DII asserts constitute trade secrets as well as

“all other sections of software code” DII has identified as its

trade secrets. 14 (JA at 125)(emphasis added) The magistrate judge

awarded Sentia the costs and attorneys’ fees associated with its

original motion to compel and its initial motion for sanctions.

On March 9, 2007, DII requested clarification as to whether

the magistrate judge contemplated production of existing

documentation only, or whether DII was expected to reverse

engineer algorithms from its current version of the source

code. 15 DII explained in its motion that, because the EU Model

was created over fifteen years ago, DII no longer had in its

possession documentation associated with the development of the

EU Model, including algorithms that would have been initially

14
An algorithm is “[a] step-by-step problem-solving
procedure, especially an established, recursive computational
procedure for solving a problem in a finite number of steps.”
THE AMERICAN HERITAGE DICTIONARY OF THE ENGLISH LANGUAGE 45 (3rd
ed. 1992).
15
The term “reverse engineer” means “to analyze a product
to try to figure out its components, construction, and inner
workings, often with the intent of creating something similar.”
WEBSTER’S NEW MILLENIUM DICTIONARY OF ENGLISH (Preview ed.
2008), available at http://www.dictionary.com.

11
relied upon. 16 DII reiterated the significant costs (as much as

$100,000) associated with reverse engineering the existing

source code for the purpose of creating new algorithms. DII

also submitted additional expert reports by Dr. Bueno de

Mesquita and Andrew Fahey. 17

On March 16, 2007, the magistrate judge issued an order

clarifying his March 5, 2007 ruling. The magistrate judge

explained that at the sanctions hearing held on February 23,

2007, the Court told DII that “it was not required to create any

algorithms, block flow diagrams, narratives, and other documents

associated with the development of its software or engage in

reverse engineering, but that [DII] should search and produce

any such responsive documents which already exist.” (JA at 144)

The order then stated that in light of DII’s representation that

no such responses exist, “this discovery matter is closed.” (JA

16
DII explained that its failure to retain such records is
not suspect. According to DII, unavailability can be attributed
to the fact that once a software code is debugged and made
operational, the original algorithms are of little value.
Similarly, as the software code is improved upon, the original
algorithms are seldom updated or referenced.
17
The Bueno de Mesquita report addressed Sentia’s claim
that certain DII processes were in the public domain and could
not be considered trade secrets. The Fahey report sought to
identify the alleged proprietary portions of the DII code also
found within the Sentia code.

12
at 144-45) Sanctions were imposed against DII for a total of

$13,256.25. 18

DII objected to the magistrate judge’s ruling. On March

30, 2007, the district court summarily affirmed the magistrate

judge’s March 5, 2007 Order, finding that the imposition of

sanctions was not “clearly erroneous or contrary to law.” (JA at

224-25)

D.

Sentia moved for summary judgment on all of DII’s causes of

action. The district court heard oral argument and opined on

June 5, 2007, that summary judgment was proper on all of DII’s

claims. Regarding Claims III through VI, the district court

found that DII failed to meet its burden as to the existence of

a trade secret. The district court likewise based his ruling on

Counts I and II, the non-trade secret claims, on DII’s failure

to identify “any confidential or proprietary information

obtained by [Appellees] while employed at DII that were

thereafter misappropriated.” (JA at 271) The district court

also ruled that DII did not have an enforceable contract to

assert against Kugler.

18
After briefing, the magistrate judge held that DII was
subject to sanctions in the amount of $13,256.25. The award to
Sentia was based upon costs in the amount of $2,956.25 and
$10,300 in attorneys’ fees.

13
 DII’s appeal is timely and we have jurisdiction pursuant to

28 U.S.C. § 1291.

II.

On appeal, DII contends the trial court erred by finding,

as a matter of law, that DII failed adequately to identify any

trade secrets relating to its software. DII also challenges the

trial court’s rulings on its contractual claims as well as the

imposition of monetary sanctions for the alleged failure to

comply with its discovery obligations.

III.

This court reviews the district court’s decision granting

summary judgment de novo. See Cont’l Airlines, Inc. v. United

Airlines, Inc., 277 F.3d 499, 508 (4th Cir. 2002).

Under Rule 56(c) of the Federal Rules of Civil Procedure,

summary judgment may be granted where Athe pleadings,

depositions, answers to interrogatories, and admissions on file,

together with affidavits, if any, show that there is no genuine

issue of material fact and that the moving party is entitled to

judgment as a matter of law.@ Fed. R. Civ. P. 56(c); Anderson v.

Liberty Lobby, 477 U.S. 242 (1986); Celotex Corp. v. Catrett,

477 U.S. 317 (1986). A genuine issue exists only if Athe evidence

14
is such that a reasonable jury could return a verdict for the

non-moving party.@ Anderson, 477 U.S. at 248.

Under Rule 56(e), “an adverse party may not rest upon the

mere allegations or denials of his pleading, but his response,

by affidavits or as otherwise provided in this rule, must set

forth specific facts showing that there is a genuine issue for

trial. . . .” Fed. R. Civ. P. 56(e). Thus, in order to

survive summary judgment, DII is required to produce evidence

setting forth specific facts that demonstrate the existence of a

genuine issue for trial. In conducting its analysis, the court

must view the evidence in the light most favorable to the non-

moving party. See Celotex Corp., 477 U.S. at 325.

A. Trade Secret Claims – Va. Code Ann. § 59.1-336

The success of DII’s appeal largely depends upon whether

DII presented sufficient evidence at summary judgment in support

of its contention that its software may be deemed a trade

secret.

Virginia’s version of the Uniform Trade Secrets Act defines

a “trade secret” as “information, including but not limited to a

formula, pattern, compilation, program, device, method,

technique, or process, that:

1. Derives independent economic value, actual or
potential, from not being generally known to, and not being
readily ascertainable by proper means by, other persons who
can obtain economic value from its disclosure or use, and

15
 2. Is the subject of efforts that are reasonable under the
circumstances to maintain its secrecy.

Va. Code. § 59.1-336 (emphasis added). 19

“The crucial characteristic of a trade secret is secrecy

rather than novelty.” Dionne v. Southeast Foam Converting &

Packaging, Inc., 397 S.E.2d 110, 113 (Va. 1990); Avtec Syss.,

Inc. v. Peiffer, 21 F.3d 568, 575 (4th Cir. 1994) (same). “The

secrecy need not be absolute; the owner of a trade secret may,

without losing protection, disclose to a licensee, an employee,

or a stranger, if the disclosure is made in confidence, express

or implied.” Dionne, 397 S.E.2d at 113 (citing Kewanee Oil Co.

v. Bicron Corp., 416 U.S. 470, 475 (1974)); Trandes Corp. v. Guy

F. Atkinson Co., 996 F.2d 655, 664 (4th Cir. 1993). “Although

the subject of a trade secret may be novel in the sense that it

is something generally unknown in the trade or business,

“[n]ovelty, in the patent law sense, is not required for a trade

secret.” Id. (citing Kewanee Oil Co., 416 U.S. at 476.)

Whether or not a trade secret exists is a “fact-intensive

question to be resolved at trial.” Hoechst Diafoil Co. v. Nan

Ya Plastics Corp., 174 F.3d 411, 419 (4th Cir. 1999); Trandes,

996 F.2d at 661; Microstrategy, Inc. v. Li, 601 S.E.2d 580, 589

19
Virginia’s Trade Secrets Act is modeled after the Uniform
Trade Secrets Act. See Dionne, 397 S.E.2d at 114; Avtec Syss.,
21 F.3d at 574 (Virginia’s statute “closely tracks the Uniform
Trade Secrets Act.”)

16
(Va. 2004)(“[T]he determination whether a trade secret exists

ordinarily presents a question of fact to be determined by the

fact finder from the greater weight of the evidence.”)

DII’s first trade secret claim is founded upon its software

as a total compilation. In addition, DII contends that twelve

specific functions within the DII Code amount to one or more

protected trade secrets. DII suggests that the proper analysis

is to evaluate each identified trade secret claim independently

as in Trandes. See e.g., Trandes, 996 F.2d 655 (4th Cir. 1993)

(applying identical Maryland’s Uniform Trade Secrets Act in

post-trial context). We agree. Because the district court did

not consider whether DII’s entire software compilation might

qualify as a “trade secret” under the Virginia statute, we

remand to the district court with guidance as follows:

1. Software Compilation Claim

The district court found that, “Plaintiff [DII] could not

provide adequate identification of its trade secrets and

confidential information, making it almost impossible for this

Court and Defendants [Sentia] to ascertain what aspects of the

EU Model are trade secrets, and which portions of the code are

publicly available.” (JA at 270) The district court did not

address whether or not the software program, as a total

compilation, could qualify as a trade secret.

17
 Understandably, identification of DII’s alleged trade

secrets presented difficulty for the court. As noted, supra, the

parties argued over DII’s actual legal theory. Sentia insisted

that because DII did not assert a copyright claim, the

application itself was not at issue. Sentia dedicated little

time to DII’s software compilation claim. Sentia’s expert

devoted only one paragraph within his original expert

declaration to this aspect of DII’s trade secret claim. Before

DII produced the flowchart and block diagram of its source code,

Dr. Alexander stated:

“I am unable to respond to the first
identification relating to the entire code to the
engine of DII’s software as a compilation.” In the
source code provided I can see numerous standard Basic
language extensions to Basic that Microsoft itself
would most likely consider proprietary. DII
proprietary contributions, if any, are not apparent
from the entirety of the code module.”

(JA at 200) Similarly, Dr. Alexander’s rebuttal report focused

almost entirely on DII’s twelve process claims. Citing no legal

authority, Sentia then faulted DII for its failure to produce

algorithms corresponding to its source code.

Our opinion in Trandes is instructive regarding DII’s

burden. Trandes involved a computer program that performed

survey calculations for the construction of subway tunnels. See

Trandes, 996 F.2d at 657. In addition to an independent software

compilation claim, the Trandes plaintiff alleged that both the

18
“specific engineering formulas and methods of calculation

embodied in the Tunnel System” and “the structure and

organization of the Tunnel system modules” constituted

additional trade secrets. Trandes, 996 F.2d at 661, 662 n.7.

These two claims were dismissed, however, because Trandes did

not provide “any information whatsoever about the formulas” and

likewise failed to explain “how the program was structured [or]

how the program was organized.” Trandes, 996 F.2d at 661-662

(plaintiff is required “to describe the subject matter of its

alleged trade secrets in sufficient detail to establish each

element of a trade secret”). Although the Trandes plaintiff was

ultimately unsuccessful on two of its trade secret claims, we

determined that Trandes presented sufficient evidence that the

software itself, which was identified by source code and

produced at trial, constituted a trade secret. Trandes, 996

F.2d at 662-663. Accordingly, we upheld the jury’s verdict that

the software compilation was a protected trade secret. Id.

Thus, Trandes teaches that a plaintiff’s alleged software

compilation trade secret is to be analyzed separate and apart

from other software trade secret claims, and that production of

source code is an acceptable method of identifying an alleged

compilation trade secret. Trandes, 996 F.2d at 661-63.

With respect to algorithms, DII represents that because its

code was created over fifteen years ago, it had none to produce.

19
However, in addition to producing a complete copy of its source

code, DII also presented detailed block diagram flow charts as

well as expert testimony in support of its position that its

source code is unique. Sentia’s own expert recognized that a

flow chart is an acceptable and “equally precise” alternative to

the production of algorithms for purposes of identification of

alleged proprietary software. (JA at 197-98)

DII produced its entire source code, as well as a flow

chart and narrative explaining its software program as a whole.

Accordingly, we remand DII’s software compilation claim to the

district court for independent consideration. On remand, should

the district court determine that DII adequately identified its

software compilation claim, the district court should then

consider the sufficiency of DII’s showing as to the existence of

a trade secret and thence a triable issue of fact. In doing so,

the district court should specifically address the relevant

criteria for establishing the existence of a trade secret under

Va. Code. § 59.1-336, namely, whether or not the compilation has

independent economic value, is generally known or readily

ascertainable by proper means, and is subject to reasonable

efforts to main secrecy. 20 If, in light of these statutory

20
On the question of whether or not DII’s software
compilation is generally known or readily ascertainable by
proper means, we refer the district court to our opinion in
(Continued)
20
criteria, the district court finds that a triable issue is

presented, it should next consider whether sufficient evidence

of misappropriation exists to survive summary judgment.

2. Twelve Process Claims

DII also alleges that each of twelve individual portions of

the program within its source code (i.e., the twelve process

claims) constitute a trade secret. The parties’ experts disagree

regarding the adequacy of identification and proprietary status.

DII attempts to identify each of the twelve individual

components by including the lines of source code that apply or

correspond to each. DII’s expert explains that the twelve

alleged trade secret functions “are not located in a single

location in the Code, and therefore cannot easily be isolated

independent of the other code as currently written . . . .” (JA

at 188) DII also contends that “the annotation of the Code

which identifies the location of each of the functions

eliminates this impediment to identifying their functions within

the Code.” (JA at 188) DII does not describe what the lines of

code teach, or how they translate to a protectable trade secret.

Servo Corp. of Am. v. Con. Elec. Co., 393 F.2d 551, 554 (4th
Cir. 1968) (recognizing that plaintiff’s trade secret “might
consist of several discrete elements, any one of which could
have been discovered by study of material available to the
public . . . .”)

21
 Sentia’s expert persuasively describes the difficulty in

analyzing the twelve processes independently. According to Dr.

Alexander, if each individual process is considered

independently, the information provided by DII is incomplete and

fragmented. We agree that the information on the twelve process

claims is presented by DII in such a way as to prohibit

meaningful analysis by Sentia, the court, or a jury. For this

reason, we find that DII has not met its evidentiary burden with

respect to the twelve process claims and we affirm the district

court’s grant of summary judgment on this issue.

3. Other Proprietary Claims

DII’s other claims seek protection of DII reports

containing marketing and research material, specific information

identified in DII’s user manual, and specific client contact

information. The district court determined that Counts III

through VI “presuppose the existence of confidential information

and trade secrets” and that DII’s “failure to identify [any such

information] with reasonable particularity” required dismissal.

(JA at 270) As a result, these specific categories of alleged

proprietary materials were not discussed by the district court

at all. Depending on the circumstances, any of this information

could be characterized as trade secrets. (See Section “III,

A.”) On remand, the district court will have an opportunity to

22
consider DII’s other proprietary claims under the statutory

criteria consistent with this court’s opinion.

B. Contractual Claims

1. Trade Secret Nondisclosure Agreements

Abdollahian entered into an Agreement on December 12, 1994,

agreeing not to disclose DII’s proprietary information. Efird

signed a similar Agreement on April 3, 1998. The Agreements

entered into by DII’s former employees contain nearly identical

language and call for the application of Virginia state law.

The Agreements include provisions for assignment of work product

to DII and the return of confidential material upon agreement

termination. The confidentiality clauses, entitled “Covenant To

Retain Confidence,” read as follows:

The Consultant / Representative 21 acknowledges that he
will, as a result of an association with Decision
Insights, Inc., have access to and be in a position to
receive information of a confidential or proprietary
nature including trade secrets. The Consultant /
Representative agrees that he will not, during the
association with Decision Insights or thereafter,
disclose to anyone whomsoever or use in any manner
whatsoever any confidential or proprietary
information, whether patentable or unpatentable,
concerning any inventions, discoveries, improvements,
processes, methods, trade secrets, research or secret
data (including but not limited to, models, formulas,
computer programs and software developments), or other
confidential matters possessed, owned, or used by

21
In their respective Agreements, Abdollahian is identified
as a “Consultant” and Efird is identified as a “Representative.”

23
 Decision Insights that may be obtained or learned by
the Consultant / Representative in the course of, or
as a result of his association with Decision Insights,
except as such disclosure or use may be required in
the normal course of doing business with Decision
Insights and pursuant to Decision Insights Inc.[‘s]
prior written consent.

(JA at 386-387) The Agreements provide that the agreement shall

continue to bind the parties after their association ends. (JA

at 387)

With respect to Abdollahian and Efird, the district court

did not discuss enforceability of the respective confidentiality

provisions. Rather, the district court relied upon an asserted

lack of evidence of a breach and simply stated that DII “failed

to come forward with any evidence identifying any confidential

or proprietary information obtained by Defendants while employed

at DII that were[sic] thereafter misappropriated.” (JA at 271)

Because this court has determined that remand is proper

with respect to DII’s claim that its software as a compilation

may constitute a trade secret, remand is also proper on the

contractual claims in order for the district court to address

DII’s contractual claims in light of its findings with respect

to the existence of a trade secret.

Kugler was presented with a similar Trade Secret

Nondisclosure Agreement but DII never executed it. DII contends

that the parties’ agreement is reflected within a document

signed by Kugler on January 30, 1998. (JA at 860-62) Per

24
handwritten additions to the typed text (initialed “JK”), the

Agreement carves out an exception for academic use of Kugler’s

work product. (JA at 861) DII posits that Kugler intended to be

bound to the terms set forth in the January 30, 1998 document,

particularly confidentiality and nondisclosure, notwithstanding

the fact that DII never executed the written contract.

DII relies upon the Virginia State Supreme Court’s opinion

in Manss-Owens, which held that “the mere fact that a written

contract was contemplated does not necessarily show that no

binding agreement had been entered into.” Manss-Owens Co. v.

H.S. Owens Son, 105 S.E. 543, 547 (Va. 1921). The rationale for

the rule is explained as follows:

The whole question is one of intention. If the parties
are fully agreed, there is a binding contract,
notwithstanding the fact that a formal contract is to
be prepared and signed; but the parties must be fully
agreed and must intend the agreement to be binding.
If though fully agreed on the terms of their contract,
they do not intend to be bound until a formal contract
is prepared, there is no contract, and the
circumstances that the parties do intend a formal
contract to be drawn up is strong evidence that they
did not intend the previous negotiations to amount to
an agreement.

If it appears from the evidence that the minds of the
parties have met; that, on the one side, there was a
proposition for a contract, which proposition has been
accepted by the other party; that the terms were in
all respects agreed upon; and that a part of the
mutual understanding was that a written contract
embodying those terms should thereafter be executed by
the respective parties – there results an obligatory
contract which neither party is at liberty to
repudiate.

25
Manss-Owens Co., 105 S.E. at 547 (quoting Boisseau v. Fuller, 30

S.E.457 (Va. 1898)); see also Charbonnages de France v. Smith,

597 F.2d 406, 414-16 (4th Cir.1979) (explaining that questions

of mutual assent and the parties’ intentions are

“quintessentially disputes about ‘states of mind’” and that

“subjective states and objective manifestations of intention

present interpretive issues traditionally understood to be for

the trier of fact.”)

DII claims a genuine issue of material fact exists with

respect to whether Kugler intended to be bound in light of

Kugler’s deposition testimony that he had an agreement with DII

that he could use DII’s technology for academic purposes. (JA at

1376-77) According to DII, it would never have shared its

proprietary information with Kugler had he not intimated

agreement not to disclose its trade secret and confidential

information. In fact, after their disassociation, on January

22, 2003, Kugler wrote to DII to assure the company that he had

no intention of disclosing or making improper use of any

confidential DII information. In the same letter, Kugler refers

to the modifications he made to the January 30, 1998 document

and states, “it is unclear to me if the agreement was

consummated or to what extent the terms of such an agreement are

even enforceable.” (JA at 864, 1378) For these reasons, we also

26
remand as to this issue. Should the district court determine on

remand that DII in fact possessed trade secret, confidential, or

proprietary information, the district court should likewise

consider whether an implied agreement existed between DII and

Kugler as suggested by DII that prohibited Kugler from

disclosing this information. 22

2. Non-Competition Clause

Efird’s Agreement contained a non-competition clause or

restrictive covenant. Paragraph 4 of the Agreement reads in

pertinent part:

[D]uring the term of this agreement and for a period
of two years after termination of association, the
representative shall not, for any reason, directly or
indirectly, enter into or engage in any business
competition with the precise business as it now exists
[or] may exist at any time during the period of the
representative’s engagement . . . .

(JA at 387) The restrictive covenant seeks to prohibit Efird

from directly or indirectly engaging in any “business

competition” with DII’s “precise business as it now exists [or]

22
Sentia’s brief is of little help. Sentia fails to cite to
the record, or any case law, in support of its argument that the
district court correctly found, as a matter of law, that no
contract existed. In addition, Sentia confuses Abdollahian and
Kugler in its terse discussion of the contractual claims.
(Sentia claims that the Abdollahian contract was never signed by
DII. That is incorrect. Rather, the Kugler agreement is the one
DII never executed.)

27
may exist at any time during the period of [Efird’s] engagement”

with DII.

Under Virginia law, the following criteria determine
the validity of non-competition agreements:

(1) Is the restraint, from the standpoint of the
employer, reasonable in the sense that it is no
greater than is necessary to protect the employer
in some legitimate business interest?

(2) From the standpoint of the employee, is the
restraint reasonable in the sense that it is not
unduly harsh and oppressive in curtailing his
legitimate efforts to earn a livelihood?

(3) Is the restraint reasonable from the
standpoint of a sound public policy?

Non-competition covenants which pass these tests in
the light of the facts of each case will be enforced
in equity.

Blue Ridge Anesthesia & Critical Care, Inc. v. Gidick, 389

S.E.2d 467, 470 (Va. 1990) (quoting Roanoke Eng’g Sales v.

Rosenbaum, 290 S.E.2d 882, 884 (Va. 1982)). In other words, the

Court must determine whether the non-competition clause or

restraint is greater than necessary to protect DII’s interest or

unreasonable in limiting Efird’s ability to obtain other

suitable employment. See Blue Ridge Anesthesia & Critical Care,

Inc., 389 S.E.2d at 470.

Virginia law does not generally favor restrictive covenants

because such covenants are a restraint on trade. See Grant v.

Carotek, 737 F.2d 410, 411-412 (4th Cir. 1984). For this reason,

restrictive covenants are strictly construed against the

28
employer. Grant, 737 F.2d at 412; accord Roanoke Eng’g Sales

Co., 290 S.E.2d 882 (Va. 1982) (other citations omitted).

Moreover, the employer bears the burden of proving that the

restraint is reasonable under the circumstances of the case.

Id. (citing, inter alia, Richardson v. Paxton Co., 127 S.E.2d

113 (Va. 1962)).

In this case, Sentia attacks this provision of Efird’s

contract as ambiguous and overbroad. Sentia’s chief criticism

of the non-competition agreement is that the contract is vague

regarding the business of the company. Indeed, the district

court found the non-compete unenforceable on this basis. 23

Construing the non-compete clause against the employer, the

district court determined that the clause was “broader than

necessary” to protect DII’s legitimate business interests and

“unduly restrictive of Efird’s efforts to pursue his

livelihood.” (JA at 273)

In evaluating the reasonableness of the restraint from the

employer’s perspective, the first inquiry necessarily requires

the court to consider the nature of the legitimate business

23
Based on the purported lack of evidence establishing that
DII’s business was conducted worldwide, the district court also
found the absence of a geographic limitation unreasonable. The
district court likewise found the two-year time limitation
unreasonable. Given our analysis, we need not discuss these
issues.

29
interest at stake, namely, whether DII possessed trade secrets

or other confidential and proprietary information. See Meissel

v. Finley, 95 S.E.2d 186, 191 (Va. Ct. App. 1956) (“The

possession of trade secrets and confidential information is an

important consideration in testing the reasonableness of a

restriction on competition.”)(citing Stoneman v. Wilson, 192

S.E. 816, 819 (Va. 1937)); But see Paramount Termite Control

Co., Inc. v. Rector, 380 S.E.2d 922, 925 (Va. 1989) (“Although

often used as a justification for non-competition agreements, it

is not necessary that the employees actually had acquired or

possessed specific information that could be legally defined as

confidential or a trade secret, . . . .”)(internal quotations

omitted). Here, the district court determined, in effect, that

DII did not have any legitimate business interests worthy of

protection. As a result, the district court’s analysis of the

restrictive covenant was likely skewed by its conclusion that

DII failed to demonstrate the existence of a trade secret.

Each of Virginia’s tests for validity of a non-competition

clause prompts a reasonableness inquiry in which the analysis

would necessarily include consideration of the existence of a

trade secret to be protected. More importantly, the competing

interests of the employer and employee must be balanced by the

court and then squared with public policy. As explained by the

Supreme Court of Virginia, “[t]hese standards have been

30
developed over the years to strike a balance between an

employee’s right to secure gainful employment and the employer’s

legitimate interest in protection from competition by a former

employee based on the employee’s ability to use information or

other elements associated with the employee’s former

employment.” Omniplex World Servs. Corp. v. U.S. Investigations

Servs., Inc., 618 S.E.2d 340, 342 (Va. 2005) (citing Worrie v.

Boze, 62 S.E.2d 876, 882 (1951)). For these reasons, we

conclude that remand is also proper on this issue so that the

requisite balancing and analysis may be conducted by the trial

court.

IV.

DII’s final argument on appeal is its challenge of the

district court’s March 5, 2007 Order imposing monetary sanctions

for alleged failure to comply with discovery obligations. The

district court’s decision to affirm the magistrate judge’s

sanctions order is reviewed for an abuse of discretion. See

Nat’l Hockey League v. Metro. Hockey Club, Inc., 427 U.S. 639,

642 (1976).

The sanctions order was driven by the district court’s

concern that DII, for strategic reasons, refused adequately to

identify its purported trade secrets. However, the record tends

to show that both the magistrate judge and district court were

31
hampered by less than thorough showings by the parties and did

not come to understand fully the contours of DII’s compilation

argument.

Rule 37 of the Federal Rules of Civil Procedure governs

imposition of sanctions for discovery violations. 24 Rule

37(a)(5)(A)(ii) provides that a district court “must not” order

sanctions if the opposing party’s nondisclosure was

“substantially justified.” A legal position is “substantially

justified” if there is a “genuine dispute” as to proper

resolution or if “a reasonable person could think it correct,

that is, if it has a reasonable basis in law and fact.” See

Pierce v. Underwood, 487 U.S. 552, 565-66 n.2 (1988).

As noted, the alleged factual basis for the imposition of

sanctions was that DII repeatedly responded inadequately to the

discovery requests of Sentia, namely, identification of what it

contended constituted trade secret material. To the extent DII

was deemed to have failed in its efforts adequately to identify

the twelve processes it contended were trade secrets, we agree

with the district court. (Section “III, A, 2,” supra) However,

the parties also legitimately disagreed about what was required

by DII in terms of identification.

24
The magistrate judge did not explain what provision of
Rule 37 he was applying.

32
 As for the software compilation, DII argues first that the

magistrate judge did not expressly direct DII to produce

existing algorithms or other developmental documents prior to

March 5, 2007. In fact, DII contends that the first time

algorithms were even requested by Sentia was during the February

23, 2007 hearing on Sentia’s motion for sanctions. The record

confirms DII’s representation. (JA at 90-91, 95) Sentia

originally requested, and the magistrate judge first ordered, a

narrative description of the alleged trade secrets. In producing

the Slack report, DII complied with the February 17, 2007 Order.

Here, the parties had a “genuine dispute” as to the method

of identifying the alleged trade secrets. Algorithms were not

designated by Sentia or the court as the preferred method of

identification prior to February 23, 2007. In addition, DII had

reasonable cause to believe its production was sufficient in

light of our holding in Trandes. See Maddow v. Proctor & Gamble

Co., Inc., 107 F.3d 846, 853 (11th Cir. 1997) (reliance on case

law is a relevant consideration in determining whether or not a

party’s actions during a discovery dispute are justified).

Accordingly, we find that DII’s failure to produce algorithms

was “substantially justified.” For these reasons, the

imposition of sanctions was not appropriate. On remand, the

district court is instructed to vacate this portion of its

33
earlier order and otherwise proceed in accordance with the

guidance herein provided.

AFFIRMED IN PART,
REVERSED IN PART,
AND REMANDED

34