Brian E. BATEMAN; Charles H. Fricker, Plaintiffs-Counter-Defendants, Appellees-Cross-Appellants, v. MNEMONICS, INC.; Harry Thompson, President; David Katz, Vice President; Parking Automation Corporation; Joe Guidage, President, et al., DefendantsCounter-Claimants, Third-party-Plaintiffs, Appellants-Cross-Appellees,

City of Miami, an entity organized and existing under the laws of the State of Florida; Department of Off-Street Parking of the City of Miami, an agency and instrumentality of the City of Miami, Florida; Generex Corporation, a Florida Corporation, Third-party-Defendants.

No. 93-3234.

United States Court of Appeals, Eleventh Circuit.

March 22, 1996.

Rehearing Denied May 30, 1996.

Michael Kahn, Michael Kahn, PA. Melbourne, Florida, Lawrence K. Nodine, Needle & Rosenberg, Atlanta, GA, for appellants. Peter M.C. Choy, Mountain View, CA, for American Committee for Interoperable Systems. Herbert L. Allen, Allen, Dyer, Doppelt, Franjola & Milbrath, Orlando, FL, for appellees.

Appeals from the United States District Court for the Middle District of Florida.

ON PETITION FOR REHEARING Before BIRCH and DUBINA, Circuit Judges, and JOHNSON, Senior Circuit Judge.

BIRCH, Circuit Judge: Pursuant to Federal Rule of Appellate Procedure 40(a), the panel, hereby grants the appellees' petition for rehearing, withdraws the previous panel opinion dated December 22, 1995, and substitutes the following opinion:

This appeal requires us to address two issues of first impression in this circuit: (1) what methodology is to be employed in analyzing claims of copyright infringement of computer software, and (2) whether interface specifications are entitled to copyright protection. For the reasons that follow, we VACATE the judgment of the district court on the two copyright infringement counts and REMAND for a new trial on these counts. In addition, we REVERSE the judgment of the district court on the trade secret misappropriation count and instruct the district court to enter judgment as a matter of law for the appellants on this count.

I. BACKGROUND Appellees-cross-appellants Brian E. Bateman and Charles H. Fricker filed an action in federal court against appellants-cross-appellees Mnemonics, Inc., Harry Thompson, David Katz, Parking Automation Corp., BCS, Inc. ("BCS"), and Robert Brunet (collectively "PAC"). The second amended complaint alleged copyright infringement in Bateman's single board computer operating system ("SBCOS") software (Count I), copyright infringement in Bateman and Flicker's hardware logic diagrams for their computer circuit board (Count II), copyright infringement in the programmable array logic ("PAL") software (Count III), false designation of origin in violation of 15 U.S.C. § 1125(a) (Count IV), common law unfair competition (Count V), and theft of trade secrets (Count VI).^{Footnote 1}

After a lengthy trial, the jury found for Bateman and Fricker on three counts, awarding them $105,000 on Count I, $120,000 on Count II, and $300,000 on Count VI.^{Footnote 2}

PAC appeals these judgments, and Bateman and Fricker cross-appeal, asserting that the district court abused its discretion in refusing to award exemplary damages to them on the trade secret claim under Fla.Stat. § 688.004(2).

Before addressing the many issues involved in this appeal, it is important to set forth the factual and technical background to this complex case. Both Bateman and Fricker are engineers. Bateman created and developed SBCOS software, and the Register of Copyrights registered his claim of copyright in this software.^{Footnote 3} Bateman and Fricker together created and developed hardware logic diagrams for a single board computer ("SBC") used in automated parking systems, and the Register of Copyrights registered their claim of copyright in these hardware logic diagrams. ^{Footnote 4}

Initially, Bateman had a business relationship with BCS and its principal, Robert Brunet, ^{Footnote 5} in the mid-1980s. BCS manufactured and sold computer boards to several companies, including Generex Corporation. ^{Footnote 6}

Generex was engaged in the parking system business and desired to purchase a computer board and operating system for use in its parking garage business. To satisfy this need, Bateman and Flicker helped Brunei design a SBC, which contained Bateman's SBCOS software as one of its components. ^{Footnote 7}

Once the SBC was fully developed. BCS sold a number of these boards to Generex. ^{Footnote 8}

After this purchase, Generex's programmers composed an application program designed to interoperate with the SBCOS. ^{Footnote 9}

To ensure compatibility, Generex's programmers relied on specifications delivered to them by BCS ^{Footnote 10} that dictated the "system calls" necessary to communicate with the operating system, the SBCOS. ^{Footnote 11}

In 1987, Generex was acquired by PAC, a wholly owned subsidiary of Mnemonics, Inc. Pursuant to the BCS/Generex agreement, BCS began delivering SBCs to PAC. In the spring of 1988, Bateman terminated BCS's right to use his SBCOS software on the SBCs that it was selling. ^{Footnote 12}

Bateman contact-

ed Joe Guidage, the president of PAC, to notify him that BCS was no longer authorized to sell the SBC boards containing his SBCOS. Guidage and Bateman decided to do business with one another and apparently bypassed BCS and Brunet. In May of 1988, Bateman negotiated a written agreement with Guidage whereby he and Fricker agreed to design a new SBC, the SBC2, which was to contain an updated version of Bateman's SBCOS software. ^{Footnote 13}

In addition, the SBC2 was to include programmable array logic ("PAL") technology, which was not included on the SBC1. ^{Footnote 14}

Days after the May agreement between Guidage and Bateman was signed, it was repudiated by Thompson, president of Mnemonics, PAC's parent company. A new contract was negotiated and signed in June of 1988, under which Bateman and Fricker agreed to design the SBC2, which would meet certain of PAC's technical requirements. Bateman and Fricker delivered the SBC2 to PAC in June of 1988. In addition, in November of 1988, pursuant to the June contract, Fricker delivered to PAC "engineering" for the board; these materials were needed in order for PAC to be able to build SBC2s. ^{Footnote 15}

The hardware logic diagrams at issue in this case were among the documents delivered to PAC by Fricker in November of 1988. Bateman and Fricker acknowledge that there was no copyright notice on the diagrams that were delivered to PAC.

Shortly after delivering the engineering to PAC, Bateman and Fricker realized that they had given PAC their only copy of these materials. Because these materials were needed to manufacture more boards, Bateman and Fricker needed to retrieve them. Fricker received back the engineering materials from PAC, signing a receipt to indicate that PAC had loaned the materials to him; Fricker later returned the materials to PAC. It is these engineering materials that are at issue in Bateman and Fricker's theft of trade secret count against PAC. ^{Footnote 16}

By January of 1989, the business relationship between Bateman/Fricker and PAC had deteriorated severely. PAC experienced problems with the SBC2s that were in the field, and the parties disputed the cause of the problems. On January 18, 1989, Bateman and Fricker met with Jack Blalock, the

general manager of PAC. PAC's payments were in arrears, while Blalock complained to Bateman and Flicker that their product was not satisfactory. Blalock requested that Bateman and Flicker deliver to PAC the "source code" ^{Footnote 17} for the SBCOS and the rights to the SBC2 circuit design in exchange for a release of claims based on the inoperability of the delivered SBC2s, for which PAC had paid Bateman and Fricker $35,000. Bateman and Fricker refused to accept PAC's offer, and, on January 26, 1989, Blalock dispatched a letter to Bateman terminating the contract between Bateman/Fricker and PAC.

After Bateman and Fricker denied PAC's request for the source code for the SBCOS and the rights to the SBC2 circuit design, PAC undertook to compose its own operating system that would interoperate with the application program that it (or its predecessor, Generex) had written to be compatible with Bateman's SBCOS. Tom Colvin, Mnemonic's senior engineer, was assigned this task. He disassembled and decompiled a portion of Bateman's SBCOS from a computer chip resident on one of the previously delivered SBC2s. ^{Footnote 18}

PAC claims that Colvin was compelled to do this because Bateman had never provided PAC with a

copy of the source code for his SBCOS. PAC also claims that Colvin identified those elements of the program that were necessary for compatibility with the preexisting application program and incorporated them into PAC's new operating system, known as the Lane Control Computer Operating System ("LCCOS"). In addition to composing the LCCOS, PA.C also designed a new circuit board, known as the Lane Control Computer ("LCC"), which was to replace Bateman and Flicker's SBC2.

Shortly after the January contract termination by PAC, Bateman and Fricker filed suit in Florida state court, alleging breach of the June 2, 1988 contract. ^{Footnote 19}

They subsequently learned that PAC was marketing a new computer circuit board that appeared to perform the same functions as their SBC2. In May of 1990, Bateman and Fricker initiated this case in federal court in Florida, alleging, inter alia, copyright infringement and theft of trade secrets.

II. ISSUES ON APPEAL On appeal, PAC raises eight issues, and Bateman and Fricker present one issue on cross-appeal. The issues on direct appeal are whether: (1) a jury instruction that limited the "successive filtration" test for copyright infringement to nonliteral similarity was confusing or misleading; (2) the district court should have instructed the jury to filter out the operating system/application program interface; (3) it was plain error to instruct the jury that a "qualitatively" small amount of copying could authorize a finding of copyright infringement; (4) the district court erred in directing a verdict striking PAC's omission of copyright notice defense because it found that one copy was a "relatively small number" under the Copyright Act, as a matter of law; (5) the district court erred in denying PAC's motions for directed verdict and JNOV as to claims of copyright infringement in hardware logic diagrams, where the only evidence of damages was the sale of purely utilitarian circuit boards built from them; (6) the district court erred in denying PAC's motions for directed verdict and JNOV as to damages based on sales of circuit boards where Bateman and Fricker failed to introduce evidence that "translations" of their schematics purportedly used in manufacturing circuit boards contained any protectable expression; (7) the district court should have vacated the jury's award of $120,000 in damages for infringement of the hardware logic diagram's copyright, where there was no evidence that any boards were manufactured after PAC had notice of the copyright claim; and (8) the district court erred in denying PAC's motions for directed verdict and JNOV on the state law trade secret claim. The sole issue on cross-appeal is whether the district court abused its discretion in refusing to award exemplary damages on Bateman and Fricker's trade secret claim of Count VI under Fla.Stat. § 688.004(2), given the evidence that supported the jury verdict in Bateman and Fricker's favor on this count.

III. DISCUSSION

A. The Feist Two-Pronged Test for Copyright Infringement At the outset, we review some applicable basic principles. To establish a claim of copyright infringement, Bateman and Flicker must prove "(1) ownership of a valid copyright, and (2) copying of constituent elements of the work that are original." .Feist Publications, Inc. v. Rural Tel. Serv. Co.,  499 U.S. 340, 361, 111 S.Ct. 1282, 1296, 113 L.Ed.2d 358 (1991); see also BellSouth Advertising & Publishing Corp. v. Donnelley Info. Publishing, Inc.,  999 F.2d 1436, 1440 (11th Cir.1993) (en bane), cert, denied, ' U.S. ------, 114 S.Ct. 943, 127 L.Ed.2d 232 (1994). To "satisfy Feist's first prong, a plaintiff must prove that the work ... is original and that the plaintiff complied with applicable statutory formalities." Lotus Dev. Corp. v. Borland Int% Inc.,  49 F.3d 807, 813 (1st Cir.1995), affd by an equally divided Court, ' U.S. -----, 116 S.Ct. 804, 133 L.Ed.2d 610 (1996) (citation omitted). In judicial proceedings, a "certificate of registration made before or within five years after first publication of the work shall constitute prima facie evidence of the validity of the copyright and of the facts stated in the certificate." 17 U.S.C. § 410(c) (1977). ^{Footnote 20}

Once the plaintiff produces a certificate of copyright, the burden shifts to the defendant to demonstrate why the claim of copyright is invalid. Bibbero Sys., Inc. v. Colwell Sys., Inc.,  893 F.2d 1104, 1106 (9th Cir.1990). At this juncture, it is incumbent upon a putative infringer to establish that the work in which copyright is claimed is unprotectable (for lack of originality) or, more specifically, to prove that the portion of the copyrighted work actually taken is unworthy of copyright protection. ^{Footnote 21}

In the case before us, Bateman has a registration of his claim of copyright in his SBCOS software, and Bateman and Flicker jointly have been issued a copyright registration in the hardware logic diagrams for the SBC2. PAC does not contest the validity of these copyright registrations and thus Feist's first prong is satisfied.

It is Feist's second prong that is a central focus of this case. To prove actionable copying, the plaintiff must first establish, as a factual matter, that the alleged infringer "actually used the copyrighted material to create his own work." Engineering Dynamics, Inc. v. Structural Software, Inc.,  26 F.3d 1335, 1340 (5th Cir.1994). Proof of copying may be shown either by direct evidence of the copying or, in the absence of such evidence, "[c]opying as a factual matter typically may be inferred from proof of access to the copyrighted work and 'probative similarity.'" Id. (citing Plains Cotton Co-op.

Ass'n v. Goodpasture Computer Serv., Inc.,  807 F.2d 1256, 1260 (5th Cir.), cert, denied, 484 U.S. 821, 108 S.Ct. 80, 98 L.Ed.2d 42 (1987)). Factual proof of copying, however, is only an element in satisfying the second prong of Feist.

It is at this point that the plaintiff must also respond to any proof advanced by the defendant that the portion of the copyrighted work actually taken does not satisfy the constitutional requirement of originality as set forth in Article I, § 8, cl. 8. See also Feist, 499 U.S. at 345-46, 111 S.Ct. at 1287-88 (noting that "[t]he sme qua non of copyright is originality," as well as emphasizing that it is a "constitutional requirement"). If the plaintiff successfully demonstrates that the portion of the copyrighted work taken satisfies the originality requirement, he must still, depending upon the proof of the defendant and the defenses asserted, surmount two additional hurdles. First, the plaintiff must prove that "the copying of copyrighted material was so extensive that it rendered the offending and copyrighted works substantially similar." Lotus, 49 F.3d at 813; see also Engineering Dynamics, 26 F.3d at 1341. Second, he may have to successfully challenge the putative infringer's evidentiary assertion that the use made of the original portion of the copyrighted work is a fair use under 17 U.S.C. § 107, where the putative infringer attempts to carry his burden and demonstrate a right to such use. ^{Footnote 22}

In this appeal, both the originality/substantial similarity aspect of .Feist's second prong and fair use are at issue, although we note that proof of originality can be the more problematic of the two hurdles to overcome; while there may be evidence of copying, not all copying is legally actionable. See Engineering Dynamics, 26 F.3d at 1341; BellSouth, 999 F.2d at 1444-45.

For the reasons discussed, we need not consider I^eist 's first prong. The factual copying aspect of Feist's second prong is not at issue either, because PAC admits that it copied portions of the SBCOS code and the SBC2 logic circuit diagrams during the design and development of the LCCOS and the LCC. Thus, the two matters directly at issue on appeal as to Count I are whether: (1) the copying done by PAC is legally actionable; that is, whether there is "substantial similarity" between the allegedly offending program and the protectable, original elements of the copyrighted works; and (2) the use of any such original elements qualifies as a fair use under 17 U.S.C. § 107. ^{Footnote 23}

At the heart of these issues is the

parties' dispute over the district court's jury instructions on literal and nonlitera] copying, to which we now turn.

B. The Jury Instructions PAC raises three separate issues on appeal concerning the district court's instructions to the jury. "We apply a deferential standard of review to a district court's jury instructions. So long as the instructions accurately reflect the law, the trial judge is given wide discretion as to the style and wording employed in the instructions." United States v. Starke,  62 F.3d 1374, 1380 (11th Cir.1995) (citing McElroy v. Firestone Tire & Rubber Co.,  894 F.2d 1504, 1509 (11th Cir.1990)). "On appeal, we examine whether the jury charges, considered as a whole, sufficiently instructed the jury so that the jurors understood the issues and were not misled." Wilkinson v. Carnival Cruise Lines, Inc.,  920 F.2d 1560, 1569 (11th Cir.1991) (citations omitted). Under this standard, if the jury charge as a whole correctly instructs the jury, even if it is technically imperfect, no reversible error has been committed. Miller v. Universal City Stvdios, Inc.,  650 F.2d 1365, 1372 (5th Cir.1981). We must reverse an erroneous instruction, however, if we are "left with a substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations." Johnson, v. Bryant,  671 F.2d 1276, 1280 (11th Cir.1982) (quoting Miller v. Universal City Studios, Inc.,  650 F.2d 1365, 1372 (5th Cir.1981)). Instructing a jury in the application of the law of copyright, which is characterized by often subtle and contradictory distinctions, is indeed a daunting task. Thus, the need for proper guidance is accentuated and heightened in directing a jury in this area of the law, which may be foreign to many jurors, particularly in this rapidly changing technological world. Applying the aforementioned standard of review, we examine seriatim the three issues on appeal relating to the jury instructions.

1. Limiting the "Abstraction-FiltrationComparison" Test to Nonliteral Similarity

PAC asserts that the district court committed reversible error by instructing the jury to filter out only nonliteral similarities in applying the "abstraction-filtrationcomparison" test for substantial similarity that the Second Circuit adopted in Computer Assocs. Int'l v. Altai Inc.,  982 F.2d 693 (2d Cir.1992). ^{Footnote 24}

PAC requested jury instructions that allowed for filtering out all nonprotectable (that is, unoriginal) expression, without regard to whether the similarity existed on the literal or nonliteral level. ^{Footnote 25}

The district

court denied PAC's request and accepted instead Bateman's proposed instructions, which confined the filtration analysis only to instances of nonliteral copying. By limiting the filtration instruction to nonliteral similarities, PAC claims that the district court effectively rendered futile its efforts to rebut Bateman's evidence of literal similarity with regard to elements of the work for which copyright was claimed, thereby nullifying several of its key defenses, including compatibility, efficiency, and standard programming techniques. Bateman and Flicker, however, assert that the district court was correct in limiting the filtration instruction to nonliteral similarities. The Second Circuit in Altai summarized its abstraction-filtration-comparison test as follows: In ascertaining substantial similarity under this approach, a court would first break down the allegedly infringed program into its constituent structural parts. Then, by examining each of these parts for such things as incorporated ideas, expression that is necessarily incidental to those ideas, and elements that are taken from the public domain, a court would then be able to sift out all non-protectable material. Left with a kernel, or possible kernels, of creative expression after following this process of elimination, the court's last step would be to compare this material with the structure of an allegedly infringing program. The result of this comparison will determine whether the protectable elements of the programs at issue are substantially similar so as to warrant a finding of infringement. Altai 982 F.2d at 706. In the case at bar, it is Altai's second step, filtration, that most concerns us. Filtration is used to separate protectable expression from nonprotectable material. It is at this step that the structural components at each level of abstraction are examined "to determine whether their particular inclusion at that level was 'idea' or was dictated by considerations of efficiency, so as to be necessarily incidental to that idea; required by factors external to the program itself; or taken from the public domain and hence is nonprotectable expression." Altai, 982 F.2d at 707. It is here that PAC claims that merger, scenes afaire, and other originality challenges to literal copying should be applied.

The district court's instruction limited the filtration analysis to nonliteral similarity, stating that: Under the Nimmer test, substantial similarity of the non-literal elements is determined by comparing with the defendants' program, that protectable expression of the copyrighted work which remains after filtering out any portion of the copyrighted work, which represents only ideas, elementsf ] dictated solely by logic and efficiency, elements dictated by hardware or software standards, computer industry programming and practices or elements which are taken from the public domain.

Even a qualitatively [sic] small amount of copied material which remains unfiltered may be sufficiently important to the operation of Mr. Bateman's operating system program to justify a finding of substantial similarity. For instance, a small portion of a structure or code of a program may nonetheless give it its distinctive features. In such a case, a finding of substantial similarity would be appropriate.

Supp. Rl-318-92-93. This instruction is a verbatim version of Bateman's proposed instruction on determining substantial similarity under the Nimmer test. At the time that this instruction was offered by Bateman,

PAC submitted its own version of a filtration instruction, one that would direct the jury to filter out all nonprotectable expression, without regard to whether the similarity existed on the literal or nonliteral level. PAC objected to Bateman's instruction when first offered to the court, Supp. Rl-330-325, and also objected to the court's decision to use Bateman's instruction, which contained the nonliteral limitation, rather than PAC's version of the filtration analysis. Supp. Rl332-74. Over both of these objections, the district court limited the filtration instruction to nonliteral similarities.

We recognize that the Altai test was formulated by the Second Circuit to help "determine the scope of copyright protection that extends to a computer program's nonliteral structure." Altai 982 F.2d at 703. In fact, other circuit courts disagree on whether its application should be limited to instances of nonliteral copying, or whether it is equally applicable to cases involving literal copying. See Lotus, 49 F.3d at 815 (noting that "[w]hile the Altai test may provide a useful framework for assessing the alleged nonliteral copying of computer code, we find it to be of little help in assessing whether the literal copying of a menu command hierarchy constitutes copyright infringement"). But see Gates Rubber Co. v. Bando Chem. Indus., Ltd.,  9 F.3d 823 (10th Cir.1993) (elaborating upon the Altai test and endorsing it as a way of determining whether "menus and sorting criteria" are protectable by copyright).

We see this disagreement between the Lotus and Gates courts as more of a matter of semantics than substance. It is undeniable that the Altai court formulated its test to address nonliteral copying of computer code. Even if the Altai test is limited to nonliteral copying, however, a parallel type of analysis must be undertaken in examining alleged instances of literal copying of computer code or screen displays. ^{Footnote 26}

Whether one chooses to call the consideration of such generally recognized challenges to literal code copying as merger and efficiency "filtration" is of little consequence; what matters is that these well-established "defenses" are considered. The Gates court simply chose to include this consideration as part of its filtration analysis, noting that "[fjiltration should eliminate from comparison the unprotectable elements of ideas, processes, facts, public domain information, merger material, scenes a faire material, and other unprotectable elements suggested by the particular facts of the program under examination." Gates, 9 F.3d at 834.

We express no opinion as to whether it is better to consider these challenges to literal copying as part of the filtration step or rather to consider them in a separate, yet parallel, analysis. The important point here is that such an analysis is necessary'it makes little difference which methodology is employed. ^{Footnote 27}

Where the district court erred, however, was in not requiring the jury to consider the connection between the Altai test and the challenges to originality, such as merger, scenes a faire, standard technique or practice, considerations of efficiency, compatibility requirements, and the like. Since the district court instructed the jury that Altai filtration was limited to nonliteral copying, the jury must have concluded that any instances of literal copying of Bateman's code by PAC were by definition acts of copyright infringement. This conclusion is a manifest distortion and misstatement of the law. There was substantial evidence presented of the literal copying of code in this case. Much of the testimony of the expert witnesses dealt with the Nimmer test and its role as the proper methodology to employ in analyzing the copyright infringement claims

of Bateman and Flicker. When this testimony was combined with the jury instruction limiting filtration analysis to nonliteral copying, it is clear that the jury was misled and did not fully understand the law of copyright relevant to this case. Under the standard established in Miller, we are "left with a substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations." 650 F.2d at 1372.

2. The Failure to Instruct the Jury on the Legal Consequences of Copying Elements Dictated by Compatibility Requirements PAC also claims that the district court erred in not instructing the jury on the legal consequences of a finding that certain instances of literal copying of code by PAC were dictated by compatibility and interoperability requirements. The district court gave no instruction on whether the application program/SBCOS interface at issue in this case was a constraint that rendered the interface either unprotectable or subject to a fair use analysis. ^{Footnote 28}

In its instructions, the court noted that "[cjomputer programs are, in essence, utilitarian articles. Articles that accomplish tasks. As such, they contain many structural and visual display elements that may be dictated by the function to be performed, by considerations of efficiency, or by external factors, such as compatibility requirements and industry demands." Supp. R1-318--95.

PAC proposed an instruction that stated, in part, that "you are to filter out as unprotectable expression portions of the program that the evidence shows were dictated by the interface with the applications code which was an external constraint." R7-216 (Instruction No. 26). Bateman objected to this instruction, noting that he had introduced evidence that PAC's application program could have been rewritten to avoid using the interface commands; ^{Footnote 29} the court, while not expressly adopting Bateman's argument, noted that the issue was far from clear and refused to charge the portion of PAC's proposed instruction quoted above. Supp. Rl-332-71.

The instruction given by the court is not technically incorrect. It fails, however, to define the important term "compatibility" ^{Footnote 30} and, more importantly, fails to instruct the jury on the legal consequences of a finding that certain copying of code by PAC was dictated by compatibility requirements. When this instruction is considered together with the "Nimmer instruction," which was limited to nonliteral similarity, it is clear that the court failed to instruct the jury that compatibility, whether it pertains to originality or fair use analysis, is a consideration that applies at the literal level.

This circuit has yet to address whether interface specifications, as a matter of law, are not entitled to copyright protection. In its briefs, PAC argues that Bateman's interface commands are legally uncopyrightable and, thus, PAC was not obligated to avoid copying them by rewriting its application program. At other times, it makes the alternative argument that the district court should have instructed the jury on the legal consequences of copying dictated by compatibility requirements. To the extent that it was making the former argument, we reject it. ^{Footnote 31}

It is an incorrect statement of the law that interface specifications are not copyrightable as a matter of law. ^{Footnote 32}

PAC is correct, however, in arguing that the district court erred in not instructing the jury on the legal consequences of copying dictated by compatibility requirements.

In analyzing elements dictated by external factors, the Second Circuit in Altai cited Professor Nimmer for the proposition that " 'in many instances it is virtually impossible to write a program to perform particular functions in a specific computing environment without employing standard techniques.' " Altai, 982 F.2d at 709 (quoting 3 Nimmer § 13.03[F][3], at 13-65). The Altai court further elaborated on this point by noting that programmers are often constrained in their design choices by "extrinsic considerations" including "the mechanical specifications of the computer on which a particular program is intended to run" and "compatibility requirements of other programs with which a program is designed to operate in conjunction." Id. at 709-10 (citing Nimmer, at 13-66-71). The Ninth Circuit has similarly noted that computer programs "contain many logical, structural, and visual display elements that are dictated by ... external factors such as compatibility requirements and industry demands," and that "[i]n some circumstances, even the exact set of commands used by a programmer is deemed functional rather than creative for purposes of copyright." Sega, 977 F.2d at 1524.

The Sega and Altai courts are certainly not alone in noting that external factors such as compatibility may work to deny copyright protection to certain portions of a computer program. See also Engineering Dynamics, 26 F.3d at 1347 n. 12.; Atari, 975 F.2d at 839. Whether the protection is unavailable because these factors render the expression unoriginal, nonexpressive per 17 U.S.C. § 102(b), or whether these factors compel a finding of fair use, copyright estoppel, or misuse, the result is to deny copyright protection to portions of the computer program. Thus, we today join these other circuits in finding that external considerations such as compatibility may negate a finding of infringement. ^{Footnote 33}

In the case at bar, we need not decide whether PAC's alleged copying constitutes an infringement. It is clear that the jury was not properly instructed on the legal principles involved in such an analysis, and thus we are "left with a substantial and ineradicable doubt as to whether the jury

was properly guided in its deliberations." Miller, 650 F.2d at 1372.

3. "Qualitatively" or "Quantitatively" Further, we address PAC's claim that the district court erred in instructing the jury that "[elyen a qualitatively small amount of copied material which remains unfiltered may be sufficiently important to the operation of Mr. Bateman's operating system program to justify a finding of substantial similarity. For instance, a small portion of a structure or code of a program may nonetheless give it its distinctive features. In such a case, a finding of substantial similarity would be appropriate." Supp. Rl-31893 (emphasis added). There is no dispute among the parties that this instruction should have contained the word "quantitatively," not "qualitatively." It is not clear, however, that the judge misspoke, because the written charges that were with the jury during their deliberations correctly read "quantitatively small amount"; the court reporter may have simply heard or transcribed the word incorrectly. Regardless of what actually happened, PAC admits that it did not object to the instruction when it was given. "[I]f no objection to the mstructions was raised at trial, we only review for plain error." Starke, 62 F.3d at 1380 (citing United States v. Andrews,  850 F.2d 1557, 1559 (11th Cir.1988), cert, denied, 488 U.S. 1032, 109 S.Ct. 842, 102 L.Ed.2d 974 (1989)). Under the plain error standard, an instruction will not be reversed "unless 'the charge, considered as a whole, is so clearly erroneous as to result in a likelihood of a grave miscarriage of justice,' or the error 'seriously affects the fairness, integrity, or public reputation of judicial proceedings.' " United States v. Pepe,