*472Mr. Chief Justice Burger

delivered the opinion of the Court.

We granted certiorari to resolve a question on which there is a conflict in the courts of appeals: whether state trade Secret protection is pre-empted by operation of the federal patent law.1 In the instant case the Court of Appeals for the Sixth Circuit held that there was preemption.2 . The Courts of Appeals for the Second, Fourth, Fifth, and ..Ninth Circuits have reached the opposite conclusion.3

*473I

Harshaw Chemical Co., an unincorporated division of petitioner, is a leading manufacturer of a type of synthetic crystal which is useful in the detection of ionizing radiation. In 1949 Harshaw commenced research into the growth of this type crystal and was able to produce one less than two inches in diameter. By 1966, as the result of expenditures in excess of $1 million, Harshaw was able to grow a 17-inch crystal, something no one else had done previously. Harshaw hád developed many processes, procedures, and manufacturing techniques in the purification of raw materials and the growth and encapsulation of the crystals which enabled it to accomplish this feat. Some of these processes Harshaw considers to be trade secrets.

The individual respondents ape former employees of Harshaw who formed or later joined respondent Bicron. While at Harshaw the individual respondents executed, as a condition of employment, at least one agreement each, requiring them not to disclose confidential information or trade secrets obtained as employees of Harshaw. Bicron was formed in August 1969 to compete with Harshaw in the production of the crystals; and by April 1970, had grown a 17-inch crystal.

Petitioner brought this diversity action in United States District Court for the Northern District of Ohio seeking injunctive relief and damages for the misappropriation of trade secrets. The District Court, applying Ohio trade secret law, granted a permanent injunction against the disclosure or use by respondents of 20 of the 40 claimed trade secrets until such time as the trade secrets had *474been released to the public, had otherwise generally become available to the public, or had been obtained by respondents from sources having the legal right to convey the information.

The Court of Appeals for the Sixth Circuit held that the findings of fact by the District Court were not clearly erroneous, and that it was evident from the record that the individual respondents appropriated to the benefit of Bicron secret information on processes obtained while they were employees at Harshaw. Further, the Court of Appeals held that the District Court properly applied Ohio law relating to trade secrets. Nevertheless, the Court of Appeals reversed the District Court, finding Ohio’s trade secret law to be in conflict with the patent •laws of the United States. The Court of Appeals reasoned that Ohio could not grant monopoly protection to processes and manufacturing techniques that were appropriate subjects for consideration under 35 U. S. C. § 101 for a federal patent but which had been in commercial use for over one year and so were no longer eligible for patent protection under 35 U. S. C. § 102 (b).

We hold that Ohio’s law of trade secrets is not preempted by the patent laws of the United States, and, accordingly, we reverse.

n

Ohio has adopted the widely relied-upon definition of a trade secret found at Restatement of Torts § 757, comment b (1939). B. F. Goodrich Co. v. Wohlgemuth, 117 Ohio App. 493, 498, 192 N. E. 2d 99, 104 (1963); W. R. Grace & Co. v. Hargadine, 392 F. 2d 9, 14 (CA6 1968). According to the Restatement,

“ [a] trade secret may consist of any formula, pattern, device or compilation of information which is used in one’s business, and which gives him an opportunity to obtain an advantage over competitors who do not *475know or use it. ft may be a formula for a chemical compound, a process of manufacturing, treating or preserving materials, a pattern for a- machine or other device, or a list of customers.”

The subject of a trade secret must be secret, and must not be of public knowledge or of a general knowledge in the trade or business. B. F. Goodrich Co. v. Wohlgemuth, supra, at 499, 192 N. E. 2d, at 104; National Tube Co. v. Eastern Tube Co., 3 Ohio C. C. R. (n. s.) 459, 462 (1902), aff’d, 69 Ohio St. 560, 70 N. E. 1127 (1903). This necessary element of secrecy is not lost, however, if the holder of the trade secret reveals the trade secret to another “in confidence, and under an implied obligation not .to use or disclose it.” Cincinnati Bell Foundry Co. v. Dodds, 10 Ohio Dec. Reprint 154, 156, 19 Weekly L. Bull. 84 (Super. Ct. 1887). These others may include those of the holder’s “employees to whom it is necessary to confide it, in order to apply it to the uses for which it is intended.” National Tube Co. v. Eastern Tube Co., supra, at 462. Often the recipient of confidential knowledge of the subject of a trade secret is a licensee of its holder. See Lear, Inc. v. Adkins, 395 U. S. 653 (1969).

The protection accorded the trade secret holder is against the .disclosure or unauthorized use of the trade secret by those to whom' the secret has been confided under the express or implied restriction of nondisclosure or nonuse.4 The law also protects the holder of a trade *476secret against disclosure or use .when the knowledge is gained, not by the owner’s volition, but by some “improper means,” Restatement of Torts § 757 (a),, which may include theft, wiretapping, or even aerial reconnaissance.5 A trade secret law, however, does not offer protection against discovery by fair and honest means, such as by independent invention, accidental disclosure, or by' so-called reverse engineering, that is by starting with the known product and working backward to divine the process' which .aided in its development or manufacture.6

Novelty, in the patent law sense, is not required for a trade secret, W. R. Grace & Co. v. Hargadine, 392 F. 2d, at 14. “Quite clearly discovery is something less than invention.” A. O. Smith Corp. v. Petroleum Iron Works Co., 73 F. 2d 531, 538 (CA6 1934), modified to increase scope of injunction, 74 F. 2d 934 (1935). However, some novelty will be required if merely because that which does not possess novelty is usually known; secrecy, in the context of trade' secrets, thus implies at least minimal novelty.7

The subject matter of a patent is limited to a “process, machine, manufacture, or composition of matter,, or ... . improvement thereof,” 35 U. "S. C. § 101, which fulfills the three conditions of novelty and utility as articulated and defined in 35 U. S. C. §§ 101 and 102, and nonobvi*477ousness, as set out in 35 U. S. C. § 103.8 If an invention xueets the rigorous statutory tests for the issuance of a patent, the patent is granted, for a period of 17 *478years, giving what has been described as the “right of exclusion,” R. Ellis, Patent Assignments and Licenses § 4, p. 7 (2d ed. 1943).9 This protection goes not only to copying the subject matter, which is forbidden under the Copyright Act, 17 U. S, C. § 1 et seq., but also to independent creation.

Ill

The first issue we deal with is whether the States are forbidden to act at all in the area of protection of the kinds of intellectual property which may make up the subject matter of trade secrets.

' Article I, § 8, cl. 8, of the Constitution grants to the Congress the power ■

“[t]o promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries . . . .”

In the 1972 Term, in Goldstein v. California, 412 U. S. 546 (1973), we held that the cl. 8 grant of power to Congress was not exclusive and that, at least in the case of writings, the States were not prohibited from encouraging and protecting the efforts of those within their borders by *479appropriate legislation. The States could, therefore, protect against the unauthorized rerecording for sa,le of performances fixed on records or tapes, even though those performances qualified as “writings” in the constitutional sense and Congress was empowered to legislate regarding such performances and could pre-empt the area if it chose to do so. This determination was premised on the great diversity of interests in .our Nation — the essentially nonuniform character of the. appreciation of intellectual achievements in the various States. Evidence for this came from patents granted by the States in the 18th century. 412 U. S., at 557.

Just as the States may exercise regulatory power over writings so may the States regulate with respect to discoveries. States may hold diverse viewpoints in protecting intellectual property relating to invention as they do in protecting the intellectual property relating to the subject matter of copyright. The only limitation on the States is that in regulating the area of patents and copyrights they do not conflict with the operation of the laws in this area passed by Congress, and it is to that more difficult question, we now turn.

IV

The .-question of whether the trade secret law of Ohio is void under the Supremacy Clause involves a consideration of whether that law “stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress.” Hines v. Davidowitz, 312 U. S. 52, 67 (1941). See Florida Avocado Growers v. Paul, 373 U. S. 132, 141 (1963). We stated in Sears, Roebuck &. Co. v. Stiffel Co., 376 U. S. 225, 229 (1964), that when state law touches upon'the area of federal statutes enacted pursuant to constitutional authority, “it is-'familiar doctrine’ that the federal policy *480‘may not be set at naught, or its benefits denied’ by the state law. Sola Elec. Co. v. Jefferson Elec. Co., 317 U. S. 173, 176 (1942). This is true, of course, even, if the state law is enacted in the exercise of otherwise undoubted state power.”

The laws which the .Court of Appeals in this case held to be in conflict, with the Ohio law.of.trade secrets were the patent laws passed by the Congress in the unchallenged, exercise'of its clear power under Art. I, §8, cl. 8, of the Constitution. The patent law does not explicitly endorse or forbid the operation of trade secret law. However, as we have noted, if the scheme of protection developed by Ohio respecting trade secrets “clashes with the objectives of the federal patent laws,” Sears, Roebuck & Co. v. Stiffel Co., supra, at 231, then the state law must-fall. To determine whether the Ohio' law “clashes” with the federal law it is helpful to examine the objectives of both the patent and trade secret, laws».

. The stated objective of the Constitution in granting the power to Congress to legislate in the area of intellectual property is. to, “promote the Progress of Science and. useful Arts.”. The patent laws promote this progress by offering a right of 'exclusion for a limited period as an incentive* to inventors to risk the often enormous costs in'-terms of time, research, and development. .The productive effort thereby fostered will have a positive effect on societyithrough the introduction of new products, and processes of manufacture into (the economy, and the emanations by why of increased employment and better lives for our citizens. In return, for the right of ex'clusion — this “reward for inventions,” Universal Oil Co. v. Globe Co., 322 U. S. 471, 484 (1944) — the patent laws impose upon the inventor a requirement of disclosure. To insure adequate and full disclosure so that upon the *481expiration of the 17-year period “the . knowledge of the invention enures to the people, who are thus enabled without restriction to practice it and profit by its use,” United States v. Dubilier Condenser Corp., 289 U. S. 178, 187 (1933), the patent laws require10' that the patent application shall include a full and clear description of the invention and “of the manner and process of making and using it” so that any person skilled in the art may make and use the invention. 35 U. S. C. § 112. When a patent is granted and the information contained in it is circulated to the general public and those especially skilled in the trade, such additions to the general store óf knowledge are of such importance to the public weal that the Federal Government is willing to pay the high price of 17 years of exclusive use for its disclosure, which' disclosure, it is assumed, will stimulate ideas and the eventual development.of further significant advances in the art. The Court has also articulated another policy of the patent law: that which is in the public domain cannot be removed therefrom by action of the States.

“[FJederal law requires that all ideas in general circulation be dedicated to the common good unless they are protected by a valid patent.” Lear, Inc. v. Adkins, 395 U. S., at 668.

See also Goldstein v. California, 412 U. S., at 570-571; Sears, Roebuck & Co. v. Stiffel Co., supra; Compco Corp. v. Day-Brite Lighting, Inc., 376 U. S. 234, 237-238 (1964); International News Service v. Associated Press. 248 U. S. 215, 250 (1918) (Brandéis, J., dissenting).

The maintenance of standards of commercial ethics and the encouragement of invention are the broadly stated policies behind trade secret law. “The necessity of good faith and honest, fair dealing, is the very life *482and spirit of the commercial world.” National Tube Co. v. Eastern Tube Co., 3 Ohio C. C. R. (n. s.), at. 462.11 In A. O. Smith Corp. v. Petroleum Iron Works Co., 73 F. 2d, at 539, the Court emphasized that even though a discovery may not be patentable, that does not

“destroy the value of the discovery to one who makes it, or advantage the competitor who by unfair means, or as the beneficiary of a broken faith, obtains thé desired knowledge without himself paying the price in lábor, money, or machines expended by the discoverer.”

In Wexler v. Greenberg, 399 Pa. 569, 578-579, 160 A. 2d 430, 434-435 (1960), the Pennsylvania Supreme Court noted the importance, of trade secret protection to the subsidization of. research and- development and to increased economic efficiency within large companies through the dispersion of responsibilities for creative developments.12

Having now in mind the objectives of both the patent and trade secret law, we turn to an examination of the interaction of these systems of protection of intellectual property-one established by the Congress' and the other by a State — to determine whether , and under7 what circumstances the latter might constitute “too great an encroachment on the federal patept system to be tolerated.” Sears, Roebuck & Co. v. Stiffel Co., 376 U. S., at 232.

As ,we noted earlier,, trade secret law protects items which would not be proper subjects for consideration for patent protection under 35 U. S. C. §101. As dn the' *483case of the recordings in Goldstein v. California, Congress, with respect to nonpatentable subject matter, “has drawn no balance; rather, it has left the area unattended, and no reason exists why the State should not be free to act.” Goldstein v. California, supra, at 570 (footnote omitted).

Since no patent is available for a discovery, however useful, novel, and nonobvious, unless it falls within one of the express categories of patentable subject matter of 35 U. S. C. § 101, the holder of such a discovery would have no reason to apply for a patent whether trade secret' protection existed or not. Abolition of trade secret protection would, therefore, not result in increased disclosure to the public of discoveries in the area of nonpatentable subject matter. Also, it is hard to see how the public would be benefited by disclosure of customer lists or advertising campaigns; in fact, keeping such items secret encourages businesses to initiate new and individualized plans of operation, and constructive competition results. This, in turn, leads to a greater variety of business methods than would otherwise be the case if privately developed marketing and other data were passed illicitly among firms involved in the same enterprise.

Congress has spoken in the area of those discoveries which fall within one of the categories of patentable subject matter of 35 U. S. C. § 101 and which are, therefore, of a nature that would be subject to consideration for a patent. Processes, machines, manufactures, compositions of matter, and improvements thereof, which meet the tests of utility, novelty, and nonobviousness are entitled to be patented, but those which do not, are not. The question remains whether those items which are proper subjects for consideration for a patent may also have available the alternative, protection accorded by trade secret law.

*484Certainly the patent policy of encouraging invention is not disturbed by the existence of another form of incentive to invention. In this respect the two; systems are not and never would be in conflict. Similarly, the policy that matter once in- the public domain must remain in the- public domain is not incompatible with the existence of trade secret protection. By definition a trade secret has not been placed in the public domain.13

The more difficult objective of the patent law to reconcile with trade secret law is that of disclosure, the quid pro quo of the right to exclude. Universal Oil Co. v. Globe Co., 322 U. S., at 484. We' are helped in this stage of the analysis by Judge Henry Friendly’s opinion in Painton & Co. v. Bourns, Inc., 442 F. 2d 216 (CA2 1971). There the Court of Appeals thought it useful, in determining whether inventors will refrain because of the existence of trade secret law from applying for patents, thereby depriving the public from learning of the invention, to distinguish between three categories of trade secrets: '

“(1) the trade secret believed by its owner to constitute a validly patentable invention; (2) the trade secret known to its owner not to be so patentable; and (3) the trade secret whose valid patentability is considered5 dubious.” . Id., at 224.

Trade .secret protection in each of these categories would run against breaches of confidence — the employee and licensee situations — and theft and other forms of industrial espionage.

As to. the trade secret known not to meet "the standards *485of patentability, very little in the way of disclosure would be accomplished by abolishing trade secret' protection. With trade secrets of nonpatentable subject matter, the patent alternative would not reasonably be available to the inventor. “There can be no public interest in stimulating- developers of such [unpatentable] know-how to flood an overburdened Patent Office with applications [for] what they do npt consider patentable.” Ibid. The- mere filing of applications doomed to be turned down by the Patent Office will bring forth no new public knowledge or enlightenment, since under federal statute and regulation patent applications and abandoned patent applications are held by the Patent Office in confidence and are not open to public inspection. 35 U. S. C. § 122; 37 CFR § 1.14 (b). .

Even as the extension of trade secret protection to patentable subject matter that the owner knows will not meet the standards of patentability will not conflict with the patent policy of disclosure, it will Rave a decidedly beneficial effect on society. Trade secret law will encourage invention in areas where patent law does not reach, and will prompt the independent innovator to proceed with the discovery and exploitation of his invention. Competition is fostered and the public is not deprived of the use of valuable, if not quite patentable, invention.14

Even if trade secret protection against the faithless employee were abolished, inventive and exploitive effort in the .area of patentable subject matter that did iiot meet the standards of patentability would continue, although at a reduced level. Alternatively with the effort that remained, however, would come an increase in the amount of self-help that innovative companies *486would employ. Knowledge would be widely dispersed among the employees of those still active in research. Security precautions necessarily would be increased, and salaries and fringe benefits of those few officers or employees who had to know the whole of the secret invention would be fixed in an amount thought sufficient to assure their loyalty.15 Smaller companies would be placed at a distinct economic disadvantage, since the costs of this kind of self-help could be great, and the cost to the public of the use of this invention would be increased. The innovative entrepreneur with limited resources would tend to confine his r ¡Search efforts to himself and those few he felt he could > mst without the ultimate assurance of legal protection against breaches of confidence. As a result, organized scientific and technological research could become fragmented, and society, as a whole, would suffer.

Another problem that would arise if state trade secret protection were precluded is in the area of licensing others to exploit secret processes. The holder óf a trade secret would not likely share his secret with a manufacturer who cannot be placed under binding legal obligation to pay a license fee or to protect the secret. The result would be to hoard rather than disseminate knowledge. Painton & Co. v. Bourns, Inc., 442 F. 2d, at 223. Instead, then,- of licensing others to use his invention and making the most efficient use of existing manufacturing and marketing structures within the industry, the trade secret holder would tend either to limit his utilization of the invention, thereby depriving the public of the maximum benefit of its use, or engage in the time-consuming and economically wasteful enterprise of *487constructing duplicative manufacturing and marketing mechanisms for the exploitation of the invention. The detrimental misallbcation of resources and economic waste that would thus take place if trade secret protection were abolished with respect to employees or licensees cannot be justified by reference to any policy that the federal patent law seeks to advance.

Nothing in the patent law. requires that States refrain from action to prevent industrial espionage. In addition to the increased costs for protection from burglary, wiretapping, bribery, and the other means used to misappropriate trade secrets, there is the inevitable cost to the basic decency of society when one firm steals from another. A most fundamental human right, that of privacy, is threatened when industrial espionage is con-, doned or is made profitable^16 the state interest in denying profit to such illegal ventures is unchallengeable.

The next category of patentable subject matter to deal with is theG invention Whose holder has a legitimate doubt as to its patentability. The risk of eventual patent invalidity by the courts -and the costs associated with that risk may well impel some with a good-faith doubt as to patentability not to take the trouble to seek to obtain and defend patent protection for their discoveries, regardless of the existence of trade secret protection. Trade secret protection would assist those inventors in the more efficient exploitation óf their discoveries and not conflict with the patent l$w. In most cases of genuine doubt as'to patent validity the potential rewards of patent protection are so far superior to those accruing to holders of trade secrets, that the holders of *488such inventions will seek patent protection, ignoring the trade secret route. For those inventors “on the line” as to whether to seek patent protection, the abolition of trade secret protection might encourage some to. apply for a patent who otherwise would not have done so. For some of those so encouraged, no patent will be granted and the result

“will have been an unnecessary postponement in the divulging of the trade secret to persons, willing to pay for it. If [the patent does issue], it may well be invalid, yet many will prefer to pay a modest royalty than to contest it, even though Lear allows them to accept a license and pursue the contest without paying royalties while the.fight goes on. The. result in such a case would be unjustified royalty payments from many who would prefer not to pay them rather than agreed fees from one or a few who are entirely willing to do so.” Painton & Co. v. Bourns, Inc., 442 F. 2d, at 225.

The point is that those who might be encouraged to file for patents by the absence of trade secret law will include inventors possessing the chaff as well as the wheat.' Some of the chaff — the nonpatentablfe discoveries — will be thrown out by the Patent Office, but in the meantime society will have been deprived of use of those discoveries through trade secret-protected licensing. Some of the. chaff may not be thrown out. This Court has noted the difference between the standards used by the Patent Office and the courts to determine patentability. Graham v. John Deere Co., 383 U. S. 1, 18 (1966).17 In Lear, Inc. v. Adkins, 395 U. S. 653 (1969), the Court thought that an invalid patent was so serious a threat to the free use of *489ideas already in the public domain that the Court permitted licensees of the patent holder to challenge, the validity of the patent. Better had the invalid patent, never been issued. More of those patents would likely issue if trade secret law were abolished. Eliminating trade secret law for the doubtfully patentable invention is thus likely to have deleterious effects on society and patent policy which we cannot say are balanced out by the speculative gain which might result from the encouragement of some inventors with doubtfully patentable inventions which -deserve patent protection to come forward and apply for patents. • There is no conflict, then, between trade'secret, law and the patent law policy of disclosure, at least insofar as the first two categories of patentable subject matter are concerned.

The final category of patentable subject matter to deal with is the clearly patentable invention, i. e., that invention which the owner believes to meet the standards of patentability. It is here that the federal interest in disclosure is at its peak;- these inventions, novel, useful and non obvious, are “ ‘the things which, are worth to the public .the embarrassment of an exclusive patent.’ ” Graham v. John Deere Co., supra, at 9 (quoting. Thomas Jefferson). The interest of the public is that the bargain of 17 years of exclusive use in return for disclosure be accepted. If a State, through a system of protection, were to cause a substantial risk that holders of patentable inventions would not seek .patents, but rather would -rely on the state protection, we would be compelled to hold that such a system could not constitutionally cbntinue to exist. In the case of trade secret law no reasonable risk of déterrence from patent application by' those who can reasonably expect' to be granted patents exists.

■ Trade Secret law provides far weaker protection in *490many'.respects than the patent law.18 While trade secret law does not forbid the discovery of the trade secret by fair and honest means, e. g., independent creation or reverse- engineering, patent law operates “against the world,” forbidding any use of the invention -for whatever purpose for a significant length of time. The holder of a trade secret also takes a substantial risk that the secret will be passed on to his competitors, by theft or by breach of a confidential relationship, in a manner not easily susceptible of discovery or proof. Painton & Co. v. Bourns, Inc., 442 F. 2d, at 224. Where patent law acts as a barrier, trade secret law functions relatively as a sieve. ' The possibility that an inventor who believes his invention meets the standards of patentability will sit back, rely on trade secret law, and after one year of use forfeit any right to patent protection, 35 U. S. C. § 102 (b), is rémote indeed.

Nor does society face much risk that scientific Or technological progress will be impeded by the rare inventor with a patentable invention who chooses trade secret protection over patent protection. The ripeness-of-time concept of invention, developed from the study of the -many independent multiple, discoveries in history, predicts. that if a particular individual had not made a particular discovery others would.have, and in probably a relatively short period of time. ' If something is to be discovered at all- very likely it will be discovered by more than one person. Singletons and Multiples in Science (1961), in R. Merton, The Sociology of Science 343 (1973); J. Cole & S. Cole, Social Stratification in Science 12-13, 229-230 (1973); Ogburn & Thomas, Are Inventions Inevitable?, 37 Pol. Sci. Q. 83 (1922).19 Even *491were an inventor to keep his discovery completely to himself, something that neither the patent nor trade secret laws forbid, there is a high probability that it will be soon independently developed. If the invention, though still a trade secret, is put into public use, the competition is alerted to the existence of the inventor’s solution to the problem and may be encouraged to make an extra effort to independently find the solution thus known to be possible. The inventor faces pressures not only from private industry, but from the skilled scientists who work in our universities and our other great publicly supported centers of learning and research.

We conclude that the extension of trade secret protection to clearly patentable inventions does not conflict with the patent policy of disclosure. Perhaps because trade secret law does not produce any positive effects in the. area of clearly patentable inventions, as opposed to the beneficial effects resulting from trade secret protection in the areas of the doubtfully patentable and the clearly unpatentable inventions, it has been suggested that partial pre-emption may be appropriate, and that courts should refuse to apply trade secret protection to inventions which the holder should have patented, and which would have been, thereby, disclosed.20 However, since' there is no real possibility that trade secret law will conflict with the federal policy favoring disclosure of clearly patentable inventions partial pre-emption is inapprbpri*492ate. Partial pre-emption, furthermore, could well create serious problems for state courts in the administration of trade secret law. As a preliminary matter in trade secret actions, state courts would be obliged to distinguish between what a reasonable inventor would and would not <correctly consider to be clearly patentable, with the holder of the trade secret arguing that the invention was not patentable and the misappropriator of. the trade secret árguing its undoubted novelty, utility, and -non-obviousness. Federal courts have a difficult enough time trying to determine whether an invention, narrowed by the patent application procedure21 and fixed in the specifications which describe the invention for which the patent has been granted,, is patentable/22 Although state courts in some circumstances must join federal courts in judging whether an issued patent is valid, Lear, Inc. v. Adkins, supra, it would be undesirable ■ to. impose the almost impossible burden on state courts to determine the patentability — in fact and in the mind of a' reasonable inventor — of a discovery which has not been patented and remains entirely uncircumscribed by expert analysis in the administrative process. . Neither complete nor partial pre-emption of state trade secret,law is justified.

Our conclusion that patent, law does not pre-empt trade secret law is in accord with prior cases of this Court. Universal Oil Co. v. Globe Co., 322 U. S., at 484; United States v. Dubilier Condenser Corp., 289 U. S., at 186—187; Becher v. Contoure Laboratories, 279 U. S. 388, 391 (1929); Du Pont Powder Co. v. Masland, 244 U. S. 100, 102 (1917); Dr. Miles Medical Co. v. Park & Sons Co., 220 U. S. 373, 402-403 (1911); Board of Trade v. Christie *493 Grain & Stock Co., 198 U. S. 236, 250-251 (1905).23 Trade secret law and patent law have co-existed in this country for over one hundred years. Each has its particular, role to play, and the operation of one does not take away from the need for the other. Trade secret law encourages the development and exploitation of those items of lesser or different invention than might be ac- ‘ corded protection under the patent laws, but which items still have an important part to play in the technological •and scientific advancement of the Nation. Trade secret law promotes the sharing of knowledge, and the efficient operation of industry; it permits the individual inventor to reap the rewards of his labor by contracting with a company large'enough to develop and exploit it. Congress, by its silence over these many years, has seen the wisdom of allowing the States to enforce trade secret protection.. Until Congress takes affirmative action to the contrary, States should be free to grant protection to trade secrets.

Since we- hold that Ohio trade secret law is not preempted by the federal patent law, the judgment of the Court of Appeals for the Sixth Circuit is reversed,, and the case is remanded to the Court of Appeals with directions to reinstate the judgment of the District Court.

It 'is- so ordered.

Mr. Justice Powell took no part in-the decision of this case.

Mr. Justice Marshall,

concurring in the result.

Unlike the Court, I do not believe that the possibility that an inventor with a patentable invention will rely *494on state trade secret law rather than apply for a patent is “remote indeed.” Ante, at 490. State trade secret law provides substantial protection to the inventor who intends to use or sell the invention himself rather than license it to others,, protection which in its unlimited duration is clearly superior to the 17-year monopoly afforded by the patent laws.' I have no doubt that the existence of trade secret protection provides in some instances a substantial disincentive to entrance into the patent system, and thus deprives society of the benefits of public disclosure of the invention which it is the policy of the patent laws to encourage. This case may well be such an instance.

But my view of sound policy in this area does not dispose of' this case. Rather, the question presented in this case is whether Congress, in enacting the patent laws, intended merely to offer inventors a limited monopoly in exchange for disclosure of their invention, or instead to exert pressure on inventors to enter into this exchange by withdrawing any alternative possibility of legal protection for their inventions. I am persuaded that the former is the case. State trade secret laws and the federal patent laws have co-existed for many, many years. During this time, Congress has repeatedly demonstrated its' full awareness of the existence of the trade secret system, without any indication of disapproval. Indeed, Congress has in a number of instances given explicit federal protection to trade secret information provided to federal agencies. See, e. g., 5 U. S. C. § 552 (b) (4); 18 U. S. C. § 1905; see generally Appendix to Brief for Petitioner. Because of this, I conclude that there is “neither such actual conflict between the two. schemes of regulation that both cannot stand in the same area, nor evidence of a congressional design to preempt the' field.” Florida Avocado Growers v. Paul, *495373 U. S. 132, 141 (1963). I therefore concur in the result reached by. the majority of the Court.

Mr. Justice Douglas,

with whom Mr. Justice Bren- ■ nan concurs, dissenting.'

Today’s decision is at war with the philosophy of Sears, Roebuck & Co. v. Stiffel Co., 376 U. S. 225, and Compco. Corp. v. Day-Brite Lighting, Inc., 376 U. S. 234. Those cases involved patents — one of a pole lamp and one of fluorescent lighting fixtures each of which was declared invalid. The lower courts held, however, that though the patents were invalid the sale of identical or confusingly similar products to the products of the patentees violated state unfair competition laws. - We held .that when an article is unprotected by a patent, state law may not forbid others to copy it, because every article not covered by a. valid patent is in the public domain.' Congress in "the patent laws decided that where no patent existed, free competition should prevail; that where a patent is rightfully issued, the right to exclude others should obtain for no longer than 17 years, and that the States may not "under some other law, such as that forbidding unfair competition, give’ protection of a kind that clashes with-the objectives of the-federal patent laws,"1 376 U. S., at 231.

The product involved in this suit, sodium iodide synthetic crystals, was a product that could be patented but was not. Harshaw the inventor apparently contributed greatly to the technology in that field by developing processes, procedures, and’ techniques that produced *496much larger crystals than any competitor. These processes, procedures, and techniques were also patentable; but no patent was sought. Rather Harshaw sought to protect its trade secrets by contracts with its employees. And the District Court found that, as a result of those secrecy precautions, “not sufficient disclosure occurred so as to place the claimed trade secrets in the public domain”; and those findings were sustained by the Court of Appeals.-

The District Court issued a permanent injunction against respondents, ex-employees, restraining them from using the processes used by Harshaw. By a patent which would require full disclosure Harshaw could have obtained ’ a 17-year monopoly against the world. By the District Court’s injunction, which the Court approves and reinstates, Harshaw gets a permanent injunction running into perpetuity against respondents. In Sears, as in the present case, an injunction against the unfair competitor issued. We said: “To allow a State by use of its law of unfair competition to prevent the copying of an article which represents too slight an advance to be patented would be to permit the State to block off from the public something which federal law has said belongs to the public. The result would be that while federal law grants only 14 or 17 years’ protection to genuine inventions, see 35 U. S. C. §§ 154, 173, States could allow perpetual protection to articles too lacking in novelty to merit any patent at all under federal constitutional standards. This would be too great an encroachment on the federal patent system to be tolerated.” 376 U. S., at 231-232.

The conflict with the patent laws is obvious. The decision of Congress to adopt a patent system was based on the idea that there will be much more innovation if discoveries are disclosed and patented than .there will be when everyone works in secret. Society thus fosters a *497free exchange of technological information at the cost of a limited 17-year monopoly.2.

A trade secret,3 unlike a patent, has no property dimension. That was the view of the Court of Appeals, 478 F. 2d 1074,1081; and its decision is supported by what Mr. Justice Holmes said in Du Pont Powder Co. v. Masland, 244 U. S. 100, 102:

“The word property as applied to trade-marks and trade secrets is an unanalyzed expression of certain *498secondary consequences of the primary fact that the. law makes some rudimentary requirements of good faith. Whether the plaintiffs have any valuable secret or not the defendant knows the facts, whatever they are, through a special confidence that he accepted. The property may be denied but the confidence cannot be. Therefore the starting point for the present matter is not property or due process .of law, but that the defendant stood in confidential relations with the plaintiffs, or one of them. These have given place to hostility, and the first thing to be made sure of is that the defendant shall not fraudulently abuse the trust reposed in him. . It is the usual incident of confidential relations. If there is any disadvantage in the fact that he knew the plaintiffs’ secrets he must take .the burden with the good.”4

A suit to redress theft of a trade secret is grounded in tort damages for breach of a contract — a historic remedy, Cataphote Corp. v. Hudson, 422 F. 2d 1290. Damages for breach of a confidential relation are not pre-empted by this patent law, but an injunction *499against use is pre-empted because the patent law states the only monopoly over trade secrets that is enforceable by specific performance; and that monopoly exacts as a price full disclosure. A trade secret can be protected only by being kept secret. Damages for breach of a contract are one thing; an injunction barring disclosure does service for the protection accorded valid patents and is therefore pre-empted.

From the findings of fact of the lower courts, the process involved in this litigation was unique, such a great discovery as to make its patentability a virtual certainty. Yet the Court’s opinion reflects a vigorous activist anti-patent philosophy. My objection is not because it is activist. This is a problem that involves no neutral principle. The Constitution in Art. I, § 8, cl. 8, expresses the activist policy which .Congress has enforced by statutes. It is that constitutional policy which we should enforce, not our individual notions of the public good.

I would affirm the judgment below.

GAJARSA, Circuit Judge.

DECISION

National Recovery Technologies, Inc. (“NRT”) appeals from the October 29, 1997 judgment of the United States District Court for the Middle District of Tennessee, No. 3-96-0154, granting summary judgment to Magnetic Separation Systems, Inc. and Garry R. Kenny (collectively “MSS”). The district court held that claim 1 of U.S. Patent No. 5,260,576 (“the ’576 patent”) was invalid for lack of enablement under 35 U.S.C. § 112, paragraph 1 (1994). We affirm the decision of the district court.

BACKGROUND

NRT is engaged in the manufacture and sale of large-scale automated recycling equipment and systems. NRT is also the assignee of the ’576 patent, issued on November 9, 1993, entitled “Method and Apparatus for the Separation of Materials Using Penetrating-Electromagnetic Radiation.” The ’576 patent addresses the problem of separating recyclable plastic materials that are virtually indistinguishable to the human eye by using penetrating electromagnetic radiation.

In recycling plastics, it is often useful to separate plastics with similar chemical compositions. A common sorting problem in the recycling industry is the separation of polyvinyl chloride (“PVC”) containers from polyester (“PET”) containers. PVC and PET containers are similar in appearance and are difficult to separate manually.1 However, PVC and PET containers have different chemical properties (hence the desire to separate them) and thus absorb different amounts of electromagnetic radiation {e.g., x-ray) when irradiated. It is generally well known that PVC containers absorb more electromagnetic radiation than PET containers for an equivalent material thickness. The difference in the ability to absorb for each of the materials can be used to differentiate between the two types of plastic. It is assumed that a high transmittance reading (low absorption) indicates a PET container, and a low transmittance (high absorption) reading indicates a PVC container.

According to the ’576 patent, the prior art systems suffered from two drawbacks. First, the prior art systems were only able to scan and classify one container at a time, greatly slowing the processing of containers. Second, the prior art systems were not able to differentiate between radiation that passed through thicker portions of the containers, such as the neck and base, and radiation that passed through the central portions of the containers.

In the separation process disclosed in the preferred embodiment of the ’576 patent, containers to be sorted are advanced along a conveyor wide enough to accommodate several containers. Each container is irradiated with a sheet-like beam of electromagnetic radiation as it progresses along the conveyor. A number of detectors spanning the width of the conveyor are positioned below the containers to measure the intensity level of electromagnetic radiation that passes through each of the containers. The patented process then uses a microprocessor to compare the detected values to preset thresholds to classify the container as being made of one type of plastic or another. The containers are then mechanically separated on this basis. If the container is classified as one type of material {e.g., PVC plastic), the container is permitted to fall off the end of the primary conveyor onto a second conveyor. If the container is classified as a second type of material {e.g., PET plastic), air valves located at the end of the primary conveyor are activated, thereby directing the container onto a third conveyor. The ’576 patent states that this process is able to classify and separate up to eighty containers per second.

*1193However, the ’576 patent recognizes that containers cannot be accurately separated simply upon the assumption that detecting a low transmittance indicates a PVC container and detecting a high transmittance indicates a PET container. Where the PET container is significantly thicker than the PVC container, or where the electromagnetic radiation passes through many layers of PET plastic before detection, the detected transmittance level can be similar to, or even lower than that of a PVC container. This can potentially cause a PET container to be misclassified as a PVC container.

There are often irregularities in container thickness due to both the shape of the container 2 and the fact that many containers are folded, crushed, or otherwise mangled by the time the containers enter the separation stage of the recycling process. These irregularities may result in a substantial variance in the thickness of the material through which the electromagnetic radiation passes. The variance in thickness may in turn cause the detected radiation transmittance to vary significantly depending on the section of the container irradiated and measured. As a result, the irregularities and variances in container thickness could cause a PET container to be erroneously classified as a PVC container where the detected transmittance level is reduced because the electromagnetic radiation passed through an abnormally thick portion of the container before detection.

The ’576 patent specifically addresses the problem of misclassification due to irregularities in container thickness. The written description of the ’576 patent discloses that containers are to be irradiated at several points along their length. Thus, several intensity measurements are recorded for each container. A microprocessor connected to the detectors compares the transmittance measurements for different portions of a particular container to one another, and a subset of the highest readings are selected for processing. The measurements selected are presumed to be measurements of electromagnetic radiation energy that did not pass through an irregularity. The selected measurements are compared to preset threshold values in order to classify the containers as being made of-either PET or PVC plastic. The containers are mechanically separated on this basis as described above.

However, given the unpredictability of container orientation and possible damage to a container’s “regular” portions, results of this process are not completely accurate in distinguishing between containers of differing materials. If a container is folded- several times, or is severely deformed by the time it reaches the scanning process, even those measurements with the highest transmittance intensity may have been taken through irregular portions, thereby leading to a potentially erroneous classification. The ideal solution, therefore, is to ensure that only the regular portions of the container are measured and to use only these measurements in classifying the container.

Claim 1 is the only independent claim at issue in this appeal. It reads as follows:

A method of distinguishing and separating material items having different levels of absorption of penetrating electromagnetic radiation, comprising the steps of:

(a) conveying a plurality of said material items in a random manner simultaneously and longitudinally along an elongated feed path;

(b) establishing a transverse region across said feed path irradiated by a sheet of penetrating electromagnetic radiation;

(c) irradiating said plurality of material items in said transverse region with said penetrating electromagnetic radiation;

(d) simultaneously measuring the amount of penetrating electromagnetic radiation passing through each material item in said transverse region at any instant of time as said items are continuously conveyed longitudinally through said trans*1194verse region to generate process signals; wherein more than one process signal is generated for each of said material items, each process signal being commensurate with the amount of penetrating electromagnetic radiation passing through a portion of each material item which is different from any other portion of said material item, and selecting for processing those of said process signals which do not pass through irregularities in the bodies' of said material items; and

(e) simultaneously analyzing said process signals to cause said process signals to actuate means for directing said items to a different destination commensurate with the amount of said penetrating electromagnetic radiation passing through each of said corresponding material items.

(emphasis added).

On February 9, 1996, NRT filed a complaint against MSS in the United States District Court for the Middle District of Tennessee alleging that MSS infringed several of NRT’s patents related to the automatic classification and separation of recyclable plastic materials. MSS defended by arguing that the patents at issue were invalid under 35 U.S.C. §§ 101-103 and 112. On October 27, 1997, the parties stipulated that only claims 1-8 and 10 of the ’576 patent remained at issue, and the other asserted patents and the remainder of the claims of the ’576 patent were dropped from the litigation. On October 29,1997, the district court issued a Memorandum and Order granting MSS’s. motion for summary judgment that concluded that claim 1 of the ’576 patent was invalid for lack of enablement under 35. U.S.C, § 112, paragraph 1. On November 7, 1997, the district court entered an Agreed Final Order that the asserted dependent claims 2-8 and 10, the remainder of the claims at issue in the case, were also invalid for lack of enablement because claims 2-8 and 10 incorporate claim 1 by reference pursuant to 35 U.S.C. § 112, paragraph 4.

DISCUSSION

On appeal, NRT argues that the district court erred in holding claims 1-8 and 10 invalid under § 112, paragraph 1 for two reasons. NRT asserts that the district court erred in its construction of the term “selecting” in claim 1. NRT also asserts that the district court erroneously required the claimed invention to work perfectly under all circumstances. We have jurisdiction over this appeal- under 28 U.S.C. § 1295(a)(1) (1994).

1. Standard of Review

We review the district court’s grant of summary judgment de novo, with all justifiable factual inferences being drawn in favor of the party opposing the motion. See Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 255, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). Summary judgment is appropriate where there is no genuine issue of material fact and the moving party is entitled to judgment as a matter of law. See Fed.R.Civ.P. 56(c). Additionally, we review the district court’s claim construction de novo. See Cybor Corp. v. FAS Technologies, Ltd., 138 F.3d 1448, 1456, 46 USPQ2d 1169, 1174 (Fed.Cir.1998) (en banc). Whether a claim is enabled under 35 U.S.C. § 112, paragraph 1 is a question of law, although based upon underlying factual findings. See PPG Indus., Inc. v. Guardian Indus. Corp., 75 F.3d 1558, 1564, 37 USPQ2d 1618, 1623 (Fed.Cir.1996); In re Goodman, 11 F.3d 1046, 1049-50, 29 USPQ2d 2010, 2013 (Fed.Cir.1993).

2. Claim Construction

NRT argues that the district court erred in incorporating its claim construction in its discussion of enablement and in interpreting claim 1 to require perfect isolation between those signals that pass through irregularities and those that do not. NRT argues that the term “select” means that only those signals that do not pass through irregularities are to be “preferred” to those signals that do. According to NRT, the term “select” does not require absolute perfection in separating signals that pass through an irregularity from those that do not. Because claim construction is a matter of law that we review de novo, we independently construe claim 1 of the ’576 patent.

*1195Claim interpretation begins with the language of the claim itself. See Bell Communications Research, Inc. v. Vitalink Communications Corp., 55 F.3d 615, 619, 34 USPQ2d 1816, 1819 (Fed.Cir.1995). The claim terms are to be given their ordinary meaning unless it is apparent that the inventor intended to use them differently. See York Prods., Inc. v. Central Tractor Farm & Family Ctr., 99 F.3d 1568, 1572, 40, USPQ2d 1619, 1622 (Fed.Cir.1996); Quantum Corp. v. Rodime, PLC, 65 F.3d 1577, 1580, 36 USPQ2d 1162, 1165 (Fed.Cir.1995). At issue on appeal is the construction of part (d) of claim 1, more specifically, the limitation that requires the claimed invention to “select[ ] for processing those of said process signals which do not pass through irregularities in the bodies of said material items.” The parties agree that the ordinary meaning of the term “select” means “chosen in preference to another or others; picked out, especially for excellence or some special quality; picked.” Webster’s New World Dictionary (3rd College Ed.) (emphasis added).

The plain meaning of the term “select,” including the dictionary meaning agreed upon by the parties, implies that some signals are to be picked or chosen and others are to be excluded based upon some special, unique or discrete quality. A specific choice is being made. Both claim 1 and the specification of the ’576 patent make clear that those signals that do not pass through irregularities are to be chosen for processing to the exclusion of those signals that do pass through irregularities. Nothing in the plain and ordinary meaning of the term “select” indicates that only a preference for signals that do not pass through irregularities is required. Furthermore, nothing in claim 1 indicates that signals are to be picked based on any other characteristic than penetrating a “normal” or regular portion of a container. Claim 1 clearly indicates that the “special quality” for which a signal is selected is whether or not that signal has passed through an irregularity. There is no indication in the plain meaning of the limitation, or from the claim as a whole, that this selection criterion is to apply only some of the time.

Nothing in the specification indicates that the inventor intended to deviate from the plain and ordinary meaning of the word “select.” Thus, the plain language of the claim when read in light of the specification is clear that the term “select” does not merely require a “preference” for certain signals. Signals are to be specifically chosen based on whether or not they pass through irregularities in the container.

3. Enablement under 35 U.S.C. § 112, Paragraph 1

Every patent is presumed valid. See 35 U.S.C. § 282 (1994). The presumption of validity includes a presumption that the patent complies with § 112. See Northern Telecom, Inc. v. Datapoint Corp., 908 F.2d 931, 941, 15 USPQ2d 1321, 1329 (Fed.Cir.1990). “Invalidity for lack of enablement is a conclusion of law and must be supported by facts proved by clear and convincing evidence. ...” Id. Because no genuine issues of material fact exist that preclude the grant of summary judgment on this issue, we determine, as a matter of law, whether the specification of the ’576 patent satisfies the requirements of 35 U.S.C. § 112, paragraph 1 with respect to claim 1.

The first paragraph of 35 U.S.C. § 112 states:

The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.

The enablement requirement of § 112 demands that the patent specification enable “those skilled in the art to make and use the full scope of the claimed invention without ‘undue experimentation.’” Genentech, Inc. v. Novo Nordisk A/S, 108 F.3d 1361, 1365, 42 USPQ2d 1001, 1004 (Fed.Cir.1997) (quoting In re Wright, 999 F.2d 1557, 1561, 27 USPQ2d 1510, 1513 (Fed.Cir.1993)); see also In re Vaeck, 947 F.2d 488, 495, 20 USPQ2d 1438, 1444 (Fed.Cir.1991). The enablement *1196requirement ensures that the public knowledge is enriched by the patent specification to a degree at least commensurate with the scope of the claims. The scope of the claims must be less than or equal to the scope of the enablement. The scope of enablement, in turn, is that which is disclosed in the specification plus the scope of what would be known to one of ordinary skill in the art without undue experimentation. See Goodman, 11 F.3d at 1050, 29 USPQ2d at 2013 (“the specification must teach those of skill in the art ‘how to make and how to use the invention as broadly as it is claimed’ ”) (citing Vaeck, 947 F.2d at 496); In re Fisher, 57 C.C.P.A. 1099, 427 F.2d 833, 839, 166 USPQ 18, 24 (CCPA 1970) (“The scope of the claims must bear a reasonable correlation to the scope of enablement provided by the specification to persons of ordinary skill in the art.”).

In the present case, the district court held that the written description of the ’576 patent did not meet the strictures of § 112, paragraph 1. In particular, it held that one of ordinary skill in the art could not “select! ] for processing those of said process signals which do not pass through irregularities in the bodies of said material items” without undue experimentation because the written description did not explain how to distinguish between signals that passed through irregular portions of the container and those that did not. See National Recovery Techs. v. Magnetic Separation Sys., No. 3-96-0154, slip op. at 12 (M.D.Tenn. Oct. 29,1997). The district court had sufficient evidence, including the testimony of one of the inventors, from which it found that NRT failed to determine where irregularities existed in the container. See id. at 10. The district court concluded that the specification merely instructed one of ordinary skill in the art to select those signals with the highest transmission measurements, not to select those signals that did not pass through irregularities as required by claim 1. See id. at 12.

NRT argues that the district court erred because it required the disclosed embodiment to work perfectly under all circumstances. According to NRT, the district court misinterpreted our decision in In re Wright, 999 F.2d 1557, 27 USPQ2d 1510 (Fed.Cir.1993) to require enablement under all possible conditions. NRT is indeed correct that a claim is not invalid for lack of operability simply because the invention does not work perfectly under all conditions. See Hildreth v. Mastoras, 257 U.S. 27, 34, 42 S.Ct. 20, 66 L.Ed. 112 (1921) (“The machine patented may be imperfect in its operation; but if it embodies the general principle and works ... it is enough.”); Decca, Ltd. v. United States, 210 Ct.Cl. 546, 544 F.2d 1070, 191 USPQ 439 (Ct.Cl.1976) (“The mere fact that the system has some drawbacks, or that under certain postulated conditions it may not work ... does not detract from the oper-ability of the disclosed equipment to perform its described function.”). However, NRT is incorrect in its characterization of the district court’s ruling on enablement as requiring perfect operation from the patented process.

Whether a patented device or process is operable is a different inquiry than whether a particular claim is enabled by the specification. In order to satisfy the enablement requirement of § 112, paragraph 1, the specification must enable one of ordinary skill in the art to practice the claimed invention without undue experimentation. Thus, with respect to enablement the relevant inquiry lies in the relationship between the specification, the claims, and the knowledge of one of ordinary skill in the art. If, by following the steps set forth in the specification, one of ordinary skill in the art is not able to replicate the claimed invention without undue experimentation, the claim has not been enabled as required by § 112, paragraph 1.

The case before us presents a classic example of a claim that is broader than the enablement as taught in the specification. The specification of the ’576 patent first acknowledges the problem: sometimes radiation intensity readings are misleading because the radiation has passed through abnormally thick portions of the scanned container. The ideal solution to this problem is clear: discard intensity measurements taken through irregularities and use only those measurements taken through the regular portions of the container. Claim 1 claims this ideal solution in the *1197step of “selecting for processing those of said process signals which do not pass through irregularities in the bodies of said material items.” However, the specification of the ’576 patent does not describe how to perform this ideal selection step. Rather, the specification instructs one of ordinary skill in the art to “use only those measurements of highest transmission rate through the item_” ’576 Patent, col. 3, 11. 40-41. The last sentence of the written description states that:

the most reliable measurements for making a classification are those measurements taken through those portions of the body of an item to be classified which exhibit the greatest rates of transmission of radiation through the item (such as those taken through a relatively thin cross section such as through an unfolded central portion of the container).

’576 Patent, col. 6, 11. 52-58. The specification is clear that in order to obtain the most reliable measurements, a good proxy for intensity measurements that do not pass through irregularities are those measurements with the highest transmission rates. However, enabling a proxy for the claimed invention is not the same as enabling the claimed invention itself.

While the written description does enable one of ordinary skill in the art to approximate the claimed function, this is not the same as enabling one of ordinary skill in the art to perform the actual selection step of claim 1 for which NRT claims patent protection. The written description does not at all purport to enable one of ordinary skill in the art to determine where irregularities exist in the containers. In fact, the ’576 patent specification points out that equipment limitations make an actual determination of the location of regular and irregular portions infeasible. It states:

We have found that, in practice, taking a measurement through only a relatively thin cross section of an item requires detailed knowledge of the geometry and orientation of the item (such as a container). Accordingly, placement of an item between a radiation source and a radiation detector such that radiation passing only through a relatively thin cross section is measured requires sophisticated and expensive materials handling means.

’576 Patent, col. 3, 11. 16-24. NRT argues that “[w]hile as a theoretical possibility it might be feasible to construct a system that ignores every single perturbation and flaw in virtually all of the items processed, the ’576 patent discloses and claims a workable, practical system, not a theoretical possibility.” However, as we have explained above, claim 1 broadly claims exactly this theoretical possibility that NRT admits is not disclosed in the specification of the ’576 patent.

The record before us does not support NRT’s contention that one of ordinary skill in the art would be able to construct a machine that is capable of selecting signals based on whether the signals pass through container irregularities without undue experimentation. While the necessity of some experimentation does not preclude enablement, the experimentation must not be unduly extensive. See PPG Industries, 75 F.3d at 1564, 37 USPQ2d at 1618. Whether making and using the claimed invention would have required undue experimentation is a legal conclusion based upon underlying facts. See Genentech, 108 F.3d at 1365, 42 USPQ2d at 1004 (citing In re Wands, 858 F.2d 731, 735, 736-37, 8 USPQ2d 1400, 1402, 1404 (Fed.Cir.1988)).

The record shows that as of October 29, 1990, the date the ’576 patent was filed, there was no known way for one of ordinary skill in the art of materials processing to distinguish x-ray readings which passed through bottle irregularities from those x-ray readings which did not pass through irregularities. The record moreover indicates that as of the time the ’576 patent was filed, even one of the listed inventors of the patent, Dr. Som-mers, believed that additional research, development, and experimentation needed to be conducted before a device could be built that would practice the invention as claimed - a device that would selectively identify signals based on whether they passed through sample irregularities. During his deposition testimony, Dr. Sommers was asked if the ’576 patent described any method, technique or algorithm that would tell someone how to *1198determine where irregularities existed in the items to be sorted. Dr. Sommers replied that he believed “what the patent discloses is the need, as a method, to determine that irregularity in the bottle” and at the time the ’576 patent was filed, NRT “did not know particularly how to do that .... [and was] still developing that process.” Dr. Sommers further noted that although analyzing the signal transmission measurements would give an indication of where irregularities might exist in the samples being sorted, if “the complete bottle is an irregularity ... you’re not going to get good readings.... ”

The ’576 patent therefore recognizes a specific need in the materials sorting field and suggests a theoretical answer to that need. It provides a starting point from which one of skill in the art can perform further research in order to practice the claimed invention, but this is not adequate to constitute enablement. See Genentech, 108 F.3d at 1366, 42 USPQ2d at 1005. The specification of the ’576 patent therefore does not enable one of ordinary skill in the art to practice the full scope of the invention embodied in claim 1 without undue experimentation. The most that NRT can be credited with is promising the ideal result in claim 1, even though the specification does not completely deliver on this promise.

CONCLUSION

Because the specification of the ’576 patent does not enable one of ordinary skill in the art to practice the invention embodied in claim 1 without undue experimentation, we affirm the district court’s grant of MSS’s motion for summary judgment that claim 1 is not enabled and thus invalid under 35 U.S.C. § 112, paragraph 1. Because dependent claims 2-8 and 10 stand or fall with independent claim 1, we affirm the district court’s judgment that these claims are also invalid under 35 U.S.C. § 112, paragraph 1.

AFFIRMED.

COSTS.

Each party to bear its own costs.

TRAGER, District Judge:

Plaintiff-Appellant Cosmetic Ideas, Inc. (“Cosmetic”) appeals the decision of the United States District Court for the Central District of California (“district court”) dismissing its claims against DefendantsAppellees IAC/InteractiveCorp, Home Shopping Network, Inc., HSN LP, and HSN General Partner LLC (collectively, “HSN”) for lack of subject-matter jurisdiction. The district court determined that Cosmetic failed to comply with the registration requirement of 17 U.S.C. § 411(a), which makes registration of a copyright a prerequisite to bringing an infringement *614suit, then concluded that lack of registration deprived it of jurisdiction. Cosmetic contends that it complied with § 411(a), because it submitted a completed registration application to the Copyright Office before instituting its infringement action. HSN argues that registration occurs only-after the Copyright Office has affirmatively granted a registration.

Background

(1)

In 199V, Cosmetic, doing business as Sweet Romance Jewelry Manufacturing, created a piece of costume jewelry known as the “Lady Caroline Lorgnette” (“the necklace”). Cosmetic began manufacturing and selling copies of the necklace in 1999, and continues to manufacture and sell copies through various stores and websites. Cosmetic claims that sometime between 2005 and 2008, HSN began manufacturing and distributing copies of a “virtually identical” necklace.

On March 6, 2008, Cosmetic submitted an application to the Copyright Office for registration of its copyright in the necklace, and received confirmation of receipt of the application on March 12, 2008. On March 27, 2008, Cosmetic filed a complaint alleging that HSN had infringed on Cosmetic’s copyright in the necklace. Although the Copyright Office ultimately issued Cosmetic a registration certificate for its copyright in the necklace, it did not do so before Cosmetic filed its complaint.1

On June 2, 2008, HSN filed a motion to dismiss for failure to state a claim and lack of subject-matter jurisdiction. HSN argued that Cosmetic did not possess a valid copyright registration when it commenced its action, and thus the court lacked subject-matter jurisdiction over the claim of copyright infringement. On June 17, 2008, the district court granted the motion to dismiss on the basis that it lacked subject-matter jurisdiction. Cosmetic timely appealed.

We have jurisdiction pursuant to 28 U.S.C. § 1291, and we reverse.

(2)

The district court’s dismissal for lack of subject-matter jurisdiction is reviewed de novo. Rattlesnake Coal v. U.S. EPA, 509 F.3d 1095, 1100 (9th Cir.2007). In this case, the district court’s decision to dismiss was necessarily — though without discussion — based on its interpretation of the Copyright Act, 17 U.S.C. § 101 et seq. (2006). Questions of statutory interpretation are reviewed de novo. Mollison v. United States, 568 F.3d 1073, 1075 (9th Cir.2009).

Discussion

(1)

We first address HSN’s argument that § 411(a)’s registration requirement-which mandates that a copyrighted work be registered before an infringement action can be brought — is a jurisdictional prerequisite to suit. In Reed Elsevier, Inc. v. Muchnick, — U.S. -, 130 S.Ct. 1237, 176 L.Ed.2d 17 (2010), the Supreme Court held otherwise.2 Specifically, Reed *615 Elsevier held that, although “[sjection 411(a)’s registration requirement is a precondition to filing a claim,” it “does not restrict a federal court’s subject-matter jurisdiction.” 130 S.Ct. at 1241. Therefore, the district court’s dismissal of Cosmetic’s complaint on the basis of lack of subject-matter jurisdiction — which must have been predicated on a conclusion that Cosmetic had not registered its copyright — was error.

(2)

However, this conclusion does not end our inquiry, as the district court still may have correctly dismissed the suit on another basis. This court “may affirm the district court on any ground supported by the record, even if the ground is not relied on by the district court.” Charley’s Taxi Radio Dispatch Corp. v. SIDA of Hawaii, Inc., 810 F.2d 869, 874 (9th Cir. 1987). Below, HSN moved to dismiss the complaint for failure to state a claim, pursuant to Federal Rule of Civil Procedure 12(b)(6). Specifically, HSN argued that Cosmetic failed to register the necklace, and post-Reed Elsevier, registration is an element of an infringement claim, rather than a jurisdictional bar.

Thus, we are asked to answer the question: What does it mean to “register” a copyrighted work? Stated another way, is a copyright registered at the time the copyright holder’s application is received by the Copyright Office (the “application approach”), or at the time that the Office acts on the application and issues a certificate of registration (the “registration approach”)? 3 Our sister circuits have split in answering this question.4 For instance, *616the Fifth and Seventh Circuits have adopted the application approach. Apple Barrel Prods., Inc. v. Beard, 730 F.2d 384, 386-87 (5th Cir.1984); Chicago Bd. of Educ. v. Substance, Inc., 354 F.3d 624, 631 (7th Cir.2003). The Tenth and Eleventh Circuits have adopted the registration approach. La Resolana, 416 F.3d at 1202-04; M.G.B. Homes, Inc. v. Ameron Homes, Inc., 903 F.2d 1486, 1489 (11th Cir.1990), abrogated in part by Reed Elsevier, 130 S.Ct. at 1243 & n. 2.

We now turn to this issue.5

A

We first note that, according to the representations of Cosmetic at oral argument and in subsequent briefing, the Copyright Office has now acted on Cosmetic’s application and issued a certificate of registration of copyright in the necklace. Once a certificate is issued, the registration dates back to the date of application. 17 U.S.C. § 410(d). Thus, Cosmetic can now allege the registration element of an infringement claim, irrespective of whether registration accrues at application or issuance. Nonetheless, we address the issue of the point at which a work is registered because it is an unsettled issue of law in this Circuit, and because it is an issue “capable of repetition yet evading review” if the Copyright Office acts on the application during litigation. See Gerstein v. Pugh, 420 U.S. 103, 110 n. 11, 95 S.Ct. 854, 43 L.Ed.2d 54 (1975) (citing Sosna v. Iowa, 419 U.S. 393, 402 n. 11, 95 S.Ct. 553, 42 L.Ed.2d 532 (1975)); United States v. Brandan, 578 F.3d 1064, 1067 (9th Cir. 2009).

B

In interpreting a statutory provision, we begin with the plain language of the statute. See K & N Eng’g, Inc. v. Bulat, 510 F.3d 1079, 1081 (9th Cir.2007). Section 411(a) of the Copyright Act of 1976 (the “Act” or “1976 Act”) provides, in relevant part: “[N]o civil action for infringement of the copyright in any United States work shall be instituted until pre-registration or registration of the copyright claim has been made in accordance with this title.”6 17 U.S.C. § 411(a). “Registration” is unhelpfully defined as “a registration of a claim in the original or the renewed and extended term of copyright.” 17 U.S.C. § 101.

Because the clause at issue gives no guidance in interpreting the meaning of “registration,” we turn to the language of the statute as a whole to determine the intended meaning. See United States v. Cruz-Gramajo, 570 F.3d 1162, 1167 (9th Cir.2009) (“While the starting point is always the language of the statute itself ... [r]ather than focusing just on the word or phrase at issue, this court looks to the entire statute to determine ... intent.” *617(internal quotation marks and citations omitted)). An examination of the statutory language here, however, reveals ambiguity that leaves us no closer to determining which interpretation to adopt.

Looking to the Act as a whole, copyright registration is addressed in five consecutive sections: §§ 408 through 412. Two subsections of the Act, § 410(a) and portions of § 411(a), contain language that suggests that registration requires some affirmative steps to be taken by the Copyright Office. Section 410(a) details the role of the Register of Copyrights (“the Register”):

When, after examination, the Register of Copyrights determines that, in accordance with the provisions of this title, the material deposited constitutes copyrightable subject matter and that the other legal and formal requirements of this title have been met, the Register shall register the claim and issue to the applicant a certificate of registration under the seal of the Copyright Office.

17 U.S.C. § 410(a) (emphasis added). This provision places an active burden of examination and registration upon the Register, suggesting that registration is not accomplished by application alone. See, e.g., Loree Rodkin, 315 F.Supp.2d at 1055 (reading the provision in this way). But see Iconbazaar, 308 F.Supp.2d at 634 (pointing out that this section “could be read to apply only to the requirements for issuance of a registration certificate, not to the requirements for instituting an action for infringement”).

The latter portion of § 411(a) can be read similarly. The portion of § 411(a) setting forth the registration requirement is followed by an exception:

In any case, however, where the deposit, application, and fee required for registration have been delivered to the Copyright Office in proper form and registration has been refused, the applicant is entitled to institute a civil action for infringement if notice thereof, with a copy of the complaint, is served on the Register of Copyrights.

17 U.S.C. § 411(a) (emphasis added). Because registration in this subsection is juxtaposed with the separate act of delivering the necessary application materials to the Copyright Office, this subsection could be read to mean that Congress intended registration to require acceptance or refusal by the Register, not mere delivery. See, e.g., Strategy Source, 233 F.Supp.2d at 3-4; Loree Rodkin, 315 F.Supp.2d at 1055-56.

Other sections of the Act, however, cast doubt on this interpretation. In particular, § 408 blurs the line between application and registration and favors the application approach. In setting forth the permissive registration requirements, § 408 states: “[T]he owner of copyright or of any exclusive right in the work may obtain registration ... by delivering to the Copyright Office the deposit specified by this section, together with the application and fee specified ....” 17- U.S.C. § 408(a) (emphasis added). This section implies that the sole requirement for obtaining registration is delivery of the appropriate documents and fee. See Prunté, 484 F.Supp.2d at 40 (concurring in this analysis and adopting the application approach); Iconbazaar, 308 F.Supp.2d at 634 (same); but see Ryan, 1998 WL 320817, at *3 (adopting the registration approach and concluding that § 408(a) means “merely that the delivery of the application is a step the applicant must take, not that delivery is sufficient by itself to obtain a registration”).7

*618One final relevant subsection of the Act, § 410(d), could be read as supporting either the application or registration approach. Section 410(d) provides: “The effective date of a copyright registration is the day on which an application, deposit, and fee, which are later determined by the Register of Copyrights or by a court of competent jurisdiction to be acceptable for registration, have all been received in the Copyright Office.” 17 U.S.C. § 410(d). Because this subsection dates a later-approved registration as of the date of its application, it supports the interpretation that application is the critical event. See, e.g., Dielsi, 916 F.Supp. at 994 n. 6; Prunté, 484 F.Supp.2d at 40. However, because this back-dating does not occur until after the Copyright Office or a court has deemed the registration acceptable, the statute could be read to require action by the Register to effect registration. See, e.g., La Resolana, 416 F.3d at 1204 n. 9 (“It is not logical to assume that the relation-back provision subsumes the explicit requirements of § 411 and § 410(a).”); Ryan, 1998 WL 320817, at *2; Brush Creek, 2002 WL 1906620, at *3-4.

We are not persuaded that the plain language of the Act unequivocally supports either the registration or application approach. Although in some places the Act seems to equate registration with affirmative approval by the Copyright Office, in others the Act suggests registration is accomplished by completing the process of submitting an application. This ambiguity makes it necessary to go beyond the Act’s plain language to determine which approach better carries out the purpose of the statute.

C

When statutory language proves unclear, we work to discern its meaning by looking to “the broader context of the statute as a whole” and the purpose of the statute. See United States v. Olander, 572 F.3d 764, 768 (9th Cir.2009) (quoting Robinson v. Shell Oil Co., 519 U.S. 337, 341, 117 S.Ct. 843, 136 L.Ed.2d 808 (1997)). An understanding of the history of § 411 is helpful in discerning the statute’s purpose. Section 411 is a product of the 1976 Act, which drastically amended the 1909 Act. See Pub.L. No. 94-553, 90 Stat. 2541 (1976); see also H.R.Rep. No. 94-1476, at 129 (1976), reprinted in 1976 U.S.C.C.A.N. 5659, 5745. Under the 1909 Act, federal copyright protection attached only upon publication, and even then, only if proper notice, registration, and deposit occurred.8 See H.R.Rep. No. 94-1476, at 129, 143, 146, 150, 1976 U.S.C.C.A.N. at 5745, 5759, 5762, 5766. Protection of unpublished works was left to the states. Id. at 129, 1976 U.S.C.C.A.N. at 5745. The 1976 Act, however, created a unitary system that attached federal copyright protection to all works — published and unpublished — at the *619time of their creation. Pub.L. No. 94-553, § 301(a) (codified at 17 U.S.C. § 301(a)); see also H.R.Rep. No. 94-1476, at 129, 1976 U.S.C.C.A.N. at 5745. It also eliminated some of the prior formalities of copyright law by relaxing notice requirements and eliminating mandatory registration. See Pub.L. No. 94-553, §§ 301, 401-412 (codified at 17 U.S.C. §§ 301, 401^12); see also H.R.Rep. No. 94-1476, at 147, 150, 1976 U.S.C.C.A.N. at 5763, 5766.9 These changes vastly increased the scope of works subject to copyright protection, thus providing an incentive to create expressive works. See, e.g., Chicago Bd. of Educ., 354 F.3d at 631.

Although registration was made optional, Congress still valued having a robust federal register of existing copyrights. See H.R.Rep. No. 94-1476, at 158, 1976 U.S.C.C.A.N. at 5774.10 To this end, it chose to encourage copyright holders to register with the Office of Copyright through various statutory incentives. See 17 U.S.C. §§ 410-412. For instance, if a copyright holder secures a registration certificate within five years after first publication, such certificate will constitute prima facie evidence of both the validity of the copyright and the facts stated in the certificate. 17 U.S.C. § 410(c). Moreover, certain remedies, such as statutory damages and attorney’s fees, are available only after registration.11 17 U.S.C. § 412. Finally, registration is required for U.S. works prior to bringing any infringement action.12 17 U.S.C. § 411(a). Indeed, because registration is not mandatory under the Act, copyright holders frequently register specifically for the purpose of being able to bring suit.

With this framework in mind, we conclude that the application approach better fulfills Congress’s purpose of providing broad copyright protection while maintaining a robust federal register.

First, the application approach avoids unnecessary delay in copyright infringement litigation, which could permit an infringing party to continue to profit from its wrongful acts. Section 411(a) allows a party, after applying for registration, to litigate the claim whether the Copyright Office accepts or rejects the registration. See 17 U.S.C. § 411(a) (permitting an applicant to bring an infringement suit after the Register has rejected his or her registration, so long as the Register is notified of the suit); see also Forasté, 248 F.Supp.2d at 77 n. 10; Prunté, 484 F.Supp.2d at 40. Under the registration approach, however, a party must wait on the Copyright Office’s affirmative acceptance or rejection, despite knowing that *620suit can proceed in either event. As the leading treatise on copyright explains, the registration approach thus creates a strange scheme: “[G]iven that the claimant ... will ultimately be allowed to proceed regardless of how the Copyright Office treats the application, it makes little sense to create a period of ‘legal limbo’ in which suit is barred.” See 2 Melville B. Nimmer & David Nimmer, Nimmer on Copyright § 7.16[B][l][a][i] (2008) (“Nimmer”). The application approach avoids this legal limbo — and avoids prolonging the period of infringement — by allowing a litigant to proceed with an infringement suit as soon as he has taken all of the necessary steps to register the copyright at issue.

Moreover, the application approach avoids delay without impairing the central goal of copyright registration. As explained above, copyright registration is now a voluntary procedure, with the prelitigation registration requirement acting as an incentive to help Congress maintain a robust national register of copyrights. See H.R.Rep. No. 94-1476, at 158, 1976 U.S.C.C.A.N. at 5774 (“Copyright registration for published works, which is useful and important to users and the public at large ... should ... be induced in some practical way.”). This goal, however, is accomplished equally by the registration and application approaches, as either approach requires a party to submit the information necessary to add the copyright to the federal registry. See 17 U.S.C. §§ 408^409 (setting forth the steps an applicant must take to register a copyright).

The registration approach’s added requirement of affirmative approval or rejection before suit thus amounts to little more than just the type of needless formality Congress generally worked to eliminate in the 1976 Act. On this point, it is telling that even many courts adopting the registration approach have recognized that “construing the statute this way leads to an inefficient and peculiar result.” Brush Creek, 2002 WL 1906620, at *4 (quoting Ryan, 1998 WL 320817, at *2); Loree Rod-kin, 315 F.Supp.2d at 1056-57 (same); Strategy Source, 233 F.Supp.2d at 3 (same). We agree. It makes little sense to dismiss a case (which will likely be refiled in a matter of weeks or months) simply because the Copyright Office has not made a prompt decision that will have no substantive impact on whether or not a litigant can ultimately proceed.

Furthermore, in addition to being generally inefficient, in the worst-case scenario the registration approach could cause a party to lose its ability to sue. The Act provides a three-year statute of limitations for copyright infringement actions. 17 U.S.C. § 507(b). Thus, under the registration approach, a plaintiff who applied for registration towards the end of the three-year period could see the statute of limitations expire during the time it took the Copyright Office to act on the application. See Nimmer § 7.16[B][l][a][i] (“[Gjiven ... a three-year statute of limitations for recovery of all damages caused by copyright infringement, the narrow[registration] approach may indeed occasion complete inability to recover damages.... ”). This result does not square well with § 410(d)’s mandate that an application’s effective registration date should be the day that a completed application is received. In specifying that registration is to be dated as of the date of application— not approval — Congress sought to ensure that an applicant did not suffer the consequences of a delay by the Register. See H.R.Rep. No. 94-1476 at 157, 1976 U.S.C.C.A.N. at 5773 (noting that the provision is designed to “take[ ] account of the inevitable timelag between receipt of the application and other material and the issuance of the certificate.... ”). Only the *621application approach fully protects litigants from any disadvantage caused by this timelag.

Finally, we find unpersuasive the argument that deference to the Register requires adoption of the registration approach. This argument posits that Congress structured the Act to require the Register’s approval or rejection of registration before suit because it wanted the Register to determine the propriety of granting copyright registration in the first instance. See, e.g., Strategy Source, 233 F.Supp.2d at 4. However, as Nimmer explains, “the pace of litigation entails that the Copyright Office will typically have granted or refused registration during its pendency.” Nimmer § 7.16[B][l][a][i]. Thus, should the Register determine that it wanted to reject an application already in litigation, it would still have an opportunity to appear. Id. Furthermore, the Register’s decision of whether or not to grant a registration certificate is largely perfunctory, and is ultimately reviewable by the courts.13 There is, therefore, no compelling justification for delaying litigation until after the Register has acted on an application. The two processes — registration approval by the Copyright Office and an underlying infringement case in the courts — can occur simultaneously with little or no prejudice to any involved parties.14

We therefore hold that receipt by the Copyright Office of a complete application satisfies the registration requirement of § 411(a). This interpretation ensures the broad copyright protection that the 1976 Act provided. It “best effectuate[s] the interests of justice and promote[s] judicial economy.” Int’l Kitchen Exhaust Cleaning Ass’n v. Power Washers of North America, 81 F.Supp.2d 70, 72 (D.D.C. 2000). This approach also fully accomplishes the central purpose of registration — the compilation of a robust national register of existing copyrights' — -and at the same time avoids unfairness and waste of judicial resources. Cf. Lau Ow Bew v. United States, 144 U.S. 47, 59, 12 S.Ct. 517, 36 L.Ed. 340 (1892) (“Nothing is better settled than that statutes should receive a sensible construction, such as will effectuate the legislative intention, and, if possible, so as to avoid an unjust or an absurd conclusion.”); United States v. Leon H., 365 F.3d 750, 753 (9th Cir.2004) (citing Lau Ow Bew, 144 U.S. at 59, 12 S.Ct. 517).

(3)

Under the application approach, Cosmetic satisfied § 411(a)’s registration requirement before it instituted this action. Cosmetic alleged in its complaint that the Copyright Office received its complete ap*622plication on March 12, 2008, weeks before Cosmetic filed its complaint. Thus, § 411(a) does not bar Cosmetic’s infringement claim, which should proceed on its merits.

Conclusion

The judgment of the district court dismissing this action is VACATED and REVERSED. This case is REMANDED to the district court for further proceedings consistent with this opinion.

RESTANI, Judge:

Planetary Motion, Inc. (“Planetary Motion” or “Appellee”) sued Techsplosion, Inc. and Michael Gay a/k/a Michael Carson (respectively “Techsplosion” and “Carson”; collectively “Appellants”) for infringement and dilution of an unregistered trademark under Section 43(a) and (c) of the Federal Trademark Act, 15 U.S.C. § 1051 et seq. (1994) (“Lanham Act”), and for violation of Florida’s unfair competition law. Fla. Stat. Ann. § 495.151 (West 2000). Finding that Planetary Motion had established priority of use and a likelihood of confusion, the United States District Court for the Southern District of Florida entered summary judgment in favor of Planetary Mo*1191tion. We affirm the judgment and vacate the award of attorney fees.

Facts

I.Development and Distribution of the “Coolmail” Softivare

In late 1994, Byron Darrah (“Darrah”) developed a UNIX-based program (the “Software”) that provides e-mail users with notice of new e-mail and serves as a gateway to the users’ e-mail application. On December 31, 1994, Darrah distributed the Software over the Internet by posting it on a UNIX user site called “Sunsite,” from which it could be downloaded for free. Darrah had named the Software “Coolmail” and this designation appeared on the announcement sent to the end-users on Sunsite as well as on the Software user-manual, both of which accompanied the release.

The Software was distributed without charge to users pursuant to a GNU General Public License that also accompanied the release. A GNU General Public License allows users to copy, distribute and/or modify the Software under certain restrictions, e.g., users modifying licensed files must carry “prominent notices” stating that the user changed the files and the date of any change. After the release of the Software, Darrah received correspondence from users referencing the “Cool-mail” mark and in some cases suggesting improvements. In 1995, Darrah released two subsequent versions of the Software under the same mark and also pursuant to the GNU General Public License.

In early 1995, a German company named S.u.S.E. GmbH sought permission from Darrah to include the Software in a CD-ROM package sold as a compilation of Unix-based programs. Darrah consented and, pursuant to the GNU licensing agreement, S.u.S.E. distributed the .Software in its compilation product and in subsequent versions thereof. S.u.S.E. sold and continues to sell the software compilation in stores in the United States and abroad, as well as over the Internet.

II. Launch of Techsplosion’s “Cool-Mail” E-mail Service

In 1998, Appellant Carson formed Tech-splosion, for the purpose of operating a business based on an e-mail service that he had developed. On April 16, 1998, Techsplosion began offering the e-mail service on the Internet under the mark “CoolMail.” Two days later, Techsplosion activated the domain name “coolmaiLto” Techsplosion delivered an e-mail solicitation under the “CoolMail” mark to approximately 11,000 members of the Paramount Banner Network, an Internet advertising network, also created and operated by Carson. Techsplosion charged no fee to subscribe to the service and generated revenues through the sale of banner advertisements on its web site.

III. Planetary Motion’s E-mail Service & Application for Trademark Registration

Appellee Planetary Motion is a computer software and telecommunications company that developed and owns an electronic mail service called “Coolmail.” As part of its service, Planetary Motion enables a person to check e-mail via telephone without logging onto a computer. On April 24, 1998, Planetary Motion filed three intent-to-use applications to register the mark “Coolmail” with the United States Patent and Trademark Office. Though Planetary Motion was aware that Darrah’s Software also bore the mark “Coolmail,” it represented in its applications that it was not aware of any mark upon which its proposed registered mark would infringe. Planetary Motion launched its Coolmail email service to subscribers on June 8, 1998.

*1192IV. Planetary Motion’s Complaint and Subsequent Acquisition of Darrah’s Rights

On April 22, 1999, Planetary Motion filed a complaint against Techsplosion. In the complaint, Planetary Motion alleged infringement of the alleged mark “Cool-mail” for use in connection with e-mail services. Planetary alleged federal trademark infringement and unfair competition under Section 43(a) of the Lanham Act, 15 U.S.C. § 1125(a), as well as injury to business reputation and dilution under Florida Statute § 495.151.

On June 10, 1999, Techsplosion filed an Answer, Affirmative Defenses, and Counterclaims. The counterclaims alleged infringement of the mark “Coolmail” for use in connection with e-mail services. Tech-splosion alleged unfair competition, false designation, description, and representation under the Lanham Act, common trademark infringement, common law unfair competition, and injury to business reputation and dilution.

In July of 1999, Planetary Motion purchased from Darrah all rights, title, and interest to the Software including all copyrights, trademarks, patents and other intellectual property rights.1 On August 31, 1999, Planetary filed an Amended Verified Complaint, adding a claim for dilution under Section 43(c) of the Lanham Act, 15 U.S.C. § 1125(c), and alleging violation of trademark rights assigned from Darrah.

V. Disposition of Planetary Motion’s Complaint

On January 31, 2000, the district court entered an Order granting Planetary Motion’s motion for summary judgment and denying Carson’s and Techsplosion’s motion for summary judgment. The district court based the Order on two findings: (1) that the alleged mark was affixed to Darrah’s software, and that Darrah’s distribution of the software over the Internet constituted a “transport in commerce,” resulting in the creation of trademark rights and priority, and (2) there was a likelihood of confusion because the marks “are essentially the same.” The district court did not reach the issue of whether Tech-splosion’s use of “CoolMail” in connection with its e-mail service diluted Planetary Motion’s mark.

On the same date, the district court entered final judgment granting Planetary Motion permanent injunctive relief. See 15 U.S.C. § 1116. The order also awarded Planetary Motion profits and damages, as well as attorney fees and costs, pursuant to section 35 of the Lanham Act, 15 U.S.C. § 1117. The district court requested a report and recommendation from a magistrate judge fixing the amounts to be awarded.

A Notice of Appeal was filed on February 15, 2000. On February 15, 2000, Tech-splosion filed an Emergency Motion to Stay Pending Appeal, reasserting that Darrah never established any rights in the alleged mark. This motion was denied on February 17, 2000.

On May 9, 2000 the magistrate judge entered his report recommending that Planetary Motion be awarded $275,508 in attorneys’ fees and $6,562.34 in costs, but that its request for damages be denied for lack of specificity. Techsplosion served its appeal brief on May 22, 2000. On June 9, 2000, the district court entered an order adopting the report and recommendation in its entirety. On July 7, 2000, Techsplo*1193sion filed a Notice of Appeal from the order adopting the magistrate judge’s report and recommendation.2

Standard of Review

Review of a district court’s grant of summary judgment is de novo, with all facts and reasonable inferences therefrom reviewed in the light most favorable to the non-moving party. Carnival Brand Seafood Co. v. Carnival Brands, Inc., 187 F.3d 1307, 1309 (11th Cir.1999).

Discussion

Section 43(a) of the Lanham Act forbids unfair trade practices involving infringement of trade dress, service marks, or trademarks, even in the absence of federal trademark registration.3 Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768, 112 S.Ct. 2753, 120 L.Ed.2d 615 (1992). Section 43(a) is remedial in nature and should be interpreted and applied broadly so as to effectuate its remedial purpose. Montgomery v. Noga, 168 F.3d 1282, 1300 & n. 29 (11th Cir.1999) (citing Warner Bros., Inc. v. Gay Toys, Inc., 658 F.2d 76, 79 (2d Cir.1981)). To prevail under this section, a claimant must show (1) that it had prior rights to the mark at issue and (2) that the defendant had adopted a mark or name that was the same, or confusingly similar to its mark, such that consumers were likely to confuse the two.4 Lone Star Steakhouse & Saloon, Inc. v. Longhorn Steaks, Inc., 106 F.3d 355, 360 (11th Cir.1997) (citing Conagra Inc. v. Singleton, 743 F.2d 1508, 1512 (11th Cir.1984)), modified, 122 F.3d 1379 (1997). Appellants argue that the district court erred in finding that Planetary Motion had established both elements. Appellants also dispute the scope of injunctive relief, as well as the award of attorney fees and costs.

I. Prior Use in Commerce

Under common law, trademark ownership rights are “appropriated only through actual prior use in commerce.” Tally-Ho, Inc. v. Coast Community College Dist., 889 F.2d 1018, 1022 (11th Cir.1989) (citation omitted). Under the Lanham Act,5 the term “use in commerce” is *1194defined in relevant part as follows:

the bona fide use of a mark in the ordinary course of trade, and not made merely to reserve a right in a mark.... [A] mark shall be deemed to be in use in commerce ... on goods when (A) it is placed in any manner on the goods or their containers or the displays associated therewith or on the tags or labels affixed thereto, or if the nature of the goods makes such placement impracticable, then on documents associated with the goods or their sale, and (B) the goods are sold or transported in commerce ....

15 U.S.C. § 1127.6 The district court found that because the statute is written in the disjunctive (i.e., “sale or transport”), Darrah’s wide distribution of the Coolmail software over the Internet, even absent any sales thereof, was sufficient to establish ownership rights in the “CoolMail” mark. Appellants contend that “transport in commerce” alone — here, Darrah’s free distribution of software over the Internet “with no existing business, no intent to form a business, and no sale under the mark” — is insufficient to create trademark rights. Appellants’ Brief at 13. Appellants’ argument lacks merit.

The parties do not make clear the two different contexts in which the phrase “use in commerce” is used. The term “use in commerce” as used in the Lanham Act “denotes Congress’s authority under the Commerce Clause rather than an intent to limit the [Lanham] Act’s application to profit making activity.” United We Stand Am., Inc. v. United We Stand, Am. N.Y., Inc., 128 F.3d 86, 92-93 (2d Cir.1997) (citation omitted), cert. denied, 523 U.S. 1076, 118 S.Ct. 1521, 140 L.Ed.2d 673 (1998); U.S. Const., Art. I, § 8, cl. 3. Because Congress’s authority under the Commerce Clause extends to activity that “substantially affects” interstate commerce, United States v. Lopez, 514 U.S. 549, 559, 115 S.Ct. 1624, 131 L.Ed.2d 626 (1995), the Lanham Act’s definition of “commerce” is concomitantly broad in scope: “all commerce which may lawfully be regulated by Congress.” 15 U.S.C. § 1127. See also Steele v. Bulova Watch Co., 344 U.S. 280, 283-84, 73 S.Ct. 252, 97 L.Ed. 319 (1952); Larry Harmon Pictures Corp. v. Williams Rest. Corp., 929 F.2d 662, 666 (Fed.Cir.) (allowing registration for an intrastate provider of restaurant services with an undefined interstate clientele), cert. denied, 502 U.S. 823, 112 S.Ct. 85, 116 L.Ed.2d 58 (1991). The distribution of the Software for end-users over the Internet satisfies *1195the “use in commerce” jurisdictional predicate. See, e.g., Planned Parenthood Fed’n of Am., Inc. v. Bucci, 42 U.S.P.Q.2d 1430, 1434, 1997 WL 133313 (S.D.N.Y.1997) (“The nature of the Internet indicates that establishing a typical home page on the Internet, for access to all users, would satisfy the Lanham Act’s ‘in commerce’ requirement.”), aff'd, 152 F.3d 920 (2d Cir.), cert. denied, 525 U.S. 834, 119 S.Ct. 90, 142 L.Ed.2d 71 (1998).

Nevertheless, the use of a mark in commerce also must be sufficient to establish ownership rights for a plaintiff to recover against subsequent users under section 43(a). See New England Duplicating Co. v. Mendes, 190 F.2d 415, 417-18 (1st Cir.1951) (after finding “use in commerce” jurisdiction predicate satisfied, court noted that “[t]he question remains whether the plaintiff has established that he was the ‘owner’ of the mark, for under [15 U.S.C. § 1051] only the ‘owner’ of a mark is entitled to have it registered.”). The court in Mendes set forth a two part test to determine whether a party has established “prior use” of a mark sufficient to establish ownership:

[EJvidence showing, first, adoption,7 and, second, use in a way sufficiently public to identify or distinguish the marked goods in an appropriate segment of the public mind as those of the adopter of the mark, is competent to establish ownership, even without evidence of actual sales.8

Id. at 418. See also New West, 595 F.2d at 1200.9

Courts generally must inquire into the activities surrounding the prior use of the mark to determine whether such an association or notice is present. See, e.g., Johnny Blastoff, Inc. v. L.A. Rams Football Co., 188 F.3d 427, 433 (7th Cir.1999) (“The determination of whether a party has established protectable rights in a trademark is made on a case by case basis, considering the totality of the circumstances.”), cert. denied, 528 U.S. 1188, 120 S.Ct. 1241, 146 L.Ed.2d 100 (2000). Under the “totality of circumstances” analysis, a party may establish “use in commerce” even in the absence of sales. “[Although evidence of sales is highly persuasive, the question of use adequate to establish appropriation remains one to be decided on the facts of each case .... ” New West, 595 F.2d at 1200 (quoting Mendes, 190 F.2d at 418). The court in New West recognized that “mere advertising by itself may not establish priority of use,” but found that promotional mailings coupled with advertiser and distributor solicitations met the Mendes “public identification” ownership *1196requirement. Id. at 1200. Thus, contrary to Appellants’ assertions, the existence of sales or lack thereof does not by itself determine whether a user of a mark has established ownership rights therein.10 Compare Marvel Comics Ltd. v. Defiant, 837 F.Supp. 546, 549 (S.D.N.Y.1993) (finding announcement of “Plasmer” title to 13 million comic book readers and promotion at annual trade convention sufficient to establish trademark ownership rights, notwithstanding lack of any sales) with WarnerVision Entm’t Inc. v. Empire of Carolina Inc., 915 F.Supp. 639, 645-46 (S.D.N.Y.) (finding toy manufacturer’s promotional efforts insufficient to establish priority of use where only a few presentations were made to industry buyers, even though one resulted in a sale to a major toy retailer), aff'd in part, vacated in part, 101 F.3d 259 (2d Cir.1996).11

Similarly, not every transport of a good is sufficient to establish ownership rights in a mark. To warrant protection, use of a mark “need not have gained wide public recognition,” but “[s]ecret, undisclosed internal shipments are generally inadequate.” Blue Bell, Inc. v. Farah Mfg. Co., 508 F.2d 1260, 1265 (5th Cir.1975).12 In general, uses that are de minimis may not establish trademark ownership rights. See, e.g., Paramount Pictures Corp. v. White, 31 U.S.P.Q.2d 1768, 1772-73, 1994 WL 484936 (Trademark Tr. & App. Bd.1994) (finding no bona fide use in ordinary course of trade where mark was affixed to a game consisting of three pieces of paper and distributed for the purpose of promoting musical group).

We find that, under these principles, Darrah’s activities under the “Coolmail” mark constitute a “use in commerce” sufficiently public to create ownership rights in the mark. First, the distribution was widespread, and there is evidence that members of the targeted public actually associated the mark Coolmail with the Software to which it was affixed. Darrah made the software available not merely to a discrete or select group (such as friends and acquaintances, or at a trade show with limited attendance), but to numerous end-users via the Internet. The Software was posted under a filename bearing the “Cool-mail” mark on a site accessible to anyone who had access to the Internet. End-users communicated with Darrah regarding the Software by referencing the “Cool-mail” mark in their e-mails. Appellants argue that only technically-skilled UNIX-users made use of the Software, but there is no evidence that they were so few in *1197number to warrant a finding of de minim-is use.

Third, the mark served to identify the source of the Software. The “Coolmail” mark appeared in the subject field and in the text of the announcement accompanying each release of the Software, thereby distinguishing the Software from other programs that might perform similar functions available on the Internet or sold in software compilations.13 The announcements also apparently indicated that Dar-rah was the “Author/Maintainer of Cool-mail” and included his e-mail address. The user manual also indicated that the Software was named “Coolmail.”14 The German company S.u.S.E. was able to locate Darrah in order to request permission to use his Software in its product under the mark “Coolmail.” Appellants do not assert that S.U.S.E. was unaware that the Software was called “Coolmail” when it contacted Darrah.

Fourth, other potential users of the mark had notice that the mark was in use in connection with Darrah’s Software. In investigating whether the mark Coolmail existed before submitting its trademark registration application for its e-mail service, Planetary Motion was able to discover that Darrah was using the mark to designate his Software product.

Fifth, the Software was incorporated into several versions of a product that was in fact sold worldwide and specifically attributed ownership of the Software to Darrah under the “Coolmail” mark. Any individual using the S.U.S.E. product, or competitor of S.U.S.E., that wanted to know the source of the program that performed the e-mail notification function, could do so by referring to the user manual accompanying the product.15 There is no support for the argument that for a trademark in software to be valid, the mark must appear on the box containing the product incorporating it, that the mark must be displayed on the screen when the program is running, or that the software bearing the mark be a selling point for the product into which it is incorporated. There is no requirement that the public come to associate a mark with a product in any particular way or that the public be *1198passive viewers of a mark for a sufficient public association to arise.

Sixth, software is commonly distributed without charge under a GNU General Public License. The sufficiency of use should be determined according to the customary practices of a particular industry. See S. Rep. 100-515 at 44 (1988) (“The committee intends that the revised definition of ‘use in commerce’ [see note 13, supra] be interpreted to mean commercial use which is typical in a particular industry.”) (emphasis added). That the Software had been distributed pursuant to a GNU General Public License does not defeat trademark ownership, nor does this in any way compel a finding that Darrah abandoned his rights in trademark. Appellants misconstrue the function of a GNU General Public License. Software distributed pursuant to such a license is not necessarily ceded to the public domain and the licensor purports to retain ownership rights, which may or may not include rights to a mark.16

Appellants cite Heinemann v. General Motors Corp., 342 F.Supp. 203 (N.D.Ill. 1972), aff'd, 478 F.2d 1405 (7th Cir.1973), for the proposition that Darrah was a “hobbyist” unworthy of common law trademark protection. Heinemann is factually distinguishable from the case at hand. The plaintiff in Heinemann used a mark in connection with his automobile before an automobile manufacturer independently had adopted the same name for a new model. The court held that the plaintiff had not established common law ownership rights based on two findings. First, the court found that because Heinemann’s purpose in using the mark was to “open [at a later date] an automobile equipment shop which would have capitalized upon the slogan,” he merely attempted to “reserve a trade or service mark pending the creation of a trade or business .... ” 342 F.Supp. at 207. The court reasoned as follows:

While the law does not require a nationwide business; an old, established business; or even a profitable business for the acquisition of property interests in trade or service marks, it does require a presently existing trade or business for such acquisition. The exhibits disclose that Plaintiff had only a desire to open a business in futuro. To hold otherwise would make a trade mark a property right in gross, instead of a right appurtenant.

Id. (emphasis in original). The Heinemann court also found that plaintiff Heinemann’s activities consisted merely of occasionally racing or displaying the automobile at fairs as a hobby, as evidenced by his testimony that he was employed at an oil company. Id. Here, Darrah did not attempt to “warehouse” the mark by promoting a product he merely intended to develop and distribute at a later date. Darrah’s use of the mark to designate the distributed Software and each subsequent version thereof indicates that his use was not mere sporadic or token use.17 Fur*1199thermore, unlike Heinemann, Darrah activities pertained to his chosen profession. Darrah is employed as a computer systems administrator, which entails the management and oversight of computer networks and systems as well as the development of software in support thereof.

Appellants also rely on DeCosta v. Columbia Broad. Sys., Inc., 520 F.2d 499, 513 (1st Cir.1975), cert. denied, 423 U.S. 1073, 96 S.Ct. 856, 47 L.Ed.2d 83 (1976), to argue that Darrah is an eleemosynary individual and therefore unworthy of protection under unfair competition laws. The DeCosta court did not hold that the that the absence of a profit-oriented enterprise renders one an eleemosynary individual, nor did it hold that such individuals categorically are denied protection. Rather, the DeCosta court expressed “misgivings” of extending common law unfair competition protection, clearly available to eleemosynary organizations, to eleemosynary individuals.18 Id. The court’s reluctance to extend protection to eleemosynary individuals was based on an apparent difficulty in establishing a line of demarcation between those eleemosynary individuals engaged in commerce and those that are not. But as the sufficiency of use to establish trademark ownership is inherently fact-driven, the court need not have based its decision on such a consideration. Mendes, 190 F.2d at 418. Common law unfair competition protection extends to non-profit organizations because they nonetheless engage in competition with other organizations.19 See Girls Clubs of Am., Inc. v. Boys Clubs of Am., Inc., 683 F.Supp. 50 (S.D.N.Y.1988), aff'd, 859 F.2d 148 (2d Cir.). Thus, an eleemosynary individual that uses a mark in connection with a good or service may nonetheless acquire ownership rights in the mark if there is sufficient evidence of competitive activity.20

*1200Here, Darrah’s activities bear elements of competition, notwithstanding his lack of an immediate profit-motive. By developing and distributing software under a particular mark, and taking steps to avoid ceding the Software to the public domain, Darrah made efforts to retain ownership rights in his Software and to ensure that his Software would be distinguishable from other developers who may have distributed similar or related Software. R2-47-Exh. 3 at 67. Competitive activity need not be fueled solely by a desire for direct monetary gain. Darrah derived value from the distribution because he was able to improve his Software based on suggestions sent by end-users. Just as any other consumers, these end-users discriminate among and share information on available software. It is logical that as the Software improved, more end-users used his Software, thereby increasing Darrah’s recognition in his profession and the likelihood that the Software would be improved even further.

In light of the foregoing, the use of the mark in connection with the Software constitutes significant and substantial public exposure of a mark sufficient to have created an association in the mind of public.

II. Likelihood of Confusion

The district court supported its determination of “likelihood of confusion” with the following findings: (1) the mark used by Planetary Motion (“Coolmail”) is “essentially the same” as that used by Techsplosion (“CoolMail”)21; (2) both marks are used in connection with e-mail services; (3) both plaintiff and defendants serve e-mail customers via the Internet; and (4) both use the Internet to promote their services. R2-62-7.22 Appellants do not dispute the accuracy of these findings. Rather, Appellants claim the district court improperly based its analysis on a comparison of the parties’ respective e-mail services, rather than on a comparison of Techsplosion’s “CoolMail” e-mail service to the “Coolmail” Software. Appellants argue that the latter comparison is required *1201because Planetary Motion acquired its rights to the “Coolmail” mark by assignment of rights in Darrah’s Software, and under the “natural expansion” doctrine this is a use unrelated to Planetary Motion’s current use of e-mail services. Thus, Appellants’ argument in essence goes to whether the scope of trademark protection enjoyed by Planetary Motion extends from the initial use (in connection with the Software) to the current use (in connection with Planetary Motion’s e-mail services).

The scope of protection enjoyed by a trademark owner is not restricted to the owner’s original use. The “natural expansion” doctrine is applied to determine the proper scope of protection where a mark owner’s previous use differs from its current use, and the junior use intervenes. Under this doctrine, the first trademark owner’s rights are limited to goods on which the mark has already been used or that lie within the realm of natural expansion; “[t]his appears to be no more than a specific application of the familiar ‘related goods’ test.” J. McCarthy, § 24:20.23 See also Carnival, 187 F.3d at 1310-11.

The court in Tally-Ho explained that a senior user’s rights “may extend into uses in ‘related’ product or service markets (termed the ‘related goods doctrine’),” and that “an owner of a common law trademark may use its mark on related products or services and may enjoin a junior user’s use of the mark on such related uses .... ” 889 F.2d at 1023 (citing J. McCarthy, § 24:1 to 24:12). This rule is limited by equitable considerations. The court in Carnival noted that “[A] trademark owner cannot by the normal expansion of its business extend the use or registration of its mark to distinctly different goods or services not comprehended by its previous use ... where the result could be a conflict with valuable intervening rights established by another through extensive use ... of the same or similar mark for like or similar goods and services.” 187 F.3d at 1310-11 (citations and internal quotation marks omitted) (emphasis added).

Courts determine the proper scope of protection of a mark in the context of intervening uses by applying the “source or sponsorship” test. Under this test, a trademark owner has “protection against use of its mark on any product or service which would reasonably be thought by the buying public to come from the same source, or thought to be affiliated with, connected with, or sponsored by, the trademark owner.” J. McCarthy, § 24:6. The public perception in this regard is determined at the time the junior user first used the mark on the product or service to which the allegedly infringing mark is affixed.24 Carnival, 187 F.3d at 1312. The court in Tally-Ho explained *1202that “related use” is “merely a facet of the likelihood of confusion test and therefore requires an inquiry into [the] seven factors affecting the likelihood of confusion ,...”25 889 F.2d at 1027.

We find that the relatedness between e-mail notification software and a service that allows users to check e-mail via telephone line is not so attenuated that the district court mistakenly failed to limit Planetary Motion’s rights in the mark to its use in connection with the Software. Consumers reasonably could attribute the Software and an e-mail service under the same name to the same source for several reasons.26 Both the Software and the e~ mail service belong to the same general field of commerce, i.e., information teehnol-ogy, and both deal more specifically with e-mail.27 Appellants do not dispute Planetary Motion’s contention that major firms in this field sell e-mail software as well as provide e-mail service.28 See Darrah Affidavit, R2-46-Exh. 2 at ¶ 34. Both the Software and the e-mail service involve sending messages over the Internet: the former provides a functionality that enables a user with e-mail capability to receive notification of new mail, and the latter enables subscribers to the service to send and receive e-mail. Both the Software and the e-mail service were promoted over the Internet to those who make use of e-mail.

Furthermore, the equities do not necessarily favor Techsplosion.29 Techsplosion’s *1203“CoolMail” e-mail service had not been in operation for an extended period of time before Planetary Motion entered the market under the name “Coolmail,” and Planetary Motion is not merely attempting to reserve the mark for a future business endeavor. Accordingly, we sustain the district court’s finding of “likelihood of confusion.”

III. Relief

Review of the district court’s award of injunctive relief, attorney’s fees and costs is for abuse of discretion. See Burger King Corp. v. Weaver, 169 F.3d 1310, 1315 (11th Cir.), cert. dismissed, 528 U.S. 948, 120 S.Ct. 370, 145 L.Ed.2d 287 (1999); Tally-Ho, 889 F.2d at 1022.

A. Injunctive Relief

Appellants assert that the injunctive relief awarded by the district court is imper-missibly vague and overbroad. Appellants contend that the language “imposes a grossly unfair burden on Techsplosion, as it does not allow for a determination of what is and is not permitted under the injunctive provisions.” Appellants’ Brief at 40. Federal Rule of Civil Procedure 65(d) states in relevant part: “Every order granting an injunction ... shall be specific in terms [and] shall describe in reasonable detail ... the act or acts sought to be restrained ....”30 Fed.R.Civ.P. 65(d). A clear and unambiguous order is one that leaves “ ‘no uncertainty in the minds of those to whom it is addressed,’ ... who must be able to ascertain from the four corners of the order precisely what acts are forbidden.’ ” King v. Allied Vision, Ltd., 65 F.3d 1051, 1058 (2d Cir.1995) (citations omitted). Optimally, the injunction

should clearly let defendant know what he is ordered to do or not to do. A court order should be phrased in terms of objective actions, not legal conclusions. An injunction which merely forbids a defendant from performing “acts of unfair competition,” or from “infringing on plaintiffs trademarks and trade secrets” adds nothing to what the law already requires.

John H. Harland Co. v. Clarke Checks, Inc., 711 F.2d 966, 984-985 (11th Cir.1983) (quoting J. McCarthy, at § 30.13) (emphasis added).

Notwithstanding these strictures, appellate courts do not set aside injunctions under Rule 65(d) “unless they are so vague that they have no reasonably specific meaning.” E. & J. Gallo Winery v. Gallo Cattle Co., 967 F.2d 1280, 1297 (9th Cir.1992). Rather than applying Rule 65(d) rigidly, appellate courts determine the propriety of an injunctive order by inquiring into whether the parties subject thereto understand their obligations under the order. Williams v. City of Dothan, 818 F.2d 755, 761 (11th Cir.1987) (citing Combs v. Ryan’s Coal Co., 785 F.2d 970, 978-79 (11th Cir.), cert. denied sub nom. Sim *1204mons v. Combs, 479 U.S. 853, 107 S.Ct. 187, 93 L.Ed.2d 120 (1986)), modified, 828 F.2d 13 (11th Cir.1987). Furthermore, the degree of particularity required depends on the nature of the subject matter. McComb v. Jacksonville Paper Co., 336 U.S. 187, 191-92, 69 S.Ct. 497, 93 L.Ed. 599 (1949) (decrees of generality are often necessary to prevent further violations where a proclivity for unlawful conduct has been shown).

Here, although several parts of the order are phrased in terms of legal conclusions, the order, read as a whole, clearly indicates what Techsplosion is forbidden from doing.31 According to the order, Techsplosion is permanently enjoined from using the name “Coolmail” in connection with “e-mail or other Internet-related services,” in connection with software, and as part of the domain name of its website. These limitations sufficiently define the scope of the injunction and give context to the arguably legally conelusory language included therein. Furthermore, the use of “any similar mark” after references to “Coolmail” does not detract from this level of specificity. See Planned Parenthood, 152 F.3d 920 (upholding preliminary injunction that prohibited defendant from using mark or “any colorable imitation of [that] mark” anywhere on the Internet and from taking actions “likely to cause confusion” among Internet users regarding Planned Parenthood’s endorsement or sponsorship of defendant’s site). Appellants do not propose any reading of the language in the injunction that would impermissibly prohibit conduct that Tech-splosion arguably has the right to do. Cf. B.H. Bunn Co. v. AAA Replacement Parts Co., 451 F.2d 1254, 1269 (5th Cir.1971) (finding injunction impermissible where broad reference to opinion’s findings of fact “[appeared to] enjoin perfectly legal acts”). But see AmBrit, Inc. v. Kraft, Inc., 812 F.2d 1531, 1547-48 (11th Cir.1986) (“An injunction can be therapeutic as well as protective. In fashioning relief against a party who has transgressed the governing legal standards, a court of equity is free to proscribe activities that, standing alone, would have been unassailable.”) (citation omitted), cert. denied, 481 U.S. 1041, 107 S.Ct. 1983, 95 L.Ed.2d 822 (1987). Nor do Appellants explain how the *1205wording of the injunction could lend itself to alternate interpretations. In the absence of any apparent contextual ambiguities or alternate readings prohibiting legal conduct, it is unlikely that Appellants will misapprehend what conduct is proscribed, or will incur liability to contempt citations for activities not contemplated by this order.

B. Attorney Fees

Awards of attorney fees are reviewable only to determine if the trial court abused its discretion in granting or denying them.32 See St. Charles Mfg. Co. v. Mercer, 737 F.2d 891, 894 (11th Cir.1983). Fla. Stat. § 495.151 provides for injunctive relief only. Under Section 35(a) of the Lanham Act, however, courts may award reasonable attorney fees to the prevailing party “in exceptional cases.” 15 U.S.C. § 1117(a). See also Montgomery, 168 F.3d at 1304. Exceptional cases are those where the infringing party acts in a “malicious, fraudulent, deliberate, or willful manner.” Burger King Corp. v. Pilgrim’s Pride Corp., 15 F.3d 166, 168 (11th Cir.1994) (quoting S.R. Rep. 93-1400 (1974), reprinted in 1974 U.S.C.C.A.N. 7132, 7133) (internal quotation marks omitted).

Here, the district court awarded attorney fees without articulating a basis for doing so, let alone the factual circumstances that would warrant such an award.33 Furthermore, there is nothing in the record to support a finding of “malicious, fraudulent, deliberate, or willful” conduct on the part of Planetary motion. Remand is therefore unnecessary on this issue. Accordingly, we find that the award of attorney fees is an abuse of discretion and vacate the award.

C. Award of Costs

Under the Lanham Act, a successful party “subject to the principles of equity” may recover: “(1) defendant’s [the infringer’s] profits; (2) any damages sustained by the plaintiff, and (3) the costs of the action.” 15 U.S.C. § 1117(a). See Babbit Elecs., Inc. v. Dynascan Corp., 38 F.3d 1161, 1182 (11th Cir.1994). To recover costs under the Lanham Act, therefore, a plaintiff need not establish that the alleged infringer acted maliciously, fraudulently, deliberately, or willfully. Because it was within the district court’s discretion to award costs, and in the absence of any evidence that would warrant a finding that the balance of equities necessarily tipped in favor of Techsplosion, we uphold the award of costs to Planetary Motion.

Conclusion

Accordingly, the district court’s order and final judgment are AFFIRMED, except that the order adopting the magistrate judge’s report and recommendation, with respect to the award of attorney fees, is VACATED.