Computer Associates International, Inc. v. Altai, Inc.

ENOCH, Justice,

delivered the opinion of the Court on Motion for Rehearing,

in which PHILLIPS, C.J., and GONZALEZ, HECHT, CORNYN, SPECTOR and BAKER, JJ., join.

The motion for rehearing is overruled. Our opinion of June 8, 1995, is withdrawn and the following is substituted in its place.

This case comes to us on certified questions from the United States Court of Appeals for the Second Circuit. Computer Assocs. Int'l, Inc. v. Altai, Inc., 22 F.3d 32 (2d Cir.1994).1 We are asked to decide two issues: (1) whether the discovery rule exception to section 16.003(a) of the Texas Civil Practice and Remedies Code applies to claims for misappropriation of trade secrets; and if not, (2) whether the application to such claims of the two-year limitations period provided by section 16.003(a) contravenes the “open courts” provision of Article I, Section 13 of the Texas Constitution. We hold that the discovery rule exception does not apply to the claim for misappropriation of trade secrets and that application of section 16.003(a) does not violate the Texas Constitution.

I

Claude Amey, a developer of computer software, was employed by Computer Associates International, Inc., in its Dallas office from 1978 until January 1984. During his employment, Arney signed an employment agreement which prohibited him from retaining or divulging Computer Associates’ trade secrets. In January 1984, Arney left Computer Associates to accept employment at Altai, Inc. In an exit interview, Arney represented that he retained no proprietary information of Computer Associates and would not divulge Computer Associates’ trade secrets to any third party. However, when Arney left Computer Associates he took copies of the computer source code for two versions of ADAPTER. ADAPTER is an operating system compatibility component of CA-SCHEDULER, which is a job scheduling program for IBM mainframe computers. ADAPTER connects CA-SCHEDULER with the three different operating systems used on IBM mainframe computers and enables CA-SCHEDULER to run on any of the IBM operating systems. ADAPTER was also used with a group of Computer Associates’ programs called the DYNAM line. However, ADAPTER is not a separate product and is not capable of operating as an independent product. Before Arney left Computer Associates, Altai developed ZEKE, a job scheduling program for IBM mainframe computers which was similar to CA-SCHEDULER. In early 1984, Arney copied approximately thirty percent of the ADAPTER source code to write OSCAR 3.4 for Altai. It is undisputed that no one at Altai (other than Arney) knew that Arney possessed the ADAPTER source code or that Amey had copied portions of the source code when he created OSCAR 3.4. OSCAR 3.4 is Altai’s operating system compatibility component which was used in several of Altai’s programs, including ZEKE. Like ADAPTER, OSCAR 3.4 is not a separate product and is not capable of operating as an independent product. From 1985 to August *4551988, Atai used OSCAR 8.4 as a component of several of its computer programs that competed with several of Computer Associates’ programs.

In July 1988, Computer Associates first discovered that Atai had copied and used the ADAPTER source code in several of its computer programs. In August 1988, Computer Associates sued Atai in federal district court for misappropriation of trade secrets and copyright infringement. Among other things, the federal district court determined that Computer Associates’ action for misappropriation of trade secrets under Texas law was preempted by the federal copyright act. Computer Assocs. Int’l, Inc. v. Altai, Inc., 775 F.Supp. 544, 563-66 (E.D.N.Y.1991). The United States Court of Appeals reversed and remanded for further consideration of Computer Associates’ misappropriation claims. Computer Assocs. Int’l, Inc. v. Altai Inc., 982 F.2d 693, 720-21 (2d Cir.1992). On remand, the district court determined that the discovery rule exception did not apply and that Computer Associates’ action for misappropriation of trade secrets was barred by section 16.003(a) of the Texas Civil Practice and Remedies Code. Computer Assocs. Int'l, Inc. v. Altai Inc., 832 F.Supp. 50, 51-52 (E.D.N.Y.1993). Subsequently, the United States Court of Appeals certified the questions to this Court. Computer Assocs. Int’l, Inc. v. Altai Inc., 22 F.3d 32 (2d Cir.1994).

II

Computer Associates argues that the discovery rule exception to the two-year statute of limitations should apply to a claim of misappropriation of trade secrets. A trade secret is any formula, pattern, device or compilation of information which is used in one’s business and presents an opportunity to obtain an advantage over competitors who do not know or use it. Hyde Corp. v. Huffines, 158 Tex. 566, 314 S.W.2d 763, 776, cert. denied, 358 U.S. 898, 79 S.Ct. 223, 3 L.Ed.2d 148 (1958) (quoting Restatement of ToRts § 757 (1939)). A cause of action for misappropriation of trade secrets accrues when the trade secret is actually used. Id., 314 S.W.2d at 769. The parties do not contest that, upon accrual, section 16.003(a) of the Texas Civil Practice and Remedies Code establishes a two-year statute of limitations for injury to the property of another or conversion of the property of another. In this case, Atai first used the source code in 1985. Not until 1988 did Computer Associates file its suit. Atai concedes that Computer Associates did not know about the use of the source code until 1988. The question is whether we should permit the discovery rule exception in these circumstances.

To answer this question, we must understand the objective of statutes of limitations. The purpose of statutes of limitations is to compel the assertion of claims within a reasonable period while the evidence is fresh in the minds of the parties and witnesses. Price v. Estate of Anderson, 522 S.W.2d 690, 692 (Tex.1975); Gaddis v. Smith, 417 S.W.2d 577, 578 (Tex.1967). “Society’s interest in repose is to have disputes either settled or barred within a reasonable time. It is based on the theory that the uncertainty and insecurity caused by unsettled claims hinder the flow of commerce.” Safeway Stores, Inc. v. Certainteed Corp., 710 S.W.2d 544, 545 (Tex.1986). The discovery rule exception defers accrual of a cause of action until the plaintiff knew or, exercising reasonable diligence, should have known of the facts giving rise to the cause of action. Trinity River Auth. v. URS Consultants, 889 S.W.2d 259, 262 (Tex.1994); Moreno v. Sterling Drug, Inc., 787 S.W.2d 348, 351 (Tex.1990). The discovery rule, in application, proves to be a very limited exception to statutes of limitations.

Similar to the discovery rule exception, where fraud is alleged, we have granted the claimant the benefit of deferring the cause of action until the claimant discovered or should have discovered the fraud. Ruebeck v. Hunt, 142 Tex. 167, 176 S.W.2d 738, 739 (1943); Quinn v. Press, 135 Tex. 60, 140 S.W.2d 438, 440 (1940). We have also permitted claimants to receive the benefit of deferring the accrual of a cause of action in cases where the facts forming the basis of an injury were concealed. Estate of Stonecipher v. Estate of Butts, 591 S.W.2d 806, 809 (Tex.1979); Owen v. King, 130 Tex. 614, 111 S.W.2d 695, 697 (1938). In short, we have said: “[F]raud *456vitiates whatever it touches, Morris v. House, 32 Tex. 492 (1870), and ... limitations begin to run from the time the fraud is discovered or could have been discovered by the defrauded party by exercise of reasonable diligence.” Estate of Stonecipher, 591 S.W.2d at 809. Although similar in effect, the discovery rule exception and deferral based on fraud or concealment exist for different reasons. Unlike the discovery rule exception, deferral in the context of fraud or concealment resembles equitable estoppel. “[F]raudulent concealment estops the defendant from relying on the statute of limitations as an affirmative defense to [the] plaintiffs claim.” Borderlon v. Peck, 661 S.W.2d 907, 908 (Tex.1983).

Unrelated to fraud or concealment, this Court has permitted the discovery rule exception in certain limited circumstances. From these cases, we can glean a unifying principle which facilitates balancing those factors that are considered before this Court has permitted application of the discovery rule exception to the statute of limitations. Generally, application has been permitted in those cases where the nature of the injury incurred is inherently undiseoverable and the evidence of injury is objectively verifiable. The requirement of inherent undiscoverability recognizes that the discovery rule exception should be permitted only in circumstances where “it is difficult for the injured party to learn of the negligent act or omission.” Willis v. Maverick, 760 S.W.2d 642, 645 (Tex.1988); see Kelley v. Rinkle, 532 S.W.2d 947, 949 (Tex.1976) (“A person will not ordinarily have any reason to suspect that he has been defamed by the publication of a false credit report to a credit agency until he makes application for credit....”); Hays v. Hall, 488 S.W.2d 412, 414 (Tex.1972) (“One who undergoes a vasectomy operation, and then after tests is told that he is sterile, cannot know that he is still fertile, if that be the case, until either his wife becomes pregnant or he is shown to be fertile by further testing.”); Gaddis, 417 S.W.2d at 580 (“It is a virtual certainty that the patient has no knowledge on the day following surgery — nor for a long time thereafter — that a foreign object was left in the incision.”).

We note that, in Willis, we also relied on the fiduciary relationship obligating an attorney to “render a full and fair disclosure of facts material to the client’s representation.” 760 S.W.2d at 645. We stated that “[f]acts which might ordinarily require investigation likely may not excite suspicion where a fiduciary relationship is involved,” and that absent the discovery rule exception the client “would have to hire a second attorney to observe the work of the first.” Id. at 645-46. But the fiduciary rationale is, in reality, a variation on the inherently undiseoverable element. Fiduciaries are presumed to possess superior knowledge, meaning the injured party, the client, is presumed to possess less information than the fiduciary. Consequently, in the fiduciary context, it may be said that the nature of the injury is presumed to be inherently undiseoverable, although a person owed a fiduciary duty has some responsibility to ascertain when an injury occurs. Courseview, Inc. v. Phillips Petroleum Co., 158 Tex. 397, 312 S.W.2d 197, 205 (Tex.1957).

We now consider whether to permit the application of the discovery rule exception to the statute of limitations in misappropriation of trade secret cases in light of the unifying principle: (1) whether the injury is inherently undiseoverable; and (2) whether evidence of the injury is objectively verifiable. Computer Associates asserts that it could not have reasonably discovered the theft of its trade secret within two years after Altai’s first actual use. Therefore, Computer Associates argues that its cause of action was inherently undiseoverable. It bolsters its contention by focusing on Altai’s oral stipulation which states that Computer Associates did not know or have reason to know of Altai’s alleged misappropriation within the two-year statute of limitations. Computer Associates misconstrues the meaning of “inherently undiseoverable.” Inherently undis-coverable encompasses the requirement that the existence of the injury is not ordinarily discoverable, even though due diligence has been used.

In this age of technological innovation, intellectual property, including trade secrets, is a jealously guarded commodity. Extensive *457precautions are taken to prevent and eliminate trade secret misappropriations, the intention being to promote the rapid detection of misappropriation. See generally 3 Milgrim ON TRADE SECRETS § 13.04[2] (1993) (identifying only forty-three out of a reported six-thousand trade secret misappropriation cases — less than one percent — that implicated the statute of limitations). While some trade secret misappropriations might not be quickly discovered, this isolated fact does not alter the reality that, in most cases, trade secret misappropriation generally is capable of detection -within the time allotted for bringing such suits. Id. In the present case, Arney misappropriated Computer Associates’ source code by absconding with a stack of “greenbar” computer paper upon which the code was printed when he left the company. As Altai points out, Computer Associates could have detected Arney’s theft by using document control logs — a common practice among high technology companies dependent upon trade secrets. See generally DoRR & Munch, Protecting Trade Secrets, Patents, Copyrights, and Trademarks §§ 1.9-1.17, at 11-29 (1990); The Law and Business of Computer Software § 3.05[b], at 3-12 to 3-17, §§ 4.00-.09, at 4-1 to 4-21 (D.C. Toedt ed., 1990); Epstein, Modern Intellectual Property 212.1-216 (1988); Pooley, Trade Secrets: How to PROTECT Your Ideas and Assets (1982).

Additionally, we live in a world of high employee mobility and easy transportability of information. Under these circumstances, it is not unexpected that a former employee will go to work for a competitor and that the competitor might thereby acquire trade secrets. Dorr & Munch, at xiii; Epstein, at 84-84.1; Seidel, What the General Practitioner Should Know about Trade Secrets and Employment Agreements § 3.01, at 21-22 (2d ed. 1984). This reality is reflected in the confidentiality agreement requested of and exit interview given to Arney by Computer Associates. Vigilance in the area of trade secrets is required, particularly because once a trade secret is made public all ownership is lost. See, e.g., Luccous v. J.C. Kinley Co., 376 S.W.2d 336, 338 (Tex.1964) (“It is self-evident that the subject matter of a trade secret must be kept secret.”); see also Schalk v. State, 823 S.W.2d 633, 638-44 (Tex.Crim.App.1991), cert. denied, 503 U.S. 1006, 112 S.Ct. 1763, 118 L.Ed.2d 425 (1992) (discussing various measures taken to insure secrecy). High employee mobility and critical interest in maintaining a proprietary interest in a trade secret are endemic to the computer software industry. Suspicions should abound when a competitor markets a product similar to that previously developed by a former employer after one of the former employer’s employees begins work for the competitor.

In the past, this Court has noted the “shocking results” of barring a plaintiffs suit before the injury has even been discovered. See Gaddis, 417 S.W.2d at 581; Hays, 488 S.W.2d at 414. This rationale offers no principled basis for distinguishing between cases in which the rule applies and those in which it does not. If carried to its logical conclusion, the rationale mandates applying the discovery rule exception to every case, thus eviscerating the whole notion of an absolute time bar to litigation. It is clear that “preclusion of a legal remedy alone is not enough to justify a judicial exception to the statute. The primary purpose of limitations, to prevent litigation of stale or fraudulent claims, must be kept in mind.” Robinson v. Weaver, 550 S.W.2d 18, 20 (Tex.1977). The discovery rule is a limited exception to strict compliance with the statute of limitations. See Trinity River Auth., 889 S.W.2d at 262. Because misappropriation of trade secrets is not a cause of action that is inherently undiscov-erable, permitting application of the discovery rule exception in these cases would do no more than permit the litigation of stale claims.

Before concluding, we note that the principle we have applied has two elements. The second element is that the alleged injury be objectively verifiable. See e.g., Gaddis, 417 S.W.2d at 581 (“[Tjhis is a peculiar type of case [leaving sponge in patient] which is not particularly susceptible to fraudulent prosecution.”). Requiring objective verifiability assures that the policy underpinnings of statutes of limitations are met — balancing the possibility of stale or fraudulent claims against individual injustice. We relied heavi*458ly on this element in deciding not to extend the discovery rule exception to cases of medical misdiagnoses: “Unlike Gaddis v. Smith there exists in the present case no physical evidence which in-and-of-itself establishes the negligence of some person.” Robinson, 550 S.W.2d at 21. And we considered this element in at least one other case. See Kelley, 532 S.W.2d at 949. Because trade secret misappropriation claims are not inherently undiscoverable, it is not necessary for us to reach the question of whether such misappropriations may be verified by objective evidence.2

Finally, we recognize that thirty-nine states and the District of Columbia have adopted the Uniform Trade Secrets Act, which provides a discovery rule exception that tolls the statute of limitations in misappropriation of trade secret cases. See Unif. TRADE SECRETS Act § 6,14 U.L.A. 462 (1990) (citing laws of Alabama, Alaska, Arizona, Arkansas, California, Colorado, Connecticut, Delaware, District of Colombia, Florida, Georgia, Hawaii, Idaho, Illinois, Indiana, Iowa, Kansas, Kentucky, Louisiana, Maine, Maryland, Minnesota, Mississippi, Montana, Nebraska, Nevada, New Hampshire, New Mexico, North Carolina, North Dakota, Oklahoma, Oregon, Rhode Island, South Carolina, South Dakota, Utah, Virginia, Washington, West Virginia and Wisconsin). No state supreme court, however, has yet adopted the discovery rule exception for trade secret cases as an exercise of its common law jurisdiction. See Pilkington Bros., P.L.C. v. Guardian Indus. Corp., 230 U.S.P.Q. (BNA) 300, 301, 1986 WL 9876 (E.D.Mich.1986) (holding that determination of applicability of discovery rule exception is a question of federal common law); Anaconda Co. v. Metric Tool, 485 F.Supp. 410, 425 (E.D.Pa.1980) (noting no reported Pennsylvania cases). We decline to be the first. We will not overlay section 16.003(a) with the discovery rule exception for misappropriation of trade secret causes of action.

Ill

Having determined that the discovery rule exception does not apply in this case, we must address Computer Associates’ challenge to the two-year statute of limitations period provided by section 16.003(a) as violating the “open courts” provision of Article I, Section 13 of the Texas Constitution. It does not. The “open courts” provision guarantees that “meaningful remedies must be afforded, ‘so that the legislature may not abrogate the right to assert a well-established common law cause of action unless the reason for its action outweighs the litigants’ constitutional right of redress.’ ” Trinity River Auth., 889 S.W.2d at 261 (quoting Texas Ass’n of Business v. Texas Air Control Bd., 852 S.W.2d 440, 448 (Tex.1993)). Therefore, a litigant alleging a violation of this guarantee must first show the abrogation of a cognizable common law cause of action and second, that the restriction is unreasonable or arbitrary when balanced against the purpose and basis of the statute. Sax v. Votteler, 648 S.W.2d 661, 666 (Tex.1983).

The traditional rule in Texas is that a cause of action accrues and the two-year limitations period begins to run as soon as the owner suffers some injury, regardless of when the injury becomes discoverable. Trinity River Auth., 889 S.W.2d at 262. Failure to permit a discovery rule exception to this traditional rule does not abrogate a common law right. Id. Even assuming that section 16.003(a) is legislative action which affects a common law cause of action for purposes of “open courts” analysis, application of a two-year statute of limitations without a discovery rule exception is not unreasonable or arbitrary when balanced against the purpose of the statute. As we explained above, a significant purpose of statutes of limitations is to prevent the litigation of stale and fraudulent claims. This policy, combined with the nature of trade secret property rights, which requires an owner to vigilantly guard the secret from the world in order to preserve its rights, makes application of the two-year statute of limitations reasonable under the circumstances.

IV

We answer the questions certified to us as follows: (1) a discovery rule exception to *459section 16.003(a) of the Texas Civil Practice and Remedies Code does not apply to claims for misappropriation of trade secrets; and (2) application of the two-year limitations period provided by section 16.003(a) does not contravene the “open courts” provision of Article I, Section 13 of the Texas Constitution.

APPENDIX

United States Court of Appeals for the Second Circuit

Docket No. 93-7957

Computer Associates International, Inc., Plaintiff-Appellant,

-against-

Altai, Inc., Defendant-Appellee.

Certificate to the Supreme Court of Texas pursuant to Tex.R.App.P. 114(a) (West 1993) (permitting certification of questions of state law for which there is no controlling precedent in the decisions of the Supreme Court of Texas).

In January 1984, Claude F. Arney, III left the Texas offices of his employer of five years, Computer Associates International, Inc. (“CA”), a developer of computer software, to go to work for a competitor, Altai, Inc. (“Altai”). See Computer Assocs. Int’l v. Altai, Inc., 775 F.Supp. 544, 553 (E.D.N.Y.1991) (“Altai P’), aff'd in part, vacated in part, and remanded, 982 F.2d 693 (2d Cir.1992). When Arney left, he took with him copies of the computer source code1 listings for two versions of a CA program known as ADAPTER (a component of another CA program called CA-SCHEDULER), in knowing violation of a CA employment agreement which he had signed that prohibited the retention of such materials. Id. at 552-53. Arney then undertook to write a program for Altai known as OSCAR. The first version of this program, which performs a task similar to CA’s ADAPTER, was designated OSCAR 3.4. Arney copied approximately thirty percent of the source code comprising OSCAR 3.4 from the source code of ADAPTER. Id. at 552. Other than Arney, no one at Altai knew that Arney possessed the ADAPTER source code or that he had copied that code in creating OSCAR 3.4. Id. at 554. From 1985 to August 1988, Altai used OSCAR 3.4 in several of its computer programs that competed with CA-SCHEDULER. Id. at 552, 554.

In late July 1988, CA first learned that Altai might have copied the ADAPTER program. Id. at 554. After confirming its suspicions, CA secured copyrights on versions 2.1 and 7.0 of CA-SCHEDULER. Id. In August 1988, CA brought suit against Altai in federal district court, invoking diversity jurisdiction.2 See 28 U.S.C. § 1332 (1988). CA alleged that Altai had infringed CA’s copyright in ADAPTER and had misappropriated CA’s trade secrets. Altai I, 775 F.Supp. at 554. Upon receiving CA’s complaint, Altai immediately investigated the allegations and learned that Arney had copied portions of the ADAPTER code. Id. James P. Williams, an employee of Altai who became its president on October 31, 1988, reviewed with Arney which portions of OSCAR Arney had copied and which he had developed independently. In conducting this review, Arney again consulted the ADAPTER code, but neither Williams nor any of Altai’s other programmers ever examined the ADAPTER code, which was locked away before Altai began rewriting OSCAR. Id.

After consulting counsel about how to proceed, Williams organized an operation to rewrite the OSCAR program. Working primarily from a different Altai program, Williams wrote descriptions of various services that the rewritten OSCAR would per*460form. Id. Williams then instructed Altai’s programmers, none of whom had worked on OSCAR 3.4, to write appropriate code to obtain those services. In doing so, the programmers were forbidden to contact Arney or to refer to OSCAR 3.4. Id. The resulting program was designated OSCAR 3.5. Upon its completion, Altai shipped OSCAR 3.5 to its new customers, and also provided OSCAR 3.5 as a “free upgrade” to all customers who had purchased OSCAR 3.4. Id.

After a bench trial, the district court found that Altai’s OSCAR 3.4 computer program had infringed the ADAPTER component of CA’s copyrighted computer program CA-SCHEDULER. Id. at 558. The court awarded CA a total of $364,444 in actual damages and apportioned profits on copyright claims stemming from OSCAR 3.4. Id. at 572. The district court also determined, however, that OSCAR 3.5 was neither substantially similar to nor copied from ADAPTER, and accordingly denied CA relief on its copyright claims regarding that version of Altai’s program. Id. at 561-62. Finally, although noting that Texas state law would govern CA’s trade secret misappropriation claim, id. at 566, the district court concluded that CA’s misappropriation claim against Altai had been preempted by the federal Copyright Act. Id. at 563-66. While recognizing that the torts of misappropriation and copyright infringement might be distinct in some cases, the district court reasoned that on the facts of this case, the claims “boil[ed] down to the same thing — a right of action for the unauthorized reproduction of, and preparation of derivative works based on, ADAPTER.” Id. at 564. Because both claims involved a copyrightable work, the district court concluded that “all claims concerning copying [are] governed, exclusively, by federal [copyright] law.” Id. at 565.

Altai did not perfect an appeal from the award of damages with respect to OSCAR 3.4, but CA appealed the denial of relief with respect to OSCAR 3.5. See Computer Assocs. Int’l v. Altai, Inc., 982 F.2d 693, 696-97 (2d Cir.1992) (“Altai II ”). We affirmed the district court’s ruling that OSCAR 3.5 did not infringe CA’s copyright in ADAPTER, id. at 714-15, and also agreed with the district court that Texas law would govern CA’s trade secret misappropriation claims (if they were not preempted by federal law). Id. at 718. As to preemption, we noted that if CA’s misappropriation claims involved an “extra element” that was not included in its copyright claims and rendered the misappropriation claims “qualitatively different” from the copyright claim, the misappropriation claims would not be preempted. See id. at 716. Concluding that such an element, beyond mere copying, might be established on the basis that Altai had constructive notice of misappropriation with respect to OSCAR 3.4, see id. at 718-19, or that Altai’s conceded knowledge of misappropriation when it developed OSCAR 3.5 may have resulted in the embodiment of CA trade secrets therein despite the preventive measures taken by Altai, see id. at 719-21, we remanded for further consideration of CA’s misappropriation claims. Id. at 720-21.

On remand, the parties briefed the trade secret issues under Texas law. Altai also asserted the affirmative defense, first invoked in its answer and also presented in its post-trial memorandum, that CA’s claims were barred by the applicable Texas statute of limitations, Texas Civil Practice & Remedies Code Atm. § 16.003(a) (Vernon 1986).3

*461See Computer Assocs. Int’l v. Altai, Inc., 882 F.Supp. 50, 51-52 (E.D.N.Y.1993) (‘Altai III”). That provision states, in pertinent part:

A person must bring suit for trespass for injury to the estate or to the property of another, conversion of personal property, [or] taking or detaining the personal property of another ... not later than two years after the day the cause of action accrues.

The Texas Court of Civil Appeals (now designated the Court of Appeals) has applied article 5526 of the Texas Civil Statutes, the forerunner of § 16.003, to claims involving the misappropriation of trade secrets. See Reynolds-Southwestern Corp. v. Dresser Indus., 438 S.W.2d 135, 140 (Tex.Civ.App.1969) (applying art. 5526 to trade secret claim); see also Coastal Distrib. Co. v. NGK Spark Plug Co., 779 F.2d 1033, 1038 (5th Cir.1986) (same) (collecting eases).

Under Texas law, the statute of limitations on tort claims generally begins to run when “the wrongful act effects an injury, regardless of when the plaintiff learned of such injury.” Moreno v. Sterling Drug, Inc., 787 S.W.2d 348, 351 (Tex.1990). This general rule of accrual applies unless modified by the “discovery rule,” which, when applicable, provides that the period of limitations for a tort action “run[s] from the date the plaintiff discovers or should have discovered, in the exercise of reasonable care and diligence, the nature of the injury.” Willis v. Maverick, 760 S.W.2d 642, 644 (Tex.1988) (citing Gaddis v. Smith, 417 S.W.2d 577, 578 (Tex.1967)).

The district court declined to address the merits of CA’s misappropriation claims on remand, instead dismissing them as barred under § 16.003(a). Although the district court had earlier opined that CA’s claims for misappropriation of trade secrets would be subject to the discovery rule under Texas law, see Altai I, 775 F.Supp. at 566, after further consideration upon remand, the court declined to apply the discovery rule in this case. The court noted that no Texas court had unambiguously applied the discovery rule to claims involving the misappropriation of trade secrets. Altai III, 832 F.Supp. at 53. Because the Supreme Court of Texas had applied the rule only under limited circumstances, each time emphasizing the special policy considerations warranting application of the rule in the case at hand, the district court “decline[d] to make an unprecedented expansion of the discovery rule by applying it to the facts of this case.” Id. at 54. The court also noted, however, its inability under Texas R.App.P. 114(a) to certify this issue to the Supreme Court of Texas, id. at 53-54, and added: “If this matter comes to the attention of the second circuit, they may feel it appropriate to certify the question to the Texas Supreme Court for an authoritative disposition of the issue.” Id. at 54.

Altai conceded at trial that CA did not know or have reason to know of Altai’s alleged misappropriation until 1988. Thus, if the discovery rule applied in this case, CA’s misappropriation claims would be timely under Texas law. However, the district court concluded that because the discovery rule did not apply, and because the allegedly wrongful act occurred in 1984 (four years before CA filed suit) when Arney began copying CA’s ADAPTER codes into OSCAR 3.4, the statute of limitations had run and CA’s claim was barred. Id. at 52, 54.

On appeal, CA contends that the district court erred in declining to apply the discovery rule in this case. CA argues that: (1) the Texas Court of Civil Appeals applied the discovery rule to a claim arising from the misappropriation of trade secrets in Reynolds-Southwestern Corp., 438 S.W.2d at 140; (2) precedent of the Supreme Court of Texas *462supports extension of the discovery rule in this instance because the misappropriation of trade secrets is inherently difficult for a potential plaintiff to discover, see Willis, 760 S.W.2d at 645 (“This court has adopted the discovery rule in cases ... in which it is difficult for the injured party to learn of the negligent act or omission.”) (collecting cases); (3) the Supreme Court of Texas presumably would follow the lead of the vast majority of states by adopting the discovery rule for trade secrets; and (4) those Texas eases that do not apply the discovery rule are distinguishable because they involved different contexts and policy considerations from those implicated by claims for the misappropriation of trade secrets. See, e.g., Moreno v. Sterling Drug, Inc., 787 S.W.2d at 354-55 (wrongful death); Seibert v. General Motors Corp., 853 S.W.2d 773, 776 (Tex.Ct.App.1993) (physical injuries). Finally, CA also argues that any construction of § 16.003(a) to bar misappropriation claims before they become discoverable would contravene the “open courts” provision of the Texas Constitution, Tex. Const. art. I, § 13 (Vernon 1984), which has been interpreted to prohibit statutes of limitations that require “a person whose injuries are not immediately discoverable” to sue “during the period of undiseoverability.” Nelson v. Krusen, 678 S.W.2d 918, 923 (Tex.1984).

Altai responds that: (1) Reynolds-Southwestern, which is in any event not a disposi-tive statement of Texas law, is distinguishable because, unlike this case, Reynolds-Southwestern involved a claim of fraud that overlapped the claim for misappropriation of trade secrets; (2) the discovery rule has primarily been applied by Texas courts to breaches of fiduciary duty by licensed professionals; (3) its application in other eases has been limited to more inherently undiscovera-ble wrongs than the alleged misappropriation of ADAPTER; and (4) CA’s failure to discover the alleged misappropriation resulted from its own laxity in safeguarding its trade secrets. Altai also points out that the “open courts” constitutional provision has only been invoked in Nelson and other Texas cases to invalidate a specific medical malpractice limitations provision that provided an inflexible two-year limitations period regardless of the discoverability of the injury inflicted by malpractice.

The decision to extend the discovery rule to a type of claim to which it has not previously been applied by the Supreme Court of Texas requires “careful[] eonsider[ation of] opposing policy considerations,” Kelley v. Rinkle, 532 S.W.2d 947, 949 (Tex.1976), weighing concerns regarding “the assertion of claims within a reasonable period,” id., against general considerations of justice. Resolution of this issue thus calls for particular sensitivity to state policy concerns and special expertise in Texas law. Accordingly, we conclude that this decision is best left to the Supreme Court of Texas. Cf. Tennimon v. Bell Helicopter Textron, Inc., 823 F.2d 68, 72 (5th Cir.1987) (per curiam) (“[I]f any court were to hold that the discovery rule is an exception to this seemingly unambiguous declaration of the Texas Legislature, it should be a Texas court in the first instance, not a federal court sitting in a diversity ease”) (discussing § 16.003(b)).

Accordingly, because CA’s claims raise unsettled and significant questions of Texas law that control the outcome of this ease, see 2d Cir.R. § 0.27, we certify the following questions of law to the Supreme Court of Texas:

1. Does the discovery rule exception to § 16.003(a) apply to claims for misappropriation of trade secrets?

2. If not, would the application to such claims of the two-year limitations period provided by § 16.003(a) contravene the “open courts” provision of article I, § 13 of the Texas Constitution?

The foregoing is hereby certified to the Supreme Court of Texas pursuant to Tex.R.App.P. 114(a) as ordered by the United States Court of Appeals for the Second Circuit.

Dated at New York, New York, this__ day of April, 1994.

George Lange, III Clerk, United States Court of Appeals for the Second Circuit By:/s/ Carolyn Clark Campbell Carolyn Clark Campbell Chief Deputy Clerk

. The certified questions are included as an appendix to the Court’s opinion.

. See S.V. v. R.V., - S.W.2d - [1996 WL 112206] (Tex.1996) (discussing the objectively verifiable standard).

. "Source code” refers to "the literal text of a [computer] program's instructions written in a particular programming language,” Gates Rubber Co. v. Bando Chem. Indus., 9 F.3d 823, 835 (10th Cir. 1993), such as COBAL, FORTRAN, BASIC, or EDL. Once the source code has been completed, it is translated or “compiled” into "object code,” which is "the binary language comprised of zeros and ones through which the computer directly receives its instructions." Computer Assocs. Int’l v. Altai, Inc., 982 F.2d 693, 698 (2d Cir. 1992).

. CA is a Delaware corporation with its principal place of business in Garden City, New York. Altai is a Texas corporation, and has its principal place of business in Arlington, Texas. See Altai I, 775 F.Supp. at 549.

. A federal court sitting in diversity applies the choice of law rules of the forum state. Klaxon Co. v. Stentor Elec. Mfg. Co., 313 U.S. 487, 496-97, 61 S.Ct 1020, 1021-22, 85 L.Ed. 1477 (1941). CA originally brought this action in the United States District Court for the District of New Jersey, whence it was transferred by stipulation to the Eastern District of New York pursuant to 28 U.S.C. § 1404(a) (1988). See Altai I, 775 F.Supp. at 566. Whether such a transfer is initiated by the plaintiff or the defendant, "the transferee court must follow the choice of law rules that prevailed in the transferor court.” Ferens v. John Deere Co., 494 U.S. 516, 519, 110 S.Ct. 1274, 1277, 108 L.Ed.2d 443 (1990). Thus, New Jersey's choice of law rules determine which state's statute of limitations applies. New Jersey has abandoned the prevalent rule that the statute of limitations is a procedural matter and therefore the forum state’s statute of limitations is applicable, regardless of which state's substantive law applies. See Heavner v. Uniroyal, Inc., 63 N.J. 130, 305 A.2d 412, 415-18 (1973); Warner v. Auberge Gray Rocks Inn, Ltee., 827 F.2d 938, 940-41 (3d Cir.1987). Instead, New Jersey applies the interest-based test normally reserved for determining which state’s substantive law *461will apply. See Schum v. Bailey, 578 F.2d 493, 495 (3d Cir.1978). In this case, the acts at issue occurred in Texas and involved the Texas offices of a Delaware corporation (CA) and a Texas corporation (Altai). Under these circumstances, we have little difficulty in concluding that New Jersey has no significant interest in applying its own statute of limitations, and would instead have recourse to the Texas statute of limitations. See Seals v. Langston Co., 206 N.J.Super. 408, 502 A.2d 1185, 1187-88 (Ct.App.Div.), certification denied, 104 N.J. 386, 517 A.2d 392 (1986); Washington v. Systems Maintenance Corp., 260 N.J.Super. 505, 616 A.2d 1352, 1354-56 (Ct.Law Div.1992).