This is a suit for misappropriation of trade secrets. Rockwell Graphic Systems, a manufacturer of printing presses used by newspapers, and of parts for those presses, brought the suit against DEV Industries, a competing manufacturer, and against the president of DEV, who used to be employed by Rockwell. The case is in federal court by virtue of the RICO ("Racketeer Influenced and Corrupt Organizations") statute. 18 U.S.C. Secs. 1961 et seq. The predicate acts required for liability under RICO are acts of misappropriation (and related misconduct, such as alleged breaches of fiduciary duty) committed by the individual defendant, Fleck, and by another former employee of Rockwell and present employee of DEV, Peloso. These acts are alleged to violate Illinois law, and in pendent counts Rockwell seeks to impose liability for them directly under that law as well as indirectly under RICO. The district judge granted summary judgment for the defendants upon the recommendation of a magistrate who concluded that Rockwell had no trade secrets because it had failed to take reasonable precautions to maintain secrecy. Therefore there had been no misappropriation, which in turn was the foundation for the predicate acts; so the RICO count had to be dismissed. With the federal claim out of the case, the district judge relinquished jurisdiction over the pendent counts, resulting in a dismissal of the entire case. 730 F. Supp. 171 (N.D.Ill.1990).

When we said that Rockwell manufactures both printing presses and replacement parts for its presses--"wear parts" or "piece parts," they are called--we were speaking approximately. Rockwell does not always manufacture the parts itself. Sometimes when an owner of one of Rockwell's presses needs a particular part, or when Rockwell anticipates demand for the part, it will subcontract the manufacture of it to an independent machine shop, called a "vendor" by the parties. When it does this it must give the vendor a "piece part drawing" indicating materials, dimensions, tolerances, and methods of manufacture. Without that information the vendor could not manufacture the part. Rockwell has not tried to patent the piece parts. It believes that the purchaser cannot, either by inspection or by "reverse engineering" (taking something apart in an effort to figure out how it was made), discover how to manufacture the part; to do that you need the piece part drawing, which contains much information concerning methods of manufacture, alloys, tolerances, etc. that cannot be gleaned from the part itself. So Rockwell tries--whether hard enough is the central issue in the case--to keep the piece part drawings secret, though not of course from the vendors; they could not manufacture the parts for Rockwell without the drawings. DEV points out that some of the parts are for presses that Rockwell no longer manufactures. But as long as the presses are in service--which can be a very long time--there is a demand for replacement parts.

Rockwell employed Fleck and Peloso in responsible positions that gave them access to piece part drawings. Fleck left Rockwell in 1975 and three years later joined DEV as its president. Peloso joined DEV the following year after being fired by Rockwell when a security guard caught him removing piece part drawings from Rockwell's plant. This suit was brought in 1984, and pretrial discovery by Rockwell turned up 600 piece part drawings in DEV's possession, of which 100 were Rockwell's. DEV claimed to have obtained them lawfully, either from customers of Rockwell or from Rockwell vendors, contrary to Rockwell's claim that either Fleck and Peloso stole them when they were employed by it or DEV obtained them in some other unlawful manner, perhaps from a vendor who violated his confidentiality agreement with Rockwell. Thus far in the litigation DEV has not been able to show which customers or vendors lawfully supplied it with Rockwell's piece part drawings.

The defendants persuaded the magistrate and the district judge that the piece part drawings weren't really trade secrets at all, because Rockwell made only perfunctory efforts to keep them secret. Not only were there thousands of drawings in the hands of the vendors; there were thousands more in the hands of owners of Rockwell presses, the customers for piece parts. The drawings held by customers, however, are not relevant. They are not piece part drawings, but assembly drawings. (One piece part drawing in the record is labeled "assembly," but as it contains dimensions, tolerances, and other specifications it is really a piece part drawing, despite the label.) An assembly drawing shows how the parts of a printing press fit together for installation and also how to integrate the press with the printer's other equipment. Whenever Rockwell sells a printing press it gives the buyer assembly drawings as well. These are the equivalent of instructions for assembling a piece of furniture. Rockwell does not claim that they contain trade secrets. It admits having supplied a few piece part drawings to customers, but they were piece part drawings of obsolete parts that Rockwell has no interest in manufacturing and of a safety device that was not part of the press as originally delivered but that its customers were clamoring for; more to the point, none of these drawings is among those that Rockwell claims DEV misappropriated.

The distinction between assembly and piece part drawings is not esoteric. A.H. Emery Co. v. Marcan Products Corp., 268 F. Supp. 289, 300 (S.D.N.Y.1967), aff'd, 389 F.2d 11, 16 (2d Cir.1968), marks it, and along with other cases declares--what is anyway obvious--that a firm's act in making public some of its documents (or part of a document) does not destroy the status as trade secrets of information contained in other documents (or another part of the same document). Alexander & Alexander, Inc. v. Drayton, 378 F. Supp. 824, 833 (E.D.Pa.1974), aff'd without opinion, 505 F.2d 729 (3d Cir.1974); Ecolaire Inc. v. Crissman, 542 F. Supp. 196, 206 (E.D.Pa.1982); Laser Industries, Ltd. v. Eder Instrument Co., 573 F. Supp. 987, 991 (N.D.Ill.1983). It is immaterial that Rockwell affixed the same legend enjoining the user to confidentiality to its assembly drawings as it did to its piece part drawings. Perhaps thinking of the doctrine of patent misuse (on which see USM Corp. v. SPS Technologies, Inc., 694 F.2d 505, 510-12 (7th Cir.1982), and cases cited there), DEV suggests that if a firm claims trade secret protection for information that is not really secret, the firm forfeits trade secret protection of information that is secret. There is no such doctrine--even the patent misuse doctrine does not decree forfeiture of the patent as the sanction for misuse--and it would make no sense. This is not only because there are any number of innocent explanations for Rockwell's action in "overclaiming" trade secret protection (if that is what it was doing)--such as an excess of caution, uncertainty as to the scope of trade secret protection, concern that clerical personnel will not always be able to distinguish between assembly and piece part drawings at a glance, and the sheer economy of a uniform policy--but also because it would place the owner of trade secrets on the razor's edge. If he stamped "confidential" on every document in sight, he would run afoul of what we are calling (without endorsing) the misuse doctrine. But if he did not stamp confidential on every document he would lay himself open to an accusation that he was sloppy about maintaining secrecy--and in fact DEV's main argument is that Rockwell was impermissibly sloppy in its efforts to keep the piece part drawings secret.

On this, the critical, issue, the record shows the following. (Because summary judgment was granted to DEV, we must construe the facts as favorably to Rockwell as is reasonable to do.) Rockwell keeps all its engineering drawings, including both piece part and assembly drawings, in a vault. Access not only to the vault, but also to the building in which it is located, is limited to authorized employees who display identification. These are mainly engineers, of whom Rockwell employs 200. They are required to sign agreements not to disseminate the drawings, or disclose their contents, other than as authorized by the company. An authorized employee who needs a drawing must sign it out from the vault and return it when he has finished with it. But he is permitted to make copies, which he is to destroy when he no longer needs them in his work. The only outsiders allowed to see piece part drawings are the vendors (who are given copies, not originals). They too are required to sign confidentiality agreements, and in addition each drawing is stamped with a legend stating that it contains proprietary material. Vendors, like Rockwell's own engineers, are allowed to make copies for internal working purposes, and although the confidentiality agreement that they sign requires the vendor to return the drawing when the order has been filled, Rockwell does not enforce this requirement. The rationale for not enforcing it is that the vendor will need the drawing if Rockwell reorders the part. Rockwell even permits unsuccessful bidders for a piece part contract to keep the drawings, on the theory that the high bidder this round may be the low bidder the next. But it does consider the ethical standards of a machine shop before making it a vendor, and so far as appears no shop has ever abused the confidence reposed in it.

The mere fact that Rockwell gave piece part drawings to vendors--that is, disclosed its trade secrets to "a limited number of outsiders for a particular purpose"--did not forfeit trade secret protection. A.H. Emery Co. v. Marcan Products Corp., 389 F.2d 11, 16 (2d Cir.1968). On the contrary, such disclosure, which is often necessary to the efficient exploitation of a trade secret, imposes a duty of confidentiality on the part of the person to whom the disclosure is made. Jones v. Ulrich, 342 Ill.App. 16, 25-26, 95 N.E.2d 113, 117 (1950); Crocan Corp. v. Sheller-Globe Corp., 385 F. Supp. 251, 253 (N.D.Ill.1974). But with 200 engineers checking out piece part drawings and making copies of them to work from, and numerous vendors receiving copies of piece part drawings and copying them, tens of thousands of copies of these drawings are floating around outside Rockwell's vault, and many of these outside the company altogether. Although the magistrate and the district judge based their conclusion that Rockwell had not made adequate efforts to maintain secrecy in part at least on the irrelevant fact that it took no measures at all to keep its assembly drawings secret, DEV in defending the judgment that it obtained in the district court argues that Rockwell failed to take adequate measures to keep even the piece part drawings secret. Not only did Rockwell not limit copying of those drawings or insist that copies be returned; it did not segregate the piece part drawings from the assembly drawings and institute more secure procedures for the former. So Rockwell could have done more to maintain the confidentiality of its piece part drawings than it did, and we must decide whether its failure to do more was so plain a breach of the obligation of a trade secret owner to make reasonable efforts to maintain secrecy as to justify the entry of summary judgment for the defendants.

The requirement of reasonable efforts has both evidentiary and remedial significance, and this regardless of which of the two different conceptions of trade secret protection prevails. (Both conceptions have footholds in Illinois law, as we shall see.) The first and more common merely gives a remedy to a firm deprived of a competitively valuable secret as the result of an independent legal wrong, which might be conversion or other trespass or the breach of an employment contract or of a confidentiality agreement. Under this approach, because the secret must be taken by improper means for the taking to give rise to liability, Ill.Rev.Stat. ch. 140, pp 352(a), (b)(1), (2)(A), (B); Restatement of Torts Sec. 757 (1939); ILG Industries, Inc. v. Scott, 49 Ill. 2d 88, 93, 273 N.E.2d 393, 396 (1971); Brunswick Corp. v. Outboard Marine Corp., 79 Ill. 2d 475, 479, 38 Ill. Dec. 781, 783, 404 N.E.2d 205, 207 (1980), the only significance of trade secrecy is that it allows the victim of wrongful appropriation to obtain damages based on the competitive value of the information taken. The second conception of trade secrecy, illustrated by E.I. duPont de Nemours & Co. v. Christopher, 431 F.2d 1012 (5th Cir.1970), and in Illinois by Ill.Rev.Stat. ch. 140, p 352(b)(2)(C), and Schulenburg v. Signatrol, Inc., 33 Ill. 2d 379, 387-88, 212 N.E.2d 865, 869 (1965), is that "trade secret" picks out a class of socially valuable information that the law should protect even against nontrespassory or other lawful conduct--in Christopher, photographing a competitor's roofless plant from the air while not flying directly overhead and hence not trespassing or committing any other wrong independent of the appropriation of the trade secret itself. See also Brunswick Corp. v. Outboard Marine Corp., supra, 79 Ill.2d at 479, 38 Ill. Dec. at 783, 404 N.E.2d at 207; Restatement, supra, Sec. 758(b).

Since, however, the opinion in Christopher describes the means used by the defendant as "improper," 431 F.2d at 1015-17, which is also the key to liability under the first, more conventional conception of trade secret protection, it is unclear how distinct the two conceptions really are. It is not as if Christopher proscribes all efforts to unmask a trade secret. It specifically mentions reverse engineering as a proper means of doing so. Id. at 1015. This difference in treatment is not explained, but it may rest on the twofold idea that reverse engineering involves the use of technical skills that we want to encourage, and that anyone should have the right to take apart and to study a product that he has bought.

It should be apparent that the two different conceptions of trade secret protection are better described as different emphases. The first emphasizes the desirability of deterring efforts that have as their sole purpose and effect the redistribution of wealth from one firm to another. The second emphasizes the desirability of encouraging inventive activity by protecting its fruits from efforts at appropriation that are, indeed, sterile wealth-redistributive--not productive--activities. The approaches differ, if at all, only in that the second does not limit the class of improper means to those that fit a preexisting pigeonhole in the law of tort or contract or fiduciary duty--and it is by no means clear that the first approach assumes a closed class of wrongful acts, either.

Under the first approach, at least if narrowly interpreted so that it does not merge with the second, the plaintiff must prove that the defendant obtained the plaintiff's trade secret by a wrongful act, illustrated here by the alleged acts of Fleck and Peloso in removing piece part drawings from Rockwell's premises without authorization, in violation of their employment contracts and confidentiality agreements, and using them in competition with Rockwell. Rockwell is unable to prove directly that the 100 piece part drawings it got from DEV in discovery were stolen by Fleck and Peloso or obtained by other improper means. But if it can show that the probability that DEV could have obtained them otherwise--that is, without engaging in wrongdoing--is slight, then it will have taken a giant step toward proving what it must prove in order to recover under the first theory of trade secret protection. The greater the precautions that Rockwell took to maintain the secrecy of the piece part drawings, the lower the probability that DEV obtained them properly and the higher the probability that it obtained them through a wrongful act; the owner had taken pains to prevent them from being obtained otherwise.

Under the second theory of trade secret protection, the owner's precautions still have evidentiary significance, but now primarily as evidence that the secret has real value. For the precise means by which the defendant acquired it is less important under the second theory, though not completely unimportant; remember that even the second theory allows the unmasking of a trade secret by some means, such as reverse engineering. If Rockwell expended only paltry resources on preventing its piece part drawings from falling into the hands of competitors such as DEV, why should the law, whose machinery is far from costless, bother to provide Rockwell with a remedy? The information contained in the drawings cannot have been worth much if Rockwell did not think it worthwhile to make serious efforts to keep the information secret.

The remedial significance of such efforts lies in the fact that if the plaintiff has allowed his trade secret to fall into the public domain, he would enjoy a windfall if permitted to recover damages merely because the defendant took the secret from him, rather than from the public domain as it could have done with impunity. Brunswick Corp. v. Outboard Marine Corp., supra, 79 Ill.2d at 479, 38 Ill. Dec. at 783, 404 N.E.2d at 207; Van Products Co. v. General Welding & Fabricating Co., 419 Pa. 248, 267-68, 213 A.2d 769, 779-80 (1965) (repudiating the interpretation of Pennsylvania law that this court had adopted in Smith v. Dravo Corp., 203 F.2d 369, 374-75 (7th Cir.1953)). It would be like punishing a person for stealing property that he believes is owned by another but that actually is abandoned property. If it were true, as apparently it is not, that Rockwell had given the piece part drawings at issue to customers, and it had done so without requiring the customers to hold them in confidence, DEV could have obtained the drawings from the customers without committing any wrong. The harm to Rockwell would have been the same as if DEV had stolen the drawings from it, but it would have had no remedy, having parted with its rights to the trade secret. This is true whether the trade secret is regarded as property protected only against wrongdoers or (the logical extreme of the second conception, although no case--not even Christopher--has yet embraced it and the patent statute might preempt it) as property protected against the world. In the first case, a defendant is perfectly entitled to obtain the property by lawful conduct if he can, and he can if the property is in the hands of persons who themselves committed no wrong to get it. In the second case the defendant is perfectly entitled to obtain the property if the plaintiff has abandoned it by giving it away without restrictions.

It is easy to understand therefore why the law of trade secrets requires a plaintiff to show that he took reasonable precautions to keep the secret a secret. If analogies are needed, one that springs to mind is the duty of the holder of a trademark to take reasonable efforts to police infringements of his mark, failing which the mark is likely to be deemed abandoned, or to become generic or descriptive (and in either event be unprotectable). 1 McCarthy, Trademarks and Unfair Competition Sec. 17:50, at pp. 778-80 (2d ed. 1984). The trademark owner who fails to police his mark both shows that he doesn't really value it very much and creates a situation in which an infringer may have been unaware that he was using a proprietary mark because the mark had drifted into the public domain, much as DEV contends Rockwell's piece part drawings have done.

But only in an extreme case can what is a "reasonable" precaution be determined on a motion for summary judgment, because the answer depends on a balancing of costs and benefits that will vary from case to case and so require estimation and measurement by persons knowledgeable in the particular field of endeavor involved. On the one hand, the more the owner of the trade secret spends on preventing the secret from leaking out, the more he demonstrates that the secret has real value deserving of legal protection, that he really was hurt as a result of the misappropriation of it, and that there really was misappropriation. On the other hand, the more he spends, the higher his costs. The costs can be indirect as well as direct. The more Rockwell restricts access to its drawings, either by its engineers or by the vendors, the harder it will be for either group to do the work expected of it. Suppose Rockwell forbids any copying of its drawings. Then a team of engineers would have to share a single drawing, perhaps by passing it around or by working in the same room, huddled over the drawing. And how would a vendor be able to make a piece part--would Rockwell have to bring all that work in house? Such reconfigurations of patterns of work and production are far from costless; and therefore perfect security is not optimum security.

There are contested factual issues here, bearing in mind that what is reasonable is itself a fact for purposes of Rule 56 of the civil rules. Cooter & Gell v. Hartmarx Corp., --- U.S. ----, 110 S. Ct. 2447, 2459, 110 L. Ed. 2d 359 (1990); Mucha v. King, 792 F.2d 602, 605 (7th Cir.1986); Nunez v. Superior Oil Co., 572 F.2d 1119, 1126 (5th Cir.1978). Obviously Rockwell took some precautions, both physical (the vault security, the security guards--one of whom apprehended Peloso in flagrante delicto ) and contractual, to maintain the confidentiality of its piece part drawings. Obviously it could have taken more precautions. But at a cost, and the question is whether the additional benefit in security would have exceeded that cost. We do not suggest that the question can be answered with the same precision with which it can be posed, but neither can we say that no reasonable jury could find that Rockwell had done enough and could then go on to infer misappropriation from a combination of the precautions Rockwell took and DEV's inability to establish the existence of a lawful source of the Rockwell piece part drawings in its possession.

This is an important case because trade secret protection is an important part of intellectual property, a form of property that is of growing importance to the competitiveness of American industry. Patent protection is at once costly and temporary, and therefore cannot be regarded as a perfect substitute. If trade secrets are protected only if their owners take extravagant, productivity-impairing measures to maintain their secrecy, the incentive to invest resources in discovering more efficient methods of production will be reduced, and with it the amount of invention. And given the importance of the case we must record our concern at the brevity of the district court's opinion granting summary judgment (one and a half printed pages). Brevity is the soul of wit, and all that, and the district judge did have the benefit of a magistrate's opinion; but it is vital that commercial litigation not appear to be treated as a stepchild in the federal courts. The future of the nation depends in no small part on the efficiency of industry, and the efficiency of industry depends in no small part on the protection of intellectual property.

The judgment is reversed and the case remanded to the district court for further proceedings consistent with this opinion (including reinstatement of the pendent counts).

REVERSED AND REMANDED.