Reverse Engineering Computer Programs Under Canadian Copyright Law

19951

S. HANDA – REVERSE ENGINEERING

Reverse Engineering Computer Programs

Under Canadian Copyright Law

Sunny Handa”

This article addresses the applicability of copyright
law to the reverse engineering of computer programs. Re-
verse engineering is a method by which programmers may
uncover the ideas and processes used within an existing
computer program, thereby allowing the construction of
compatible computer programs. Recently, both the Euro-
pean Community and the United States have accepted re-
verse engineering as an exception to copyright infringement.
The author argues that Canada should follow suit. It is
claimed that such an exception is justified insofar as it is
consistent with the underlying policy goals of copyright law
within an economiclutilitarian framework. The author posits
that reverse engineering fosters the creation of standards
which increase societal wealth. The existence of a reverse
engineering exception is consistent with the goal of balanc-
ing the economic rights of individual authors with societal
technological progress, which copyright in general seeks to
encourage.

According to the author, an effective exception should
be statutorily based. The existing fair dealing exception in
the Canadian Copyright Act is juridically underdeveloped
and too uncertain to provide an effective solution to the re-
verse engineering problem. A legislative solution would
send a clear message to the software industry as well as to
the courts, and could prohibit contracting out of the Copy-
right Act, a potential loophole, should a judicial solution be
sought. The statutory exception should broadly allow the
process of reverse engineering, rather than limiting it to
cases where compatibility is sought.

Ce memoire vise A analyser l’application des droits
d’auteurs au processus d’analyse et de recomposition des
logiciels, aussi appel6 decompilation. Ce terme designe une
methode utilisee par les programmeurs pour extraire les
ides et le processus utilises dans un logiciel existant afin de
produire de nouveaux programmes compatibles avec celui-
ci. Recemment, tant les pays de l’Union europeenne que les
ttats-Unis ont exclu la decompilation des infractions ii ]a
Loi sur les droits d’auteurs. Dans ce m moire, l’auteur sou-
tient que le Canada doit lui aussi adopter une solution qui
exclut le processus de d6compilation des droits d’auteurs. II
est argue que l’implantation d’une telle exception est justi-
fie par ‘analyse des principaux objectifs vises par la loi
concemant les droits d’auteurs dans le contexte d’un enca-
drement economique/utilitaire. Le processus de ddcompila-
tion aide t crder des normes qui, selon l’argumentation
soumise, contribuent As l’enrichissement de Ia societ6.
L’exemption de ce processus constitue une demarche logi-
que dans le cadre d’un &luilibre entre les droits &conomi-
ques des auteurs individuels et le progrss technologique de
la societ6 que la Loi sur les droits d’auteurs cherche A sou-
tenir.

tre confir

Par ailleurs, l’auteur suggsre qu’une exception effi-
e par une loi. I1 est soutenu que
cace devrait
l’exemption pour l’utilisation equitable contenue dans la Loi
sur les droits d’auteurs canadienne est juridiquement insuf-
fisante et trop vague pour offrir une solution efficace au
problime du processes de decompilation. Une solution legis-
lative transmettrait un message clair h, l’industrie des logi-
ciels ainsi qu’aux
l’option
d’exclusion volontaire (vcontracting out>) de la Loi sur les
droits d’auteurs. Cette exclusion serait possiblement accor-
dde si une solution judiciaire 6tait privilegi~e. En plus, il est
suggere que l’exemption statutaire devrait englober tout le
processes de decompilation plutht que de le limiter aux cas
oh) une compatibilite et recherchee.

tribunaux et prohiberait

“B. Comm. (McGill), LL.B. (University of Toronto), LL.M., Doctoral Candidate (McGill); Barris-
ter and Solicitor (of the Ontario Bar). Sessional Lecturer, Faculty of Law, McGill University. The
author wishes to express his gratitude to Rosemary Shuttleworth, without whose ongoing support this
article would have never been written.

McGill Law Journal 1995
Revue de droit de McGill
To be cited as: (1995) 40 McGill L.J. 621
Mode de rrf~rence: (1995) 40 R.D. McGill 621

MCGILL LAW JOURNAL/REVUE DE DROITDEMCGILL

[Vol. 40

Synopsis

Introduction

I. Concepts

A. Computer Programs
B. Computer Languages
C. Computer Memory
D. Operating Systems
E. Reverse Engineering

1.
2.
3.
4.

Intermediate Copying
Reverse Engineering and Piracy
Using the Results Obtained Through Reverse Engineering
Using Reverse Engineering in the Creation of
Compatible Programs
Conclusion

II. Intellectual Property Protections for Computer Programs

A. The Law of Copyright

The Copyright Act
a.
b.
c.
d.
e.
f

The Idea/Expression Dichotomy
,Formalities, Term and Ownership
Originality
Fixation
Infringement of Copyright
Exceptions to Infringement
i.
ii.
iii.
iv.

Translation – Modification Exception
Making Backup Copies
Fair Dealing Under the Copyright Act
Public Interest Exception

B. Is Reverse Engineering an Infringement of Copyright Law?
C. Reverse Engineering Under Other Legal Regimes

Trade Secrets
Enhancing Copyright Protection Through Licensing/Contract Law
a.
b.
c.
Competition Law
Patents
Semiconductor Chip Protection

1.
2.

3.
4.
5.

D. Conclusion

1995]

S. HANDA – REVERSE ENGINEERING

III. Existing Approaches to the Reverse Engineering Problem

A. United States

B. The European Union

1.
2.
3.

1.
2.

E.E Johnson Co. v. Uniden Corp. of America
Atari Games Corp. v. Nintendo of America Inc.
Sega Enterprises Ltd. v. Accolade Inc.

The E.E.C. Directive on Computer Programs
Implementation of the Directive’s Provisions in the
United Kingdom

C. Australia

Autodesk Inc. v. Dyason

D. Conclusion

IV. Justifying the Reverse Engineering of Computer Programs

A. The Economics of Intellectual Property Rights
B. The Economics of Reverse Engineering

1.
2.

Software Compatibility
The Costs and Benefits of a Standardized Computing
Environment

C. Developing a Solution to the Problem
Creating a Statutory Exception
Relying on the Fair Dealing Exception
Alternatives to Copyright

1.
2.
3.

D. The Scope of a Reverse Engineering Right

Conclusion

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

Introduction

Originality is nothing but judicious imitation. Thre
most original writers borrowed one firn another
The instruction we find in books is like fire. We fetch
it from our neighbors, kindle it at home, communi-
cate it to others, and it becomes the property of all.
Voltaire

Nothing can with greater propriety be called a
man’s property than the fruit of his brains. The
property in any article or substance accruing to hin
by reason of his own mechanical labour is never
denied him: the labour of his miind is no less ardu-
ous and consequently no less worthy of the protec-
tion of the law.

Copinger and Skone James on Copyright

Laws that do not embody public opinion can never
be enforced.

Elbert Hubbard

If we desire respect for the law, we nmust first imake
the law respectable.

Louis D. Brandeis

The field of computer law has grown in leaps and bounds in recent years. In
fact, the very use of the title “computer law” to denote a separate body of law con-
tinues to be regarded with some skepticism by traditionalists. A number of law
firms, however, have developed departments or groups that specialize in this fast-
growing field, and some practitioners hold themselves out as computer lawyers.
What distinguishes computer law from accepted traditional legal categories such as
copyright or, even more, broadly, intellectual property law, is its interdisciplinary
nature. It incorporates many established areas of law including, but not limited to,
copyright, patents, trade secrets, semiconductor chip, contract, criminal, and tort
law.

1995]

S. HANDA – REVERSE ENGINEERING

In terms of civil and criminal protection against illicit copying of computer
programs, copyright law has clearly evolved as the standard form of protection
throughout the world. It is generally accepted that international copyright instru-
ments such as the Berne Convention’ and the Universal Copyright Convention,2
while not explicitly referring to computer programs as protectable works, do protect
computer programs as literary works. The field of copyright law, not often consid-
ered a particularly dynamic area of law, has in the past decade been the subject of
increased activity as a result of its application to computer programs. While com-
puter programs generally fit into the underlying framework of copyright law, they
do not, by their very nature, lend themselves perfectly to many of the idiosyncratic
jurisprudential concepts of copyright law as traditionally understood. Accordingly,
copyright law as it applies to computers is rapidly developing its own wealth of ju-
risprudence.

One of the most topical and difficult decisions facing legislators and jurists in
this area is the permissibility of reverse engineering of computer programs under
copyright law. Reverse engineering, or decompilation as it is sometimes called, in-
volves working backwards from a finished product in order to gain a better under-
standing of how the product was produced. The issue of whether software products
protected by copyright may be legally reverse engineered without the copyright
holder’s consent remains unchallenged under Canadian law. Indeed, only a few ju-
risdictions have yet to deal with the issue. Most notably, the European Community,
in its Software Directive of May 14, 1991,3 took a bold step forward in declaring
that computer programs may be decompiled for the purposes of achieving interop-
erability (compatibility) with other computer programs. More recently, several
American courts have had to deal with whether, and under what circumstances, re-
verse engineering of computer programs would be permitted under American
copyright laws.4 To date, the issue of reverse engineering of copyrighted works has
not been dealt with in Canada, either legislatively or judicially. A broad reading of
the Canadian Copyright Act suggests that prima facie the reverse engineering of a
copyrighted work without the copyright holder’s consent is prohibited.

The American copyright legislation, however, while appearing very similar to

Revised Berne Convention, being Schedule I1 of the Copyright Act, R.S.C. 1985, c. C-42
[hereinafter Bente Convention]. The North American Free Trade Agreement [hereinafter NAFTA], to
which Canada is a party and which came into force on January 1, 1994, requires that the contracting
member states accede to the (latest) 1971 Paris revision to the Berne Convention (An Act to Imple-
ment the North American Free Trade Agreement, S.C. 1993, c. C-44, s. 5).

2 6 September 1952, 216 U.N.T.S. 132.
‘EC, Council Directive (EEC) No 91/250 of 14 May 1991 on the legal protection of computer pro-

grams, O.J. Legislation (1991) No L122 at 42 [hereinaftet Directive].

‘ E.g. Sega Enterprises Ltt v. Accolade hic., 977 F2d 1510 (9th Cir. 1992) [hereinafter Sega]; Atari
Gaines Corp. v. Nintendo of America Inc., 975 F.2d 832 (Fed. Cir. 1992) [hereinafter Atari]. For cases
involving the reverse engineering of data tables, see e.g. E.E Johnson Co. v. Uniden Corp. of Amer-
ica, 623 FSupp. 1485 (D.C. Minn. 1985) [hereinafter Uniden]; Autodesk Inc. v. Dyason (1992), 22
I.PR. 162 (Aust. H.C.), rev’ing (1990), 18 I.P.R. 399 (Aust. Fed. CL) [hereinafterAutodesk].

MCGiLL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

the Canadian legislation, has been interpreted by American courts as allowing re-
verse engineering in certain circumstances. Although Canadian courts are not
bound by decisions of their American counterparts, much is borrowed from Ameri-
can law in the computer law field.5 There are also strong public policy arguments in
favour of permitting the reverse engineering of computer programs. Reverse engi-
neering allows for the creation of interoperable or compatible computer programs
in cases where the copyright holder will not release the program’s technical specifi-
cations. The balance between public benefit and the protection of an individual’s
right to claim rewards associated with a work’s distribution and use is at the core of
intellectual property rights protection. Accordingly, in considering whether to allow
reverse engineering, the societal benefits to be gained must be weighed against the
potential risks to the copyright holder.

This article will suggest that a broad right of reverse engineering with respect
to computer programs should be permitted under Canadian law. This claim will be
based on a cost/benefit or economic analysis of the law and of the outcomes asso-
ciated with various proposed reverse engineering scenarios. Justifications for the
use of law and economics as a theoretical basis for allowing reverse engineering,
and as an explanation of intellectual property protections more generally, will also
be discussed.

I. Concepts

Before proceeding any further, it is necessary to define concepts such as reverse
engineering, computer programs and other related terminology for purposes of both
clarity and consistency. This is not meant to be a comprehensive review of com-
puter technology, but rather a cursory overview of the technological concepts in-
volved in the reverse engineering debate.

5 In Compo Co. v. Blue Crest Music Inc. (1979), [1980] 1 S.C.R. 357, 105 D.L.R. (3d) 249

[hereinafter Compo cited to S.C.R.], Estey J. stated:

The United States Copyright Act, both in its present and earlier forms, has, of course,
many similarities to the Canadian Act, as well as to the pre-existing Imperial Copyright
Act. However, United States court decisions, even where the factual situations are
similar, must be scrutinized very carefully because of fundamental differences in
copyright concepts which have been adopted in the legislation of that country…. That
is not to say that we may not find some assistance in examining the experience in the
United States (Compo, ibid at 366-67).

O’Leary J. further recognized the value of American jurisprudence in Delrina Corp. v. Triolet Sys-
tems Inc. (1993), 47 C.P.R. (3d) I at 28,9 B.L.R. (2d) 140 (Ont. Ct. (Gen. Div.)) [hereinafter Delrina
cited to C.P.R.], stating: “The U.S. Copyright Act differs somewhat from the Canadian Copyright Act,
but nevertheless American copyright decisions were heavily relied on by both the plaintiff and defen-
dants in this case and are of great assistance on the issues before me.”

1995]

S. HANDA – REVERSE ENGINEERING

A. Computer Programs

Section 2 of the Canadian Copyright Act defines a “computer program” as “a
set of instructions or statements, expressed, fixed, embodied or stored in any man-
ner, that is to be used directly or indirectly in a computer in order to bring about a
specific result.” This definition is virtually identical to that set out in section 101 of
the American Copyright Act, which defines “computer program”f as “a set of state-
ments or instructions to be used directly or indirectly in a computer in order to
bring about a certain result.”6 Subsection 342.1(2) of the Canadian Criminal Code7
defines a computer program as “data representing instructions of statements that,
when executed in a computer system causes the computer system to perform a
function.”

Together, these definitions identify a computer program as an arrangement of
instructions that is used in a computer to solve a particular problem. Simple data
that does not in itself instruct a computer to perform calculations towards a given
end does not qualify as a computer program. Other similar definitions can be found
scattered throughout the jurisprudence dealing with the copyrightability of com-
puter programs.

Computer programs can be broadly categorized into two types: operating sys-
tem programs and application programs. Reed J. articulated a particularly clear
definition of each in Apple Computer Inc. v. Mackintosh Computers Ltd.:

Application programs are designed for a specific task, such as the playing of a
video game, preparation of a tax return, or the writing of a text. Operating sys-
tem programs are designed primarily to facilitate the operation of application
programs and perform tasks common to any application program. Without
them each application program would need to duplicate their functions.’

B. Computer Languages

Also essential to understanding the issues surrounding reverse engineering is a
minimal knowledge of how a computer program is built, compiled and executed, as.
well as an understanding of the related jargon. Computer programs are written in
computer languages which vary in their degree of resemblance to “ordinary
mathematics and English (or other common languages).” 9 A higher level computer
language is said to be closer to “common languages” in its vocabulary than is a
lower level language. The level, also referred to as the generation, of the computer

” 17 U.S.C. 101 (1994) [hereinafter CopyrightAct (U.S.)].
7 R.S.C. 1985, c. C-46.
‘ (1986), [1987] 1 F.C. 173, 10 C.PR. (3d) 1 at 11 (T.D.) [hereinafter Apple cited to C..R.], var’d
(1987), [1988] 1 EC. 673, 44 D.L.R. (4th) 74 (C.A.), aff’d [1990] 2 S.C.R. 209, 30 C.ER. (3d) 257
[hereinafter Apple (S.C.C.) cited to C.PR.].

9 Ibid. at 7.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40
language depends “upon the ease with which it can be read” by human beings.’ In
order for a computer to process the instructions of any given language, the instruc-
tions must first be compiled, or translated, into a language or notation that the com-
puter’s processor can understand. This latter notation is known as the lowest, or
first level language.

The term “source code” refers to the written form of program that the user
physically produces in a given computer language. Programmers today have a
wealth of higher languages from which to choose, including PASCAL, COBOL,
FORTRAN, BASIC, and ASSEMBLER.” Once the program’s source code has
been written, the programmer will “compile” the source code into machine-
readable object code using another computer program, known as a compiler, to per-
form the conversion. Object code is generally in binary form, a language made up
exclusively of “l”s and “O”s, and is used directly by the computer.

C. Computer Memory

Another concept central to the use of computers is that of memory. The Cana-
dian Copyright Act’s definition of “computer program” requires that a set of in-
structions be “expressed, fixed, embodied or stored” in order to qualify as a com-
puter program. In the United States, legislation defines this fixation as existing
where a tangible mode of expression is embodied in a form which is sufficiently
permanent and stable to be “perceived, reproduced, or otherwise communicated,
either directly or with the aid of a machine or device.”‘2 Traditionally, the fixation
of copyrighted literary works was done on paper. 3 This type of storage, while also
valid for computer programs, is not the only possible type of fixation. Computer
programs stored in a computer’s memory also qualify as being stored for the pur-
poses of the Copyright Act. 4

For the purposes of convenience, computer memory devices can be separated
into three categories: internal, external and archival.”5 Internal memory consists of

10 Ibid.
” ASSEMBLER is a second generation, or intermediate, language whereas PASCAL, COBOL,

BASIC and FORTRAN are considered to be third generation languages.
2 Copyright Act (U.S.), supra note 6, 102.
13 See Part H.1.d, below, for a discussion of fixation.
” In Apple, the Court held that object code stored on a silicon microchip was a reproduction in a
material form of copyrightable source code and was therefore protectable as a “computer program”
under the Copyright Act. According to the Court, it was irrelevant that the object code was not neces-
sarily in human-readable form.

” R.R. Panko, End User Computing (New YorI: John Wiley & Sons, 1988) at 315. A simple, but

useful, definition of “memory” can be found in Delrina:

[Memory is defined as a]n area of the computer’s circuitry that holds applications and
any data generated with those applications. Information held in Random Access Mem-
ory (RAM) is erased whenever the computer is turned off. Information held in Read
Only Memory (ROM) is retained even when the computer is off.

1995]

S. HANDA – REVERSE ENGINEERING

those memory devices, built into the structure of the computer, such as ROM-chips
(read-only memory microchips) and RAM-chips (random access memory micro-
chips), that are necessary for the computer to operate at its most basic level. Exter-
nal memory includes devices, such as disks and CD-Roms, that provide additional,
non-essential, but more permanent storage at a relatively cheap cost. Archival
memory, consisting of devices such as tape-backup machines, does not impact on
the reverse engineering debate in any meaningful way and will thus not be dis-
cussed in this article.

D. Operating Systems

In order for a computer to execute its programs and interact with the user, an
operating system is required. The purpose of the operating system is to set up and
manage the computer system’s environment and resources. These systems are com-
puter programs the specifications of which are essential to computer application
programmers who. are writing programs designed for certain computer systems.

E. Reverse Engineering

Reverse engineering, as the name suggests, is the opposite to the above de-
scribed process of constructing a computer program. Reverse engineering “involves
going backwards from a finished product and determining how the product
works.”‘ 6 Another definition holds that reverse engineering occurs where “one in-
spects or takes apart a new product … by translating the unreadable object code of a
program into source code that may be studied.”” 7 The terms “disassembly” and
“decompilation”, often used as synonyms of “reverse engineering”, are actually
subsets
thereof. Decompilation of a computer program occurs where one
“convert[s] the machine code version [of the program] into a high level language,”‘ S
whereas “[d]isassembly of a computer program is done by translating the machine
or object code into humanly-readable assembly language.”‘ 9 The only difference
between decompilation and disassembly is the product obtained at the end of the
process. In the former case it involves converting the machine code into a high
level language, whereas in the case of disassembly the final product is in
ASSEMBLER, an intermediate level language.

Memory usually refers to the high speed semiconductor storage within a computer
that is used to temporarily store data while it is being processed or examined. The term
“memory” is also generically extended to refer to data that is stored externally on disks
and tapes (Delrina, supra note 5 at 51-52).

“Sega Enterprises Ltd. v. Accolade Inc., 785 F. Supp. 1392 at 1394 (N.D. Cal. 1992).
‘7 G.R. Ignatin, “Let the Hackers Hack: Allowing the Reverse Engineering of Copyrighted Com-

puter Programs to Achieve Compatibility” (1992) 140 U. Penn. L. Rev. 1999 at 2010.

D.I. Bainbridge, “Computer Programs and Copyright: More Exceptions to Infringement” (1993)

56 Modem L. Rev. 591 at 593.

‘9 Uniden, supra note 4 at 1490.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

Practically speaking, most reverse engineering is of the disassembly variety, as
the computer programs that are used in performing disassembly are easier to create
and more flexible than are decompilers. Furthermore, software engineers and com-
puter programmers involved in reverse engineering are generally comfortable in an
ASSEMBLER language environment and do not need to visualize the program in a
higher level language.

Intermediate Copying

The copying of computer programs as it relates to the reverse engineering
process can occur in several ways. The first instance of copying that results from
reverse engineering occurs during the deconstruction process. Whether the reverse
engineering is conducted through a manual inspection of the program code which is
reassembled on paper, or through the more common method of disassembly, the
process invariably results in copying. 0 This copying does not result in a completely
verbatim copy of the original code, but rather entails translating the original pro-
gram code several times, each time moving closer to an ASSEMBLER, or higher
level, language translation of the original object code.2 Whether intermediate cop-
ies produced during the disassembly process violate copyright rules is not alto-
gether clear.

Reverse Engineering and Piracy

Piracy, the most widely publicized form of illicit copying, concerns the direct
reproduction of a computer program, usually by a user, without the author’s con-
sent. Because computer programs are stored digitally (as ‘T’s and “O”s), flawless
reproductions can be made at little cost to the copier.23 As a result, computer pro-
grams, and more recently digitally-stored audio recordings, have increased intellec-

Copying, for the purposes of copyright, is subject to the copies being “fixed” in some form. See
2′ A disassembler makes several “passes” over the original code, gradually building towards a final

Part HA. 1.d, below, for a discussion of fixation.’

translation in ASSEMBLER language.

2 See text accompanying note 76, below, for a discussion of intermediate copying that remains
stored only in RAM. Where the intermediate copies qre stored in a more permanent manner, it is
more likely that they will violate copyright laws (see Part H., below). See also paragraph 3(l)(a) of
the Copyright Act, which prohibits unauthorized translations of protected works; see also infra note
201 and accompanying text for the decision of the United States Court of Appeals (Ninth Circuit)
with respect to intermediate copying.

2 In its final report to Congress concerning the copyrightability of computer programs, the United
States National Commission on New Technological Uses of Copyrighted Works (C.O.N.T.U.) stated:

The cost of developing computer programs is far greater than the cost of their duplica-
tion. Consequently, computer programs … are likely to be disseminated only if … the
creator may spread its costs over multiple copies of the work with some form of pro-
tection against unauthorized duplication of the work (National Commission on New
Technological Uses of Copyrighted Works, Final Report (Washington: Library of
Congress, 1979) at 11).

1995]

S. HANDA – REVERSE ENGINEERING

tual property protection needs as the economic incentives to purchase the original
product have diminished.24 Typically, with this form of direct copying, a user will
make an unaltered copy of an original computer progran or of an existing copy
thereof, and will use the copy instead of purchasing the original computer pro-
gram.” Generally, these copies can be easily produced using simple operating sys-
tem commands, although, in an effort to deter this practice, some computer pro-
gram manufacturers have attempted to use copy protection schemes.

Copy protection schemes vary in their functioning and, because of their very
purpose, cannot be standardized. These schemes, however, can currently be divided
into three basic types:2 7 (1) the program is stored in such a way that copying pro-
grams cannot copy all the necessary parts; (2) the program prompts the user for a
code or other piece of information that can only be found in the original packaging;
and (3) the program comes with a hardware device that attaches to the computer
and sends the program signals or information which it seeks prior to functioning.
The first type of protection is often defeated by copying programs developed by
third parties, which can copy the required parts. With respect to the latter two
schemes, program modifications, known as “cracks”, designed to defeat the protec-
tion will invariably appear soon after, or sometimes even before, a program’s public
release.”

Program cracking for the purpose of defeating a copy protection scheme often
involves some disassembly of the protected computer program. Cracking programs,
or enhanced disassemblers with specialized features to assist a cracker, are also
generally available as either shareware or freeware. 29 Recently, the reverse engi-

, W.M. Landes & R.A. Posner, “An Economic Analysis of Copyright Law” (1989) 18 J. Legal

Stud. 325 at 327.

” In this context, a copy is produced in the absence of authority of the copyright holder, while an

original is a copy produced with permission.

2″ ‘Persons who have not paid for a software copy cannot be excluded from using a program, and
use of a program copy by one person does not necessarily diminish the supply of copies available for
use by others” (D.A. Rice, “Public Goods, Private Contract and Public Policy: Federal Pre-emption of
Software License Prohibitions Against Reverse Engineering (1987) 53 U. Pitt. L. Rev. 543 at 545).

27 A comprehensive review of more eclectic forms of copy protection is beyond the scope of this ar-

ticle.

‘ A “crack” is often distributed in either printed form (as a set of instructions on how to modify the
program to defeat the protection) or as a “patch”. A patch is a small computer program, or portion of
computer code, that applies itself to the protected program and replaces the required code with in-
structions that defeat the protection. An example of a simple crack is a set of instructions that tells the
program to skip over the code that executes the protection checks. Cracks are commonly distributed
on various computer bulletin boards and are easily available on the Internet.

9 Shareware refers to computer programs which may be used for a trial period without infringing
copyright, after which a licence fee is payable to the copyright holder for continued use. Shareware
programs are also freely distributable in their unaltered state to other users (hence the “share” in
shareware) who may try them out for the trial period without payment. Freeware refers to computer
programs where the copyright holder waives his or her rights to any economic return for its use.
Waiving of economic returns does not mean a waiver of moral rights which would allow users to
modify and to alter the original work. Any such waiver is independent of the free or shareware desig-

MCGILL LAw JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

neering debate was brought in front of the American courts with respect to video
game cartridges which contained computer programs protected with a program
check (akin to protection type (3) set out above).1

Using the Results Obtained Through Reverse Engineering

Another form of illicit copying occurs where the copier, usually a programmer,
alters or uses parts of the original program in his or her own work. The amount of
modification involved varies greatly in these situations, and may or may not be
substantial enough to violate copyright. 3 Copying a work in this manner may not
require actual direct copying of the original program code (known as “literal copy-
ing”).32 “Non-literal” elements of a computer program are “those aspects that are
not reduced to a written code, 33 and include “components such as general flow
charts as well as the more specific organization of inter-modular relationships, pa-
rameter lists, and macros. ‘ Screen displays also fall within the definition of non-
literal elements. 3
1 Copying these non-literal elements may also infringe copyright,
and may be accomplished without dissecting the program as previously discussed.
For example, copying the layout of a screen may simply involve visual examination
of the original program and subsequent replication using entirely new program-
ming. Similarly, copying the order of the keystrokes used in the operation of a
computer program, known as “command sequences”, may not involve actual literal
copying of the original code. The issue of non-literal copying of computer pro-
grams has been highly topical in recent years, and is far from being resolved.”

Generally, reverse engineering will entail a deconstruction of the original pro-
gram’s literal code to uncover its underlying ideas, rather than non-literal copying.
Whether the reverse engineering will be sufficient to uncover these ideas, or
whether it will simply stop at the point of uncovering expression will depend on
each individual case. This exercise is then followed by an attempt to construct a dif-
ferent program using the results of the reverse engineering. The degree to which the
reprogramming will involve copying of the original program code will also vary
greatly. This genre of copying may include some forms of non-literal copying, such

nation.

Sega, supra note 4; Atari, supra note 4.
Copyright only protects expressions and not their underlying ideas. See Part HA.l, below, for a

discussion of copyright principles.

32 Rice, supra note 26 at 567.
3 ContputerAssociates International Inc. v. Altai Inc., 23 U.S.PQ. (2d) 1241 at 1244 (2d Cir. 1992)

[hereinafter Altai].
Ibid. at 1249.

3 Screen displays are protected as parts of a computer program “except in the case of programs
whose very purpose is to produce screen displays for use in playing of games or for some artistic or
other like purpose” (Delrina, supra note 5 at 32).

36 See Delrina, ibid.; Systimes informatisis Solartronix v. Cggep de Jonqui~re (1988), 22 C.I.P.R.
101 (Que. Sup. Ct.); Lotus Development Corporation v. Paperback Softivare International, 740 F.
Supp. 37 (D. Mass. 1990) [hereinafter Paperback Software]; Altai, supra note 33; John Richardson
Computers Ltd v. Flanders, [1992] F.S.R. 391 (Ch.) [hereinafter Richardson].

1995]

S. HANDA – REVERSE ENGINEERING

as reproducing the layout of the program’s subroutines (structure), as well as literal
copying, such as copying parts of the original program’s code either directly or
through a translation into another language.

The results obtained by reverse engineering can be used for a number of pur-
poses including: the programming of cracks to defeat a program’s copy protection;
the creation of a similar program or of a program that uses the same routines as the
original program in order to save time and expense; academic study of the pro-
gram’s underlying ideas and the techniques used in their expression; or the creation
of compatible programs. While the first two applications do not usually gamer
much support as they involve an element of thievery, the last two goals are more
readily perceived as morally acceptable justifications of reverse engineering com-
patible with the underlying spirit of copyright law.”

Using Reverse Engineering in the Creation of Compatible
Programs

Although no statutory definition of “interoperable” or “compatible” program
currently exists, the concepts are synonymous and simple to define. Compatibility
is a measure of the degree to which one program will function in conjunction with
another. In order to create a compatible program, “a programmer mut have a
complete specification of the other program’s ‘interface’ –
a precise description of
how the program receives, stores and/or outputs information.”38 Traditionally, com-
patible programs were written by the same company because their programmers
had access to the necessary specifications.

A more modem approach to the design of compatible programs is to create
more robust operating systems which can handle a greater number of functions, and
to provide application programmers with consistent specifications for all types of
data structures (objects) controlled by each operating system. Programmers writing
computer programs for use with such operating systems are usually given access to
the operating system’s specifications by the operating system designers at minimal
cost. Objects created by computer programs that follow the standardized specifica-
tion are then useable in other programs that also follow the specification. The pro-
grams are therefore made compatible without the programmers ever having seen or
used each other’s computer programs because the operating system acts as the
standardizing link. Examples of such object-oriehted operating systems are Micro-
soft’s WindowsT and Next’s NextStepT1 operating systems. 39 The move towards

3, Ignatin, supra note 17 at 2022. Allowing reverse engineering for the creation of interoperable
programs is recognized in Article 6 of the European Community’s Software Directive, supra note 3 at
45, as the sole justification for reverse engineering.

S Ignatin, ibid. at 2023.
‘9 Under the popular Microsoft WindowsTM operating system, a system of standardized objects, ob-
ject linking and embedding (O.L.E.), is used. O.L.E. allows users to share information created in
other applications with the application they are using. For example, users of Microsoft’s WordTM may
either link or embed objects created with Microsoft’s ExcelTM spreadsheet program inside their

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

creating standards through operating systems is a sound one; however, the creation
of new types of objects not contemplated by the operating system designers may
arise. In such cases, the standards are once again protected as the property of the
application’s designers. 0

Conclusion

As the law currently stands, the newly constructed program or parts
thereof, will sometimes be considered a copy and thereby infringe the original pro-
gram’s copyright, whereas in other cases, the new program will not involve a suffi-
cient degree of copying to amount to a copyright violation. This determination is
made irrespective of whether the new program is compatible or whether it is con-
structed to compete with the original as a similar product.4’ A more immediate
question, however, is whether the actual reverse engineering of the computer pro-
gram is an infringement in itself. Both of these issues will be addressed in the fol-
lowing discussion.

II. Intellectual Property Protections for Computer Programs

Computer programs are protected from illicit copying under a number of legal
regimes. Copyright, patent, trade secret, and semiconductor chip laws all provide
intrinsic measures of protection against the unauthorized copying of a computer
program. The term “intrinsic” is used to distinguish the protections granted by these
regimes from those contractual provisions fashioned by private parties. The protec-
tions provided by each of the regimes listed above are enforceable at law notwith-

WordTM document. A large part of the linkage and embedding of these objects is a function of the
WindowsT operating system, and not of any specific design created by either application’s pro-
grammers.

‘ There is also the question whether operating system designers, often members of the same com-
pany that designs various applications for use with their operating systems, will release all of the
specifications required to make the most effective use of the operating system environment. Clearly,
there is motive to withhold some of the technical information so as to provide one’s own company
with a competitive advantage in the application program market. In 1993, the United States Depart-
ment of Justice stepped up a Federal Trade Comnmission anti-trust investigation of Microsoft Corpo-
ration. This investigation was recently concluded (1995) in favour of Microsoft. Among the charges
investigated were claims by competitors in the application program market “that Microsoft unfairly
uses secret features known as ‘undocumented calls’ and its advance knowledge of changes to MS-
DOS and the related Windows software to place its competitors at a disadvantage” O.K. Gotts,
“Regulators Focusing on Antitrust Issues” The National Law Journal (24 January 1994) S12.

” Prima facie copyright law is currently oblivious to any such distinction. A determination of
copying under the Copyright Act concerns whether the copied code is “substantial” in quality and not
in quantity. However, if a newly constructed program is directly competing with an original work
from which information was reverse engineered, it is more likely that the parts used will be consid-
ered substantial. This occurs because the quality of the parts used in a competing program may appear
of greater import than if they had been used in a compatible program which, to a court, seems differ-
ent in both appearance and structure (SAS Institute, Inc. v. S & H Computer Systems, Inc., 605 F.
Supp. 816 (N.D. Tenn. 1985) [hereinafter SAS]).

1995]

S. HANDA – REVERSE ENGINEERING

standing the absence of specific contractual agreements between parties.

A. The Law of Copyright42

The CopyrightAct

Following the House of Lords’ decision in Donaldson v. Becket,43 which effec-
tively abolished the common law of copyright, the British Parliament passed a se-
ries of Copyright Acts. 44 This body of legislation was eventually the source of the
first Canadian Copyright Act which came into force on January 1, 1924.46 The Ca-
nadian Copyright Act has continued in the tradition of its Imperial forebearers and
explicitly states that no copyright or similar right shall exist in Canada other than
those recognized under the Copyright Act.47

Today’s copyright laws are no longer limited to the protection of published and
unpublished manuscripts. Copyright currently protects dramatic, musical and artis-
tic works, as well as a broad category of literary works including tables, compila-
tions and translations. Section 2 of the Copyright Act classifies computer programs
as literary works for the purposes of copyright protection. This protection permits
the copyright holder to make copies of the work, while prohibiting others from do-
ing So* Copyright holders may, however, freely license or assign their economic
rights.

42 Copyright law, also referred to as Anglo-American copyright law or the common law of copy-
right, is to be differentiated from droit d’auteur or continental regimes, which protect similar works
but are primarily used in civilian jurisdictions. Copyright, as discussed in this paper, refers to those
regimes with common law origins.

” (1774), 4 Burr. 2408, 2 Bro. Parl. Cas. 129.
4 Copyright Act, 1814 (U.K.), 54 Geo. 3, c. 156; Copyright Act, 1842 (U.K.), 5 & 6 Vict., c. 45;

Copyright Act, 1911 (U.K.), 1 & 2 Geo. 5, c. 46.
45 S.C. 1921, c. 24.
” The 1927 Canadian Copyright Act abrogated all previous Copyright Acts enacted by the Imperial
Parliament.
47 Copyright Act, supra note 1, s. 63. Prior to passage of the 1921 Act, a common law of copyright
had existed in Canada. This right was replaced by a statutory right under section 42 of the 1921 Act.
In Conipo, Estey J. stated:

[Clopyright is neither tort law nor property law in classification, but is statutory law. It
neither cuts across existing rights in property of conduct nor falls between rights and
obligations heretofore existing in the common law. Copyright legislation simply creates
rights and obligations upon their terms and in the circumstances set out in the statute
(Compo, supra note 5 at 372-73).

See B.B. Sookman, Computer Law: Acquiring and Protecting Information Technology (Toronto:
Carswell, 1989) at 3-1; H.G. Fox, The Canadian Law of Copyright and Industrial Designs, 2d ed.
(Toronto: Carswell, 1967) at 1-2.

‘9 Copyrighted works consist of two components: economic rights and moral rights. The former re-
fers to the right of the copyright holder to reap economic benefits for authorizing use of the work,
whereas the latter refers to the author’s (as opposed to the copyright holder’s) right to the integrity of
the work as well as the right to be associated with the work “in certain circumstances” (Copyright Act,
supra note 1, s. 14.1(1)). “Moral rights may not be assigned butmay be waived in whole or in part”

MCGiLL LAW JOURNAL/REVUE DE DROIT DE MCGILL

[Vol. 40

The Idea/Expression Dichotomy

Copyright protects the expression of ideas, but does not extend to the ideas
themselves. 0 This separation is referred to as the idea/expression dichotomy. Find-
ing the line that separates idea from expression is not an easy task, and the diffi-
culty is even more pronounced when dealing with computer programs which are by
their very nature utilitarian works and hence intertwindd with the ideas they seek to
express. 5′ The difficulty inherent in creating a test to distill expression from idea
lies in defining the scope of expression. By adopting an overly liberal view of ex-
pression, one risks granting monopoly protection to the first authors of certain types
of programs:

[Such a view] would thereby inhibit other creators from developing improved
products. [Conversely, d]rawing the line too conservatively would allow pro-
grammers’ efforts to be copied easily, thus discouraging the creation of all but
modest incremental advances 2

In order to devise a sound test, American jurisprudence suggests that

[t]he court must be faithful to the statutory language and mindful of both the
ultimate goal of copyright law –
and
Congress’ chosen method of achieving this goal –
private reward to the indi-
vidual author.

the advancement of public welfare –

The consequences of an imperfect test are illustrated by the decision in Whelan
Associates, Inc. v. Jaslow Dental Laboratory, Inc.5 and its short-lived, but highly

(Copyright Act, ibicL, s. 14.1(2)).
‘0 In Moreau v. St. Vncent, [1950] Ex. C.R. 198 [hereinafter Moreau], Thorson P. stated:

[An elementary principle of copyright law [is] that an author has no copyright in ideas
but only in his expression of them. The law of copyright does not give him any mo-
nopoly in the use of the ideas with which he deals or any property in them, even if they
are original (Moreau, ibid. at 203).

5, See Delrina, supra note 5 at 33. The doctrine of merger holds that if an expression is necessary to
the function or efficiency of a given idea, it is considered necessarily incidental to the idea and is not
protectable as an expression. Such a form of expression is said to be purely functional, resulting in the
merger of idea and expression. For example, where a program requires the user to type the word
“print” followed by the command “full page” in order to direct the output of a program to a printer,
the command sequence will not be protected as it is necessarily incidental to the idea of printing.

2 Paperback Software, supra note 36 at 53, quoting P.S. Menell, “An Analysis of the Scope of

55Paperback Softvare, ibid.

797 F2d 1222 (3d Cir. 1986) [hereinafter Whelan]. A medical software developer, after designing
a dental laboratory program for the plaintiff, started up her own company and developed a similar
program. The Third Circuit Court of Appeals found that the defendant’s program copied the structure,
sequence and organization of the plaintiff’s program, and that this was enough to constitute an in-
fringement of the plaintiff’s copyright. This decision was revolutionary in that it formally extended
software copyright protection beyond the literal copying of source code to non-literal elements, in this
case the structure, sequence and organization of the program. The Court held that since the plot of a

1995]

S. HANDA – REVERSE ENGINEERING

controversial legacy in the United States. In terms of the idea/expression dichot-
omy, the Whelan Court decided that “[t]he purpose or function of a utilitarian work
would be the work’s idea, and everything that is not necessary to that purpose or
function would be part of the expression of the idea.”55 The case was subsequently
criticized as having a “somewhat outdated appreciation of computer science” and
for ignoring “practical considerations”.56 In Altai, decided after Whelan, the Second
Circuit of the United States Court of Appeals devised a three-part, abstraction-
filtration-comparison test that would better distinguish idea from expression in
computer programs.57 According to Walker J. of the Altai Court,

[i]n ascertaining substantial similarity under this approach, a court would first
break down the allegedly infringed program into its constituent structural parts
[(abstraction)]. Then, by examining each of these parts for such things as in-
corporated 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 [(filtration)]. Left with a kernel, or possibly
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 al-
legedly infringing program [(comparison)]!’

The three-part abstraction-filtration-comparison test has gained widespread accep-
tance in the United States,59 and has also been adopted in computer program copy-
right cases in Canada and the United Kingdom. 60

Fonnalities, Term and Ownership

In order to avail itself of copyright protection, a work. may be either published
1 Under Canadian copyright law there is no registration62 or mark-

or unpublished. 6

story or play is protected by copyright, so should be the sequence and organization of programs. The
Court reasoned that there were many possible ways in which to organize the idea of a dental lab pro-
gram, and therefore the particular manner chosen by the plaintiff was a copyrightable expression of
that idea.

5 Ibidl at 1236.
56Altai, supra note 33 at 1252.

In CMAX Cleveland, Inc. v. UCR, Inc., 4 CCH Computer Cases 46,752 at para. 94 (M.D. Ga.
1992) [hereinafter CMAX], Fitzpatrick D.J. stated that the decision in Whelan is “conceptually over-
broad” and “descriptively inadequate”, preferring instead to follow the three-part test outlined in Al-
tai.

5 8Altai, supra note 33 at 1252-53.
“‘ See CMAX, supra note 57; Lotus Development Corporation v. Borland International, Inc., 788

F.Supp. 78 (D. Mass. 1992) [hereinafter Borland].

‘ See the decisions in Delrina, supra note 5 and in Richardson, supra note 36 with respect to Can-

ada and the United Kingdom respectively.

61 CopyrightAct, supra note 1, s. 3.
62

Although it is not mandatory that one register one’s copyright in the work, it is consid-
ered prudent to register a work for several reasons. First, registration provides an evi-
dentiary record of the work should a dispute as to its authorship ever arise, and second
the copyright holder may have a broader range of remedies available should there be
an infringement of the copyright (S. Handa & J. Buchan, “Copyright as It Applies to

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

ing requirement” for a work to be copyrightable; copyright is said to subsist upon
the creation of the work.6 The term for which copyright attaches is the life of the
author, or in the case of joint authorship of a work65 the life of the longest surviving
author, plus fifty years.6 The author of a work is presumed to be the first owner of
the copyright therein, except where a work is created under a contract of service
whereby the employer is presumed to be the first owner of the work.

Originality

Merely expressing oneself in one of the protected forms enumerated above may
not in itself be sufficient to obtain copyright protection. The Copyright Act requires
that protectable works, at a minimum, demonstrate a modicum of originality.
Originality in a work refers to the degree of the author’s creative or inventive
thought, and is a comparatively less rigorous requirement in common law copyright
jurisdictions as compared with continental droit d’auteur jurisdictions such as
Germany or France. 9 In effect, under common law copyright systems, in order to
demonstrate originality one need only show that the work originated from the
author and was not a copy of an existing work.

the Protection of Computer Programs in Canada” (1995) 26 I.I.C. 48 at 51-52).

If a computer program copyright is registered, there is no requirement that the copyright holder file
the detailed source code specification with the copyright office. In fact the Canadian Copyright Office
will not accept attachments when registering the copyright in a work.

67 “Canada is a long-standing signatory to the Beme Convention on Copyright (Rome Revision,
1928), which prohibits any requirement that works be registered or that the “” symbol be used in
conjunction with expressions of the work in order for copyright protection to apply” (Handa & Bu-
chan, ibid at note 11, p. 51).

64Copyright Act, supra note 1, s. 5.
‘Joint authorship exists where the “contribution of one author is not distinct from the contribution

of the other author or authors” (ibid, s. 2.9).

“Ibid., ss. 6,9(1).
67 Ibid., s. 13(3). A “contract of service”, which denotes an employment relationship in the tradi-
tional sense, is to be differentiated from a “contract for services”, which refers to an independent con-
tractor who has arranged to produce a work under a specific contract as opposed to a general em-
ployment contract. In either case the presumption created is rebuttable at law (see Orbitron Software
Design Corp. v. M.C.R. Systems Ltd (1990), 48 B.L.R. 147, 32 C.PR. (3d) 414 (B.C.S.C.); Positron
Inc. v. Desroches, (1988] R.J.Q. 1636 (S.C.)).

68 Copyright Act, ibid., s. 5.
“Handa & Buchan, supra note 62 at 52.
7′ The word “original” does not in this connection mean that the work must be the expression of a
creative or inventive thought Copyright acts are not concerned with the originality of ideas, but with
the expression of thought, and, in the case of “literary work”, with the expression of thought in print
or in writing. The originality which is required relates to the expression of the thought (University of
London Press Ltd v. University Tutorial Press Ltd, [1916] 2 Ch. 601 at 608; see also Delrina, supra
note 5 at 32).

1995]

S. HANDA – REVERSE ENGINEERING

Fixation

In order to be worthy of copyright protection, a work must also be fixed or
stored in some manner.7 For computer programs, section 2 of the Act explicitly re-
quires that a computer program be “expressed, fixed, embodied or stored in any
manner” for copyright protection to apply. The American Copyright Act is slightly
more specific, in that it requires a work to be expressed in a form which is suffi-
ciently permanent and stable such that it may be “perceived, reproduced, or other-
wise communicated, either directly or with the aid of a machine or device.’ 72

Issues of fixation arise frequently in relation to computer programs and without
further legislative or juridical guidance, a good deal of legal uncertainty will per-
sist. Recent jurisprudence addressing the protection of non-literal elements, such as
computer screens, suggests that fixation in volatile memory devices, such as a
video interface’s RAM,
is sufficient to meet the fixation requirement. For exam-
ple, computer screens are protected because the screens are said to exist under the
umbrella of the underlying computer program’s copyright.75 A more difficult ques-
tion is whether transitory combinations of data, such as the results of a database
search conducted at the direction of a user, are sufficiently fixed for the purposes of
copyright. These results are often only stored in volatile memory, and cannot be
said to be part of the underlying search engine as the user’s search criteria are en-
tered only upon use.

In terms of the reverse engineering debate, the issue of fixation is important
since in order to reverse engineer a computer program, one must first “copy” that
program into a computer’s RAM memory to allow for disassembly. As the disas-
sembler performs its passes through the program, it makes increasingly precise
translations of the program, which are often stored only in RAM. Once the disas-

7, In Canadian Admiral Corp. v. Rediffusion Inc., [1954] Ex. C.R. 382 at 394 [hereinafter Rediffu-
sion], the Court ruled that “for copyright to subsist in a ‘work’ it must be expressed to some extent at
least in some material form, capable of identification and having a more or less permanent endur-
ance.”

” Copyright Act (U.S.), supra note 6, 102.
7′ RAM, also known as volatile or dynamic memory, is a form of internal memory that stores in-
formation as long as electrical impulses are being fed through it. Once the power is cut off, the RAM
microchips lose all their stored information. RAM is the functional memory that allows a computer to
operate, since the programs or parts thereof which the computer processes must be stored in RAM
during the operation of the computer (with the exception of ROM programs). If programs are other-
wise stored on extemal memory devices, they must be copied into RAM memory in order to be exe-
cuted.

7
1 See Delrina, supra note 5.
‘5 Note, however, that a fleeting image broadcast on television has been held not to fulfil the re-
quirements of fixation for the purposes of copyright. If the television program is otherwise fixed, such
as on some form of video tape, then the fleeting image described may be protected as part of the un-
derlying copyright of the fixed program (Rediffusion, supra note 71 at 396).
76 Depending on the disassembler and the size of the program code being disassembled, these in-
termediate translations may be temporarily stored on disk, at which point fixation issues become ir-

MCGILL LAW JOURNAL/REvuE DE DROIT DE MCGILL

[Vol. 40

sembly is completed, the results are usually stored on more permanent media such
as a disk, and in print-out form. It is, however, possible to disassemble a program,
or parts thereof, without storing either the intermediate copies or final result in any
media other than RAM. Although this may seem impractical since the results will
remain within the dynamic RAM of the computer, it may provide a technical way
around the difficulties of copyright infringement since without adequate fixation,
copyright laws may not consider that a copy has indeed been made. Furthermore, if
it is deemed that, even where the final product of disassembly is fixed, the inter-
mediate copies stored in RAM do not constitute infringing copies for want of fixa-
tion, the exception to copyright found in paragraph 27(2)(1) of the Canadian Copy-
right Act, discussed below, may apply.”

Infringement of Copyright

Once copyright subsists in a work, it will be infringed by “any person who,
without the consent of the owner of the copyright, does anything that, by th[e
Copyright Act], only the owner of the copyright has the right to do. 7
1 It must be
stressed that copyright only prevents the copying of a work, or a substantial part
thereof;n the monopoly rights granted by the Copyright Act do not extend to situa-
tions where a person independently creates a similar work. Accordingly, the juris-
prudence dealing with copyright infringement contends that the onus to demon-
strate copying rests on the plaintiff, who must establish both a substantial similarity
between his or her work and that of the defendant, as well as a causal connection
between the two worksY The causal connection requirement can be satisfied by
demonstrating that the defendant had access to the original work. Once the plaintiff
has discharged the burden, a rebuttable presumption arises whereby the onus of
proving independent creation shifts to the defendant. The defendant may also at-
tempt to rely on the exceptions to infringement contained in subsection 27(2) of the
Copyright Act.

relevant.

Paragraph 27(2)(1) allows a single reproduction of an authorized copy of a computer program to
be made for purposes of compatibility. In order to use that paragraph to exempt disassembly from in-
fringing copyright, the intermediate copies would have to somehow be exempt. If intermediate copies
were not qualified as fixed, they could thus be exempt from copyright (see Part HA. 1.f.i, below).

“‘ Copyright Act, supra note 1, s. 27(1).
‘ “Substantial” refers not to a strict percentage, but rather to the quality of the part taken (Breen v.
Hancock House Publishers Ltd. (1986), 6 C.I.PR. 129 at 133, 6 C.PR. (3d) 433 (F.C.T.D.)
[hereinafter Breen cited to C.I.P.R.]). In SAS, supra note 41 at 822, the Court found that 44 examples
of copying had occurred out of a total of approximately 186,000 lines of computer source code. It
held that this constituted a substantial taking, and that the small number of instances of copying did
not render the copying trivial.

‘0 Gondos v. Hardy (1982), 38 O.R. (2d) 555, 64 C.ER. (2d) 145 (H.C.J.); Francis Day & Hunter

Ltd. v. Bron, [1963] Ch. 587 (C.A.).

1995]

S. HANDA – REVERSE ENGINEERING

Exceptions to Infringement

Translation – Modification Exception

If the intermediate copies or final results of the disassembly process are indeed
considered fixed, the copying of a computer program for use by a computer may be
exempted by paragraph 27(2)(1) of the Copyright Act. The crucial issue is whether
such use includes disassembly. This argument is strengthened where the intermedi-
ate copies avoid infringement due to lack of fixation, and only the single final dis-
assembled product is fixed. Paragraph 27(2)(1) states that infringement does not oc-
cur as a result of

the making by a person who owns a copy of a computer program, which copy
is authorized by the owner of the copyright, of a single reproduction of the
copy by adapting, modifying or converting the computer program or translat-
ing it into another computer language if the person proves that

(i)

the reproduction is essential for the compatibility of the computer pro-
gram with a particular computer,
the reproduction is solely for the person’s own use, and

(ii)
(iii) the reproduction is destroyed forthwith when the person ceases to be the

owner of the copy of the computer program.”

In drafting this paragraph, “[t]he House of Commons Sub-Committee on the Revi-
sion of Copyright recognized that it is common in the industry for computer pro-
grams to be adapted or modified to meet the particular needs of end users.””2 Ac-
cordingly, disassembly was not contemplated as being within the scope of the sec-
tion.

Nevertheless, this paragraph may, in fact, also be used to support the contention
that, in order to function, computer programs must be copied into parts of a com-
puter (usually into RAM) and translated by the central processing unit into mi-
crocode.8. Although it can be argued that paragraph 27(2)(1) protects a computer

8, Computer programs, however, are generally licensed and only rarely sold by the copyright holder.
See infra note 87 and accompanying text for a discussion of licensing, as contrasted with sale, of
software products.

Sookman, supra note 48 at 3-203.
A similar exception was placed into the American Copyright Act in 1978. The purpose of section
117 was to give authorized users the right to use a computer program without technically infringing
the copyright in the program. The section reads:

Notwithstanding the provisions of section 106, it is not an infringement for the owner
of a copy of a computer program to make or authorize the making of another copy or
adaptation of that computer program provided:

(1) that such a new copy or adaptation is created as an essential step in the utiliza-
tion of the computer program in conjunction with a machine and that it is used
in no other manner…

The American exception has, however, been held not to apply to the reverse engineering of com-

puter programs (Sega, supra note 4 at 1520-21; see also text accompanying note 204, below).

MCGILL LAW JOURNAL/REVUE DE DRO1TDE MCGILL

[Vol. 40

program’s use from being declared an infringement, it is more tenuous to argue that
this paragraph also applies where a program is copied into RAM and subsequently
onto more permanent media for the purposes of dissection by a disassembler.4
Where the simple use of a program is concerned, however, even if paragraph
27(2)(1) does not apply, it is unlikely that the copying of the program for the pur-
poses of use will be declared an infringement because the conversion of object code
into electrical signals may not be a “reproduction of the work in a material form.””
The fact that the simple intended use of an authorized copy of a computer program
could potentially constitute a copyright infringement provides a good illustration of
the inappropriateness of copyright as the principal form of intellectual property
protection for computer programs.”

ii.

Making Backup Copies

In addition to the translation/adaptation/modification exception contained in
paragraph 27(2)(1), there are two other exceptions to copyright infringement that
also apply to computer programs. The more specific of these exceptions is in para-
graph 27(2)(m) of the Copyright Act, and allows the owner of an authorized copy
of a computer program to make a single copy for backup purposes, but requires that
this backup be destroyed if the person ceases to be the owner of the authorized
copy. Practically speaking, this provision, in addition to paragraph 27(2)(1), has
proven to be of limited use because software companies seldom transfer the owner-
ship of their software, preferring instead to license it to users. Licensing effectively
allows the software companies to supersede certain provisions of the Act, such as
this backup exception to infringement, by manipulating terms of their licensing
agreement.87 Most software companies do, nonetheless, permit the making of a

The right [under paragraph 27(2)(1)] to convert a computer program or translate it into
another computer language will probably give a person who owns an authorized copy
of a computer program the right to convert the program from one higher-level language
to another. It might also give such a person the right to convert a program from a
higher-level language to machine language, and vice-versa, but the meaning of the
term ‘translation’ in the Act is still uncertain … and so the scope of [27(2)(1)] … is still
not known (Sookman, supra note 48 at 3-204 [emphasis added]; see also infra note
107 and accompanying text).

Apple, supra note 8 at 30.

“Sookman, supra note 48 at note 17 and accompanying text, p. 3-3.
” As a result of the mass production of off-the-shelf software, it has become impossible for soft-
ware publishers and vendors to individually negotiate licence terms with each prospective purchaser;
“shrink-wrap” licences were therefore developed. These licences consist of a list of licensing terms
placed visibly on a product (usually under the cellophane wrapper). Typically, the licence contains a
clause that states that if the purchaser does not agree to abide by the terms of the licence he or she
should not purchase the product, and, further, that by opening the package the purchaser agrees to
abide by its terms. It is unclear whether shrink-wrap licences are enforceable in Canada even though
they are an extremely common industry practice. In an oft-cited passage from Betts v. Wilntott (1871),
L.R. 6 Ch. App. 239, the Court held that “[w]hen a man has purchased an article he expects to have
the control of it, and there must be some clear and explicit agreement to the contrary to justify the
vendor in saying that he has not given the purchaser his licence to sell the article, or to use it wherever

1995]

S. HANDA – REVERSE ENGINEERING

backup copy in their licensing agreement as it avoids the inconvenience of replac-
ing programs if and when the carrying media become defective.,

Licensing terms can also be used to prevent the reverse engineering of the
computer program by the user. If the user is a licensee rather than the owner of a
copy of the computer program, then prima facie the licence terms will prevail
notwithstanding the permissibility of reverse engineering under the Copyright Act.
If a licence is silent or ambiguous as to its interpretation, the court will fill in the
gaps using implied terms reflective of industry practice. Licence terms that run
afoul of the Copyright Act have not been tested in Canadian courts; however, the
experience in the United States suggests that such terms may not be enforceable.,

iii.

Fair Dealing Under the Copyright Act

Another more general copyright exception is the fair dealing exception found in
subsection 27(2) of the Copyright Act. According to that subsection, “any fair
dealing with any work for the purposes of private study [or] research” will not
constitute copyright infringement. The Act does not further detail the breadth of the
fair dealing exception, and very little jurisprudence has elaborated upon its interpre-
tation. Similar exceptions exist in British copyright legislation, where the term “fair
dealing” is also used, and more importantly for the purposes of reverse engineering
computer programs, in the American copyright legislation where the term “fair use”
is used.9′ As a result of the paucity of Canadian jurisprudence on the subject of fair
dealing, it is impossible to say how similar our exception will be to those of other
jurisdictions. 92 In the context of Canadian cases “related to copyrightability of
computer programs, American authorities have been cited and, notwithstanding the
differences between the wording of the Acts, have been given qualified approval.”93

he pleases as against himself.” In North American Systeinshops Ltd. v. King (1989), 97 A.R. 46, 68
Alta. L.R. (2d) 145 (Q.B.), the Court found unenforceable a shrink-wrap licence agreement not visi-
ble to the purchaser at the time of purchase. It remains to be seen whether visible shrink-wrap licences
fulfil the requirement of a “clear and explicit agreement”.

s Unlike in the Canadian and American Copyright Acts, a purchaser cannot contract out of similar
backup copy provisions contained in the British Copyright, Designs and Patents Act 1988 (U.K.),
1988, c. 48 (see ifra note 229 and accompanying text).

” As with any new industry, the difficulty with implied licences in the computer software industry
is that business practices often are only partially established at the point when a court is forced to de-
cide upon a relationship with implicit terms.
90 See Part ll.C. 1, below, for a detailed discussion of trade secrets and licensing.
9′ United Kingdom: Copyright, Designs, and Patent Act 1988, supra note 88, s. 29(1); United
States: Copyright Act, supra note 6, 107 (see infra note 193 for the text of the section).
92 It is widely believed that “fair dealing” and “fair use” are different concepts. With respect to the
home taping issue decided by the United States Supreme Court in Sony Corporation of America v.
Universal City Studios, hzc., 464 U.S. 417, 78 L. Ed. (2d) 574 (1984) [hereinafter Sony cited to U.S.],
“there are sufficient differences between the American ‘fair use’ defence and Canada’s ‘fair dealing’
to conclude that, if an action were brought in Canada, home taping would be found to constitute a
copyright infringement” (M. HMbert, Copyright Act Refonn, rev. ed. (Ottawa: Research Branch, Li-
brary of Parliament, 1987) at 11).

93Sookman, supra note 48 at 3-6.

MCGILL LAW JOURNAL/REVUE DE DROIT DE MCGILL

[Vol. 40

No court has yet performed a comprehensive comparative analysis of the Canadian
“fair dealing” and the American “fair use” exceptions.

The earliest Canadian case to deal extensively with the fair dealing exception
was Zamaco’s v. Douville.94 In that case, the Exchequer Court laid out the basic
principles that govern fair dealing: a verdict of fair dealing must depend on the
specific facts of each case; the copying of an entire work cannot qualify as fair
dealing; short of copying the entire work, the quantity of the work copied is not
solely determinative of fair dealing; and “in considering whether a dealing with a
particular work [is] fair, it would have to be considered whether any competition
[is] likely to exist between the two works.” 95 The other two Canadian cases to deal
with the defence of fair dealing both faced the question whether an abridgment or
summary of a work could avoid being declared an infringement through a claim of
fair dealing. 96 In both cases, the courts held that merely summarizing a work with-
out the addition of some further comment is not fair dealing.

In Hubbard v. Vosper, the United Kingdom Court of Appeal stated:

[I]t is impossible to define what is ‘fair dealing’. It must be a question of de-
gree. You must consider first the number and extent of the … extracts. Are they
altogether too many and too long to be fair? Then you must consider the use
made of them. If they are used as a basis for comment, criticism or review, that
may be a fair dealing. If they are used to convey the same information as the
author, for a rival purpose, that may be unfair. Next, you must consider the
proportions. To take long extracts and attach short comments may be unfair.
But, short extracts and long comments may be fair. Other considerations may
come to mind also. But, after all is said and done, it must be matter of impres-
sion.9

In Beloff v. Pressdram Ltd.,9 the United Kingdom Chancery Court expanded
the application of the fair dealing defence to unpublished and confidential informa-
tion. In Beloff, the defendant newspaper company published an internal office
memorandum written by the plaintiff without obtaining the plaintiff’s authorization.
The defendant claimed that its purpose in publishing the plaintiff’s work was to
criticize it and that it was therefore covered by the fair dealing exception under the
Copyright Act. The Court held that the fact that the memorandum had not been
published was not in itself enough to find in favour of the plaintiff; however, it was
an aggravating factor to be taken into account in assessing the defendant’s conduct.
With respect to the confidentiality of the information used, the Court held that

9′ (1943), 3 Fox Pat. C. 44, [1943] 2 D.L.R. 257 (Ex. Ct.) [hereinafter Zanacors cited to Fox Pat.

C.].

9 lbiL at 72-74.
96Breen, supra note 79 at 133; R. v. James Lorimer & Co., [1984] 1 EC. 1065 at 1077-78, 77 C.P.R.

(2d) 262 (C.A.) [hereinafter Lorimer cited to F.C.].

[1972] 1 All E.R. 1023 at 1027 (C.A.) [hereinafter Hubbard].
[1973] 1 All E.R. 241 (Ch.) [hereinafter Beloff].

1995]

S. HANDA – REVERSE ENGINEERING

[t]he vice of the leak of the publication in this case was, to my mind, clearly
unjustifiable for the authorised purposes of criticism, review and news, and
clearly in my view constituted dealing which was not fair within the statute….
This ground is ample to defeat the defence of fair dealing.99

Each of the aforementioned cases dealt with fair dealing in the context of tradi-
tional literary works. No case in Canada or in the United Kingdom has applied the
defence of fair dealing in the context of a computer program. The American experi-
ence with the fair use exception has been quite different and will be discussed in
further detail in Part I.

iv.

Public Interest Exception

A final potential exception to copyright infringement in reverse engineering
cases is the public interest defence. This defence is judicially created, and does not
expressly appear in Canadian copyright legislation.” The public interest defence
has not yet been raised successfully in Canada with respect to copyright, and has
only gained judicial recognition in one small passage in Lorimer. In that case, Ma-
honey J. stated:

I have no doubt that a defence of public interest as enunciated in the English
cases is available in proper circumstances against an assertion of Crown copy-
right. … [However t]his is not a “public interest” case in the same sense as the
English decisions nor, really, in the sense the defence was advanced here.”

The English cases to which Mahoney J. was referring were Hubbard and Beloff. In
both of those cases, the courts recognized that a common law defence of public in-
terest was available notwithstanding its lack of statutory support. In this regard,
Ungoed-Thomas J. in Beloff stated that the defences of public interest and fair
dealing are

separate defences [which] are governed by separate considerations. Fair deal-
ing is a statutory defence limited to infringement of copyright only. But public
interest is a defence outside and independent of statutes, is not limited to copy-
right cases and is based on a general principle of common law.0 2

The discussion of a public interest defence in the context of a copyright in-
fringement claim was further examined in Lion Laboratories Ltd. v. Evans.’0 3 In
that case, the United Kingdom Court of Appeal had to decide whether the theft of

Ibid. at 264.

‘ Subsection 171(3) of the United Kingdom’s Copyright, Designs and Patents Act 1988 expressly
recognizes the existence of a public interest defence and states: “Nothing in this Part affects any rule
of law preventing or restricting the enforcement of copyright, on grounds of public interest or other-
wise.”

“‘ Loriner, supra note 96 at 1078-79.
12 Beloff, supra note 98 at 259.
‘0’ [1985] Q.B. 526 (C.A.).

MCGILL LAW JOURNAL/REVUE DE DROITDE McGILL

[Vol. 40

confidential literature that was subsequently published without the copyright
holder’s permission, could be held to be in the public interest. The literature in
question was an internal memorandum suggesting that a breathalyzer device manu-
factured by the plaintiff was capable of giving false readings which may have led to
the conviction of innocent persons. The defendants did not deny that they took the
confidential information without the plaintiff’s permission, or that its publication
did not prima facie infringe the plaintiff’s copyright; the only defence presented
was one of public interest. In deciding how to weigh the public interest against the
copyright infringement and breach of confidence perpetrated by the defendants,
Stephenson L.J. stated:

To be allowed to publish confidential information, the defendants must do
more than raise a plea of public interest; they must show “a legitimate ground
for supposing it is in the public interest for it to be disclosed.” … [W]e “should
not restrain it by interlocutory injunction, but should leave the complainant to
his remedy in damages” 4

Griffiths L.J. agreed, and in assessing the applicability of the public interest defence
to copyright infringement, stated: “I am quite satisfied that the defence of public
interest is now well established in actions for breach of confidence and, although
there is less authority on the point, that it also extends to breach of copyright.””

Although the jurisprudence dealing with public interest is solidly in place in the
United Kingdom, the sole mention in Canada of a public interest copyright defence
considered its possible existence only in the context of Crown copyright. It would,
however, be nonsensical if it did not also extend to defend against claims by private
copyright holders. As it stands, a public interest defence seemingly exists in Can-
ada, independently of any statutory exception to copyright. ”

It will be shown that both the fair dealing exception and the public interest de-
fence are essential to an argument in support of reverse engineering. With respect to
the former, the American fair use exception has been successfully applied in re-
verse engineering cases, and notwithstanding the differences between the respective
exceptions, fair dealing presently remains the most likely candidate to allow the re-
verse engineering of computer programs under Canadian copyright law. If the fair
dealing defence fails to support an argument allowing reverse engineering, then a
more tenuous, although nonetheless plausible, claim may be made under the prin-
ciple of public interest. Copyright, however, may not be. the sole hurdle to the re-
verse engineering of computer programs. Additional protections, such as those

,04 Ibid. at 538, quoting from Khashoggi v. Smith (15 January 1980), No. 58 at 15 (C.A.), Sir David

Cairns and Woodward v. Hutchins, [1977] 1 W.L.R. 760 at 764 (C.A.), Lord Denning M.R.

‘o5 Ibid. at 550.
‘6Although section 63 of the Copyright Act effectively abolishes common law copyright, it makes
no mention of common law defences to copyright infringement. The existence of an implied public
interest defence may be rooted in the modem-day general underpinnings of copyright law protection,
which seek to balance the public’s right to knowledge with the individual’s right to be remunerated
for work (Sookman, supra note 48 at note 2 and accompanying text, p. 3-1).

1995]

S. HANDA – REVERSE ENGINEERING

raised under trade secrets law or by licensing provisions which seek to expand the
scope of copyright protection, may also frustrate any attempt to reverse engineer a
computer program.

B. Is Reverse Engineering an Infringement Of Copyright Law?

With the fundamental principles in place, we are now prepared to consider
whether reverse engineering is indeed an infringement of copyright law, and if so,
how. As mentioned above, when reverse engineering a computer program, the dis-
assembler must first load a copy of the program into a computer’s memory.’0 This
results in the first potentially infringing copy. As the disassembler makes passes
over the program, it will continue to produce translations of the work which consti-
tute potentially infringing copies of the program according to paragraph 3(1)(a) of
the Copyright Act. These copies are referred to as intermediate copies.’ Once the
disassembler has completed its task, it will produce an ASSEMBLER language
version of the computer program. This ASSEMBLER source code constitutes yet
another potential infringement of the computer program’s copyright, although the
assembler will not generally reconstitute a program in the exact fashion in which it
was written. The resulting source codes will be a translation rather than a reproduc-
tion;’09 but either would potentially infringe the original work’s copyright.” Finally,
making a hard, or printed, copy of the work for further examination of the com-
puter program’s operating principles will also potentially infringe copyright.”‘ The
difficulty with reverse engineering a computer program lies in the fact that the pro-
gram must be put into memory for decompilation. The act of reverse engineering is
thus not in itself a violation of copyright, but rather the means which it employs
are.”‘ This article posits, however, that because of the need to develop standards
and achieve program compatibility, reverse engineering can be contemplated within

,07 Although the operating system also copies a computer program io execute the program, this ac-

tion is not regarded as an infringement of copyright (see supra notes 84-85 and accompanying text).

“‘z Atari, supra note 4 at 842.
‘9 In the case of computer programs, the distinction between a reproduction and a translation for the

purposes of copyright law is irrelevant (Apple (S.C.C.), supra note 8 at 261).

“o In the United States, the copies of the computer program made by the disassembler in final form

would be termed “derivative works”. A derivative work is

a work based upon one or more preexisting works, such as a translation, musical ar-
rangement, dramatization, fictionalization, motion picture version, sound recording, art
reproduction, abridgment, condensation, or any other form in which a work may be re-
cast, transformed, or adapted. A work consisting of editorial revisions, annotations,
elaborations, or other modifications which, as a whole, represent an original work-of
authorship, is a “derivative work” (Copyright Act (U.S.), supra note 6, 101).

… Ignatin, supra note 17 at 2011-12.
“2 Reverse engineering of other products can be accomplished without any copying, simply by ob-
taining the necessary examples, taking them apart, studying and testing them (ibiei at note 50, p.
2011). The idiosyncratic application of copyright law to computer program products makes the in-
spection of these products a potential infringement Once again, it is apparent that copyright may not
provide the optimal manner by which to protect these products.

MCGILL LAW JOURNAL/REVUE DE DROIT DE MCGILL

[Vol. 40

the scope of the fair dealing provision in the Copyright Act.”” Furthermore, it will
be argued that intermediate copies produced during reverse engineering should be
exempted from copyright through a statutory exception, since the prohibition of re-
verse engineering falls beyond the scope of copyright protection. The rationale for
these arguments will be more fully discussed in Part IV.

C. Reverse Engineering Under Other Legal Regimes

Trade Secrets

The law governing trade secrets is intertwined with the law of copyright insofar
as it applies to reverse engineering. Trade secret law is judge-made law that pro-
tects commercial confidences.14 It can coexist with other intellectual property pro-
tections, and in the case of literary, dramatic, artistic or musical matter, may even
protect the underlying ideas which are not in themselves copyrightable.” ‘ Trade se-
cret protection covers a broader range of informational elements than do other
forms of intellectual property protection, but does so in a more limited manner.

Canadian and British courts have applied three general requirements for a suc-
cessful trade secret claim: (I) the information must have the necessary quality of
confidence; (2) the information must have been imparted in circumstances import-
ing an obligation of confidence (i.e. a “special relationship” must exist between the
parties); and (3) there must be an unauthorized subsequent use of that information
to the detriment of the party communicating it.”6 The burden of proving the secrecy
of the information falls on the plaintiff.” 7 However, once the information loses its
quality of secrecy through authorized disclosure, legitimate independent research or
any other honest means, its purveyor may no longer avail him or herself of trade
secret protection.

… In 1984-85, the Canadian government did consider implementing in the Copyright Act what ef-
fectively amounted to a statutory exception which would have allowed reverse engineering of com-
puter programs. This proposal was rejected (see infra note 257).

“‘ The possessor of the secret neither has to go to unreasonable lengths to maintain secrecy
(Creditel of Canada Ltd. v. Faultless (1977), 18 O.R. (2d) 95, 81 D.L.R. (3d) 567 (H.C.J.)), nor does
he or she have to guard against unanticipated, undetectable or unpreventable methods of discovery
(International Corona Resources Ltd v. Lac Minerals Ltd. (1986), 53 O.R. (2d) 737, 25 D.L.R. (4th)
504 (H.C.J.)). See also Sookman, supra note 48 at 4-29.

“‘ According to the Restatement of Torts 757b (1939), “[a] trade secret may consist of any for-
mula, 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 know or use it.” This defini-
tion was accepted in Canada in R.L Crain Ltd. v. Ashton (1949), O.R. 303 at 308, [1949] O.W.N. 246
(H.C.J.), aff’d [1950] O.R. 62, [1950] 1 D.L.R. 601 (C.A.).

5 See Q-Co Industries, Inc. v. Hoffnan, 625 FSupp. 608 (S.D. N.Y 1985).
116 See Lac Minerals Ltd. v. International Corona Resources, [1989] 2 S.C.R. 574, 69 O.R. (2d) 287
[hereinafter Lac Minerals cited to S.C.R.]; Sofhvare Solutions Associates Inc. v. Depow (1989), 99
N.B.R. (2d) 110, 250 A.P.R. 110 (Q.B.); Coco v. A.N. Clark Ltd., [1969] R.P.C. 41 (Ch.); Ridgewood
Resources Ltd v. Henuset (1982), 35 A.R. 493, 18 Alta. L.R. (2d) 68 (C.A.).

”

19951

S. HANDA – REVERSE ENGINEERiNG

Trade secret protection is often considered to be based on the laws of property,
contracts and/or trusts. In Canada, Sopinka J. stated that the legal basis for trade se-
cret actions is “sui generis relying of each of these areas to enforce the policy of the
law that confidences be respected.”” 8 This law of confidentiality arises from an ob-
ligation of good faith in commercial settings or from a fiduciary relationship.
Where no fiduciary relationship exists, the courts will look for a fiduciary quality to
the relationship which would require an element of confidence. For example, the
relationships between manufacturers and designers, licensors and licensees, joint
venturers intending to do business with one another, and employees may all fall
within the ambit of trade secret laws.

Section 63 of the Copyright Act states that the Act is to be the sole source of
copyright, but the section also contemplates the existence of trade secret laws that
may operate independently of the copyright legislation.”‘ With respect to the re-
verse engineering debate, the applicability of trade secret laws is generally ensured
through licensing agreements or contracts between the parties.

Although section 63 most certainly allows trade secret laws to supplement
copyright principles, it is not clear whether trade secret laws would be paramount in
a situation of conflict between the two laws.2 It has been suggested that in the case
of reverse engineering computer programs, the true aim of trade secret restrictions
“is not preservation of confidentiality or security against disclosure to third parties;
it is foreclosure of competition.”

Clearly, an agreement purported to restrain competition, rather than to protect
against a breach of trust or confidence, is not contemplated by section 63 and will
infringe both copyright and competition laws. 2 Practically speaking, the intent of
most contracts that restrain reverse engineering supports both objectives. In such a
case, it is not clear whether a court would rely solely on the primary purpose of the
agreement, or whether an incidental objective to restrain a breach of confidence or
trust would suffice to support the paramountcy of the terms of the agreement over
copyright law.

“a Lac Minerals, supra note 116 at 615. Aside from the common law basis for dealing with trade
secrets, several American states have enacted trade secret statutes (see M.D. Stein, “The Importance
of a Trade Secret as a Supplement to Copyright Protection of Computer Software” (1993) 12:1 I.EL.
Newsletter 28 at 29).

“9 The text of section 63 reads:

No person is entitled to copyright or any similar right in any literary, dramatic, musical,
or artistic work otherwise than under and in accordance with this Act, or of any other
statutory enactment for the time being in force, but nothing in this section shall be
construed as abrogating any right or jurisdiction to restrain a breach of trust or confi-
dence.

See infra note 154 and accompanying text.
Rice, supra note 26 at 623.

‘ See Part ILC.2.a, below.

MCGILL LAW JOURNAL/REvuE DE DROITDE MCGILL

[Vol. 40

When computer programs are mass produced, however, the situation is differ-
ent because no negotiated agreements are involved. If an object code copy of com-
puter software is released to the public without a licensing agreement, there is no
“special relationship” between the parties, and for the purposes of trade secrects
law, members of the public are free to reverse engineer the product to determine the
source code. In the absence of a copyright prohibition on reverse engineering
(imparting an obligation of confidence), anyone not in a “special relationship” with
the computer program owner is legally free to attempt to reverse engineer a lawful
copy of the computer program.’ Generally, however, mass-marketed computer
programs use shrink-wrap license agreements to enhance their copyright.124

Enhancing Copyright Protection Through Licensing/Contract Law

Difficulties immediately arise with the use of shrink-wrap licences to enhance
copyright protection. First, with mass-marketed software it is unclear whether
shrink-wrap licensing is valid. For the purposes of a trade secrets argument, it
seems unlikely that a shrink-wrap agreement would sufficiently constitute the
“special relationship” required between the software developer and the purchaser
for trade secret protection. This type of relationship is more likely to exist where
the licensing terms have been negotiated between the vendor and purchaser. Sec-
ond, if reverse engineering is deemed to be in the public interest under copyright
legislation, trade secret protection, as a creature of the common law, might not be
extended to cover the reverse engineering of computer programs. This would be
especially likely if the purpose of the licence agreement were found to be a restraint
of competition. The following three subsections set out defences that may be used
where a copyright holder attempts to prevent a user from reverse engineering a
computer program through licensing or contract provisions, assuming that such re-
verse engineering is permitted under the Copyright Act.

The general question whether contract/licensing law can indeed supersede
copyright law provisions in a situation of conflict is yet another issue that has still
to be comprehensively addressed by the courts. Specifically, if copyright law sup-
ports a limited reverse engineering exception, it is debatable whether protection

“In GEAC Canada Ltd v. Prologic Computer Corp. (1989), 35 B.C.L.R. (2d) 136, 25 C.PR. (3d)
91 (S.C.), the plaintiff alleged that one of the defendants, who was in a software licensing relationship
with the plaintiff, had violated “its common-law duty of confidentiality” to the plaintiff by allowing
another of the defendants to reverse engineer the licensed software, and to use the results in the crea-
tion of compatible and even competing products. The plaintiff’s claim was rejected; however, the
Court based its ruling solely on the fact that the plaintiff was guilty of laches and acquiescence. The
Court did not express an opinion on the issue of reverse engineering as it relates to trade secrets per
se, although it did state that, broadly speaking, the claims raised “serious issues to be tried.” No claim
concerning a potential infringement of copyright was made by the plaintiffs, and, accordingly, the
Court in no way addressed any issue related to copyright.

“4 See supra note 87 for a discussion of shrink-wrap licensing.

19951

1 S. HANDA – REVERSE ENGINEERING

obtained through licensing can be used to override the copyright exception. Al-
though Canadian courts have not yet faced this problem, American courts have de-
veloped a “copyright misuse doctrine” that may potentially be used to pre-empt the
“enforcement of software license terms that prohibit reverse engineering.”” An
attempt to widen copyright law beyond its limited statutory scope through contract
would, under the misuse doctrine, render any attempt to enforce one’s copyright in-
valid. 2

The copyright misuse doctrine was recently applied by the United States Fed-
eral Court of Appeals in Lasercomb America, Inc. v. Reynolds,27 where the Court
held that the misuse doctrine extends to render “a copyright unenforceable against
any person regardless of whether they entered into a contract containing the offend-
ing term.”‘2 18 Under the ruling in Lasercomb, merely attempting to widen the scope
of copyright protection beyond its accepted limits is a bar to its use. Although the
United States Supreme Court has not applied the copyright misuse doctrine, it did
acknowledge its existence in United States v. Loew’s Inc.’29 The fact that the misuse
defence is grounded in equity means that claimants must show clean hands in order
to use it,I and that the defence is potentially available to Canadian litigants.’ 3′

In fact, the existence of a copyright misuse defence has been implicitly
acknowledged by the Supreme Court of Canada in Massie & Renwick Ltd. v.
Underwriters’ Survey Bureau Ltd. 32 In that case, the defendants, accused of copyright
infringement, alleged that the plaintiffs’ withholding of their plans and insurance rating
schedules constituted “a combine and conspiracy” under both the Combines
Izvestigation Act’ and the Criminal Code. T he trial judge stated that he did not have
to rule on the misuse defence because the defendants failed to demonstrate that the
plaintiffs were guilty of anti-competitive behaviour. In obiter dictum, the trial judge
held:

‘2 Rice, supra note 26 at 551.
126 Ibid.
‘ 911 F.2d 970 (4th Cir. 1990) [hereinafter Lasercomb].
‘”Rice, supra note 26 at note 24, p. 551.
“9 371 U.S. 38,9 L. Ed. (2d) 11 (1962). See also Atari, supra note 4 at 846.
‘”See Atari, ibid.
“‘ Principles of both common law and equity apply to actions concerning federal laws in Canada

(Aldrich v. One Stop Wdeo Ltd. (1987), 13 B.C.L.R. (2d) 106, 17 C.PR. (3d) 27 (S.C.)).

“3’ [1940] S.C.R. 218, [1940] 1 D.L.R. 625 [hereinafter Massie (1940) cited to S.C.R.], var’g [1938]
Ex. C.R. 103, [1938] 2 D.L.R. 31 [hereinafter Massie (1938) cited to Ex. C.R.]. A defence based on
misuse has also been raised in several Canadian patent cases (see e.g. Philco Products Ltd. v. Thermi-
onics Ltd., [1939] Ex. C.R. 147, [1939] 3 D.L.R. 133, aff’d [1940] S.C.R. 501; Thennionics Ltd. v.
Philco Products Ltd., [1941] Ex. C.R. 209, 1 Fox Pat. C. 166, var’d [1943] S.C.R. 396, 3 Fox Pat. C.
92; Eli Lilly & Co. v. Marzone Chemicals Ltd. (1976), 29 C.PR. (2d) 253 (F.C.T.D.), aff’d [1977] 2
F.C. 104,29 C.P.R. (2d) 255 (C.A.)).

‘”R.S.C. 1927, c. 26.
‘”R.S.C. 1927, c. 36, s. 498.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

Even if the wrongs imputed against the plaintiffs were established in fact, I do
not think that would deprive them of their right to protect their copyright; their
copyrights would not perish because they had offended against another stat-
ute.'”

On appeal, the Supreme Court of Canada upheld the trial judge’s ruling, but with
respect to the statement as to misuse, the Court added:

[I]f the plaintiffs in an action for the infringement of copyright are obliged, for
the purpose of establishing the existence of, and their title to, the copyright to
rely upon an agreement and that agreement constitutes a criminal conspiracy,
and their title rests upon such agreement and upon acts which are criminal acts
by reason of their connection with such an agreement, then it would be diffi-
cult, on general principles to understand how such an action could succeed.’36

in Bell Canada

The Supreme Court’s ruling in Massie (1940) suggests that some form of
copyright misuse defence may be available in Canada. This can be reconciled with the
Federal Court of Appeal’s decision
v. hItra Canada
7 which seemed to reject a defence of misuse by recognizing
Telecommunications Ltd. 1
that the former case dealt with a criminal conspiracy while the latter dealt solely with
anti-competitive behaviour under the Combines Investigation Act.’
It has often been
thought that the misuse defence, though grounded in equity, has its origins in principles
which encourage competition. It is not clear from these cases, however, whether a mere
finding of anti-competitive behaviour will suffice to invoke the misuse defence, or
whether some other violation, such as a criminal act, is required. The current
Competition Act, which has replaced the Combines Investigation Act as the guardian of
competitive behaviour, has its own remedies for intellectual property misuse.’ ”

Pre-emption of Conflicting Laws

The United States Court of Appeals’ decision in Lasercomb, while clearly ac-
knowledging the existence of a copyright misuse defence, did not deal with reverse
engineering per se. The only American case to deal with the conflict between
copyright and contract in the context of reverse engineering was Vault Corp. v.
Quaid Software Ltd.’40 In that case, Heebe J. held that contract terms which widened
the scope of rights granted by the Copyright Act and by the legislation that sanc-
tioned the use of these terms, were pre-empted by the federal copyright legisla-

’35 Massie (1938), supra note 132 at 119.
‘6 Massie (1940), supra note 132 at 244.
’37 (1982), 70 C.PR. (2d) 252 (F.C.A.), rev’g (1982), 62 C.PR. (2d) 21 (F.C.T.D.).
33 Pratte J., facing a claim that anti-competitive behaviour by the plaintiff, in contravention of the
Combines Investigation Act, R.S.C. 1970, c. C-23, serves as a defence in an action for copyright in-
fringement, stated, ibid at 253: “We entertain serious doubts that they constitute a valid defence to the
action.”

‘9Competition Act, R.S.C. 1985, c. C-34, s. 32. See Part 11.C.3, below.
‘ 655 F.Supp. .750 (E.D. La. 1987), aff’d 847 E2d 255 (5th Cir. 1988) [hereinafter Vault].

1995]

S. HANDA – REVERSE ENGINEERiNG

tion.’ 4′ The case concerned the validity of the Louisiana Software License Enforce-
inent Act,142 a state statute that allowed for the enforceability of shrink-wrap li-
cences. Heebe J. ruled that the state legislation widened the protections granted by
section 106 of the American Copyright Act, and allowed contractual terms to im-
pede the archival copy privilege conferred on authorized users by section 117 .43
The Court of Appeals (Fifth Circuit) upheld the District Court’s decision, stating
that the state legislation was pre-empted “because it touched upon federal copyright
in a manner that set federal policy at naught.”‘ 44

In a Canadian context, one would argue that any provincial law that expressly
permits restrictions on rights and privileges guaranteed by the Copyright Act is
ultra vires because copyright is an exclusively federal power by virtue of subsection
91(23) of the Constitution Act, 1867.4
1 This, however, does not guarantee that
provinces will not create legislation in their own areas of competence which impacts
indirectly on copyrighted works. The law of contract, as well as the power to create
and administer property rights, falls within the property and civil rights powers given
to the provinces.4
4 Accordingly, “the publication, distribution and sale of … [many]
forms of literature may be regulated by the province within which the publication,
distribution or sale occurs. These are matters within property and civil rights in the
province.”‘ 47 As a result, much of the licensing of computer programs may be said to
fall generally within the competence of the provinces although the intellectual property
aspect remains within federal competence. The potential for conflict, as evidenced by
the litigation arising from the Louisiana licensing statute, also exists in Canada.

Where such a conflict exists in a federal system, the doctrine of federal para-
mountcy is used to resolve the dispute. Under this principle, “where there are in-
consistent (or conflicting) federal and provincial laws, it is the federal law which pre-
vails. … The doctrine of paramountcy applies where there is a federal law and a pro-
vincial law which are (1) each valid, and (2) inconsistent.””4 Therefore provincial
legislation and common law principles impacting upon copyright may not enhance

“‘ Rice, supra note 26 at 612-13.
” 27B La. Rev. Stat. Ann. 51: 1962ff (West 1987).
“‘ The relevant portion of 117, that could potentially be upset by the state legislation, reads:
Notwithstanding the provisions of section 106, it is not an infringement for the owner
of a copy of a computer program to make or authorize the making of another copy or
adaptation of that computer program provided: …

(2) that such new copy or adaptation is for archival purposes only and that all ar-
chival copies are destroyed in the event that continued possession of the com-
puter program should cease to be rightful.

‘Rice, supra note 26 at 612-13.
“‘(U.K.), 30 & 31 Vict., c. 3. Any residual powers, not specifically enumerated by the Constitution
Act, 1867 remain with the federal government unlike the case in the United States where they are left
with the states (P.W. Hogg, Constitutional Law of Canada, 3d ed. (Toronto: Carswell, 1992) at 108).

“, Constitution Act, 1867, ibid., s. 92(13).
“7 Hogg, supra note 145 at 601.
“‘ Ibid. at 418-19.

MCGILL LAW JOURNAL/REVUE DE DRO1TDE McGILL

[Vol. 40

or restrict rights and privileges granted under federal copyright law. This argument
exists independently of any argument made by virtue of the wording of section 63
of the Copyright Act, which expressly prohibits the granting of any rights similar to
copyright other than under the Act.

Statutory Paramountcy

Arguably, the linchpin that most effectively secures the pre-emption of licens-
ing provisions that conflict with federal legislation is section 301 of the American
Copyright Act. It is interesting to note that this section was not directly used in the
Vault decision. Section 301 states that the Act is the sole grantor of copyright; any
similar rights granted by the common law or by state legislation are pre-empted by
the Act.” 9 By virtue of paragraph 301(b),50 trade secret provisions continue to ap-
ply, but only insofar as they are consistent with the provisions of the Act.’ Accord-
ing to one commentator,

[t]he far-reaching public policy Section 301 implements clearly requires pre-
emption of contract-based protection of expression as expression where the ef-
fect is to secure rights in that expression which are greater than, equal to, or
supplemental of those which Section 106 secures. … The inescapable conclu-
sion is that contractual reverse engineering prohibitions cannot survive a Sec-
tion 301(a) challenge.’ 2

While section 301 of the American Copyright Act is markedly similar to section
63 of the Canadian Copyright Act, its construction is slightly different with respect
to the pre-emption of trade secret laws. The American Act bases itself on the
proposition that once copyright exists, nothing may exist that is inconsistent with
it.'” Section 63 of the Canadian Copyright Act, however, states that nothing in that
Act “shall be construed as abrogating any right or jurisdiction to restrain a breach of
trust or confidence.” It might therefore be argued that trade secrets are paramount

“9 Where other laws compete with copyright by creating similar rights or restricting copyright, there
are provisions in the Canadian, American and British copyright legislation which expressly state that
no other law may create a copyright or similar rights (Canada: Copyright Act, supra note 1, s. 63;
United States: Copyright Act, supra note 6, 301; United Kingdom: Copyright, Designs and Patents
Act 1988, supra note 88, s. 171).

“0 The text of the paragraph reads:

Nothing in this title [101 et seq.] annuls or limits any rights or remedies under the
common law or statutes of any State with respect to –

(1) subject matter that does not come within the subject matter of copyright as
specified by sections 102 and 103, including works of authorship not fixed in
any tangible medium of expression; or…

(3) activities violating legal or equitable rights that are not equivalent to any of the
exclusive rights within the general scope of copyright as specified by section
106.

15 Rice, supra note 26 at 605-606.
… Ibid at 614-16.
,51 See supra note 149.

1995]

S. HANDA – REVERSE ENGINEERING

under Canadian copyright law even where a conflict exists. If we apply this line of
reasoning to reverse engineering, the determination that a protectable trade secret
exists could lead a Canadian court to ignore the Copyright Act and its possible
sanctioning of such an exercise.5 4 Conversely, under the American Act, if the exer-
cise of a trade secret claim were to conflict with the rights and privileges granted
under the Act, the claim would fail.

While reverse engineering may be legally prohibitable under trade secret law,
the case for claiming copyright paramountcy is much stronger where no trade secret
exists. Unlike trade secret terms, simple contractual terms are not expressly ex-
empted from copyright principles by section 63 of the Copyright Act. Furthermore,
the application of section 63 is not restricted to other statutory instruments, and is
presumably wide enough to encompass contractual rights as well. Accordingly, a
public interest argument that stresses both paramountcy and copyright misuse is
likely to prevail where non-trade-secret contractual terms attempt to bar rights that
are otherwise allowed under the Copyright Act.

Competition Law

Even without the copyright misuse doctrine, or a statutory pre-emption of legis-
lative or contractual terms in favour of copyright, it is arguable that an attempt to
restrict reverse engineering through contractual terms may run afoul of Canadian
competition law. Grants of intellectual property protections such as patents and
copyrights are exempted from competition laws because of their statutory basis . 55
Any misuse of these rights that results in the lessening of market competition may,
however, be prohibited by the courts. Section 32 of the Competition Act, deals with
the misuse of intellectual property rights, stating:

(1) In any case where use has been made of the exclusive rights and privileges

conferred by … a copyright … so as to
(a) limit unduly the facilities for … producing, manufacturing, supplying …
or dealing in any article or commodity that may be a subject of trade or
commerce,

(b) restrain or injure, unduly, trade or commerce in relation to any such ar-

ticle or commodity,

(d) prevent or lessen, unduly, competition in the … sale … or supply of any

such article or commodity … , or

such article or commodity …

(2) The Federal Court, on an information exhibited by the Attorney General of

Canada, may, … make one or more of the following orders,
(a) declaring void, in whole or in part, any agreement, arrangement or li-

cence relating to that use;

Subject to the arguments previously discussed (see text accompanying notes 120-21, above).

t Subsection 79(5) of the Competition Act, subject to section 32 of that Act, expressly exempts in-
tellectual property rights obtained under the Copyright Act from being deemed “anti-competitive” and
thereby subject to a prohibitive order by the Competition Tribunal.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

(b) restraining any person from carrying out or exercising any or all of the

terms or provisions of the agreement, arrangement or licence;

(c) directing the grant of licences under any such … copyright … to such
persons and on such terms and conditions as the court may deem
proper… ; …

(e) directing that such other acts be done or omitted as the Court may deem

necessary to prevent any such use.

This section gives the Federal Court broad powers to restrict the misuse of any of
the intellectual property rights set out therein. If reverse engineering is indeed
permitted as an exception under the copyright regime, then any licence term that
would restrict reverse engineering would arguably be a use of the privileges of
copyright to “unduly limit or lessen” the production of other computer programs.
To date, there have yet to be any copyright disputes under the Competition Act, so
the breadth of section 32 remains unclear.

In summary, although it is a common practice in the computer software indus-
try to prohibit reverse engineering through licensing agreements, these prohibitions
cannot be supported by trade secret laws where there is no fiduciary or similar rela-
tionship between the parties. If copyright law is found to be broad enough to allow
reverse engineering, this type of licensing agreement provision will become unen-
forceable for a host of reasons, including copyright misuse, federal paramountcy,
statutory paramountcy and competition laws. If a valid trade secret is found to ex-
ist, the argument to prohibit reverse engineering remains strong, threatened only by
the potential paramountcy of competition law principles.

Patents

Patent law is another intellectual property regime that is increasingly important
to computer program protection. Patents are monopoly rights granted by the federal
government to inventors. 5 Obtaining a patent does not entitle one to specific sums
of money or to any other positive act, but rather creates a right to exclude others
from making, using or selling the invention for the duration of the patent. Patents
are distinct from other forms of intellectual property protection because they are
limited to functional products or processes that create a tangible product. Patents
are granted not for the underlying ideas of these products or processes, but only for
the physical manifestations of the ideas. Traditionally, it was thought that patents
were not applicable to computer programs; however, in recent years the Canadian
and American Patent Offices and the courts have allowed software patents for cer-
tain products.

To obtain a patent, the invention must possess novelty, utility and some
measure of inventive step (also known as “non-obviousness”). 57 Once a patent is

… Patents are governed in Canada by the Patent Act, R.S.C. 1985, c. P-4.
‘5’ The inventive step does not require that the invention be a revolutionary development: rather, it
may improve upon already existing technology. The rule of thumb for this test is that the invention

1995]

S. HANDA – REVERSE ENGINEERiNG

granted, its object is protected in Canada for twenty years from the date of filing the
patent application, after which the invention falls into the public domain.

Although it is widely believed that software patents are prima facie prohibited
under patent legislation, both the Canadian and American Patent Offices grant patents
for computer programs the operation of which results in a real world manifestation.
That is, the computer program must be characterizable as “something more than a mere
algorithm … [and can neither be] merely directed to making calculations nor to the
presentation of an algorithm and its solution.”’58 This view was articulated by the
Federal Court of Appeal in Schlumberger Ltd. v. Canada (Patent Commissioner),’59
which now stands as the authoritative Canadian decision in the field of software
patents. In that case, the Court held that the involvement of a computer could not
render patentable that which was unpatentable, and since mathematical formulae are
not patentable under the Patent Act,’6 they could not become patentable merely
because they took the form of a computer program.”” The Court did not, however, state
that a computer program was, as a result of its form, unpatentable. Presumably, if a
computer program embodied otherwise patentable subject matter, the program would
be patentable. The Canadian Patent Office, rejecting its earlier blanket ban on
computer program patents, seized upon this interpretation of the Schlumberger
decision.

The debate with respect to reverse engineering in the realm of patents is virtu-
ally non-existent. Any use of a patented invention, within the scope of the patent
claims filed, will require the payment of a royalty fee to the patent holder. If a pat-
ented invention is reverse engineered, any use of the material discovered through
the reverse engineering process will still potentially be covered by the claims in the
patent document. Accordingly, royalties would still be payable to the patent
holder if the use were to fall within the patent claims.

Many commentators continue to believe that the functionality of computer pro-
grams makes patent, and not copyright, law best equipped to provide intellectual
property protection.”‘ Providing copyright protection to computer programs over-

must elicit some reaction of marvel or amazement (i.e. “why didn’t I think of that?”) on the part of
others in the industry. If the patent is issued for an improvement to an existing technology, any pro-
duction of the new invention must be authorized by the patent holders of the existing technology. This
permission is usually in the form of licensing agreements.

’58R. Trudeau, “Software Patents” (1992) 9 C.I.P.R. 234 at 238-39.

(1981), [1982] 1 F.C. 845, 56 C.P.R. (2d) 204 (C.A.) [hereinafter Schlumberger].

“‘ No patent may be granted for mere scientific principles or abstract theorems (Patent Act, supra

note 156, s. 27(3)).

“‘ Sookman, supra note 48 at 6-21.
161 The patent granting process requires full disclosure concerning the manufacture of the invention.
Reverse engineering is therefore unnecessary since anyone may obtain a copy of this disclosure and
the specifications contained therein.

,63 See Z. Hau, “Securing Patent Protection for Computer Program-Related Inventions” (March
1993) Patent World 34; S.A. Becker, “Drafting Patent Applications On Computer-Implemented In-
ventions” (1991) 4 Harv. J. L. & Tech. 237; Trudeau, supra note 158; Sookman, supra note 48 at 6-2.

MCGiLL LAW JOURNAL/REVUE DE DROITDE MCGLL

[Vol. 40

protects them because it improperly treats them as literary works and not solely as
functional works of technology.’ 6 The limitations of copyright to adequately reflect
the policy goals associated with technological works have become clearer with the
advent of computer programs, and have forced governments to disregard copyright
when implementing protections for other technological works, such as semiconduc-
tor chips.

Semiconductor Chip Protection

In the past decade, both Canada and the United States have enacted sui generis
legislation to protect semiconductor chips.’65 An examination of these statutes informs
our discussion of copyright and reverse engineering as it provides some insight into
how intellectual property legislation can be customized to address the needs of a par-
ticular technology. The recently enacted semiconductor legislation may be contrasted
with more general catch-all intellectual property legislation, such as the copyright and
patent statutes. Moreover, semiconductor legislation is also relevant because it contains
specific reverse engineering provisions.’66

The protection afforded by both the American Semiconductor Chip Protection
Act t 7 and the Canadian Integrated Circuit Topography Act’6 extends to the “mask
work fixed in a semiconductor chip product”, or “topography”, as it is referred to under
the Canadian Act. Mask works, or integrated circuit topographies, are generally defined
as a series of images that represent in their totality a three-dimensional rendering of the
chip product 69 Each image represents a layer of a chip that is conceptually peeled

‘ Ignatin, supra note 17 at 2021.
16 Semiconductor or microchips which contain computer programs, such as ROM chips, may re-
ceive two-tiered protection as the computer programs are further protected by copyright legislation
(see Apple, supra note 8).

’66 The over-protection of software, as compared to semiconductor chips, is particularly
relevant because of the great similarities between the two technologies, both in the way
they are developed and the way they operate. … [In fact,] the dividing line between
hardware (such as semiconductor chips) and software is extremely fuzzy (Ignatin, su-
pra note 17 at 2020 and at note 83).

267 17 U.S.C. 901ff (1994) [hereinafter S.C.P.A.].
‘6 S.C. 1990, c. 37 [hereinafter I.C.T.A.].
“s Section 2 of the I.C.T.A. defines topography as:

the design, however expressed, of the disposition of,

(a) the interconnections, if any, and the elements for the making of an integrated

circuit product, or

(b) the elements, if any, and the interconnections for the making of a customization
layer or layers to be added to an integrated circuit product in an intermediate
form.

Section 901(a)(2) of the S.C.P.A. defines mask work as:

a series of related images, however fixed or encoded

(A) having or representing the predetermined, three-dimensional pattern of metal-
lic, insulating, or semiconductor material present or removed from the layers
of a semiconductor chip product; and

1995]

S. HANDA – REVERSE ENGiNEERING

away from the chip so as to be represented in two dimensions. These images are pro-
tected by the respective Acts.

Both Acts protect the topographies of original works for ten years.70 In keeping
with the American and Canadian Copyright Acts, the protection afforded by the
S.C.P.A. and I.C.T.A. extends only to the expression of the topography and not to any
underlying “idea, procedure, process, system [or] method of operation” embodied in
the work. 7′ The S.C.EA. requires that the mask work be fixed in order to receive pro-
tection,’ but the I.C.T.A. has no such requirement. Under both Acts, the required stan-
dard of originality falls somewhere between the standard of originality under the
Copyright Act’ 73 and the standard of novelty under the Patent Act. 74 Originality for the
purposes of semiconductor chip protection under the I.C.T.A. requires that in addition
to the work not being a “mere reproduction of another topography,” it must also be the
“result of an intellectual effort and … not … commonplace among creators of topogra-
phies or manufacturers of integrated circuit products.”‘ 75

The reverse engineering provisions, appearing in similar fashion in both Acts, dis-
tinguish these Acts from the copyright provisions that apply to computer programs.
Paragraph 6(2)(a) of the I.C.T.A. allows a person, to undertake any act “in relation to
that registered topography for the sole purpose of analysis or evaluation or of re-
search,” that does not “commercially exploit the topography or any substantial part
thereof.” Thus, any act that does not copy the whole topography or a substantial part
thereof and that is commercially exploitive is allowed under the I.C.T.A. This level of
similarity should be much more loosely construed than are similar enquiries as to
copying under the Copyright Act. According to the proposal that resulted in the imple-
mentation of the I.C.T.A.,

[t]he reverse-engineered chip and the protected chip might legitimately be
identical in electronic function and external fit. But, reverse engineering would
produce a chip with a three-dimensional layout neither identical nor virtually
identical to the topography embodied in the protected chip. The proposed re-
verse-engineering measure would legitimate the creation of a substantially
similar topography for a fully compatible chip, potentially offering for exam-
ple:

” an improved signal/noise ratio;
” fewer fabrication steps;
* greater thermal stability;
* decreased die size;

(B) in which series the relation of the images to one another is that each image has

the pattern of the surface of one form of the semiconductor chip product.

70 S.C.P.A., supra note 167, 904 (a) – (b); LC.T.A., supra note 168, paras. 5 (a) – (b).
‘7’ S.C.P.A., ibid., 902(c); I.C.T.A., ibid., s. 3(3).
,’ S.C.PA., ibid., 902(a)(1).
‘.. See Part HA. L.c, above, for a discussion of this issue.
7, See text accompanying note 157, above.
,75 I.C.T.A., supra note 168, s. 4(2). Similar protection is afforded by S..C.A., supra note 167,

902(b).

McGiLL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

*
*

faster performance; and
lower manufacturing cost.’76

Infringement of the rule would occur only where the reverse engineering resulted in
the “production and marketing: of an exact copy of a protected topography; or of a
chip embodying a topography virtually identical to the protected topography or to a
substantial portion thereof.”” 77 A sui generis regime was favoured over simply extend-
ing copyright protection to semiconductor chips because the underlying policy recog-
nized that this genre of “legislation must … suppress chip piracy without creating un-
necessary obstacles to a free market in semiconductor chips and to the spread of chip
technology.”‘ 78 The use of

pure copyright principles to prevent the unauthorized copying of a chip topog-
raphy would not meet the needs of the semiconductor industry … who wish
[sic] to make an unauthorized copy of all of their competitor’s topography for
analysis; and to manufacture a substantially similar chip derived from their
competitor’s topography.”

The test for valid reverse engineering under both the I.C.T.A. and the S.C.P.A. requires
a “two-step inquiry”:

First, if it is determined that a competitor has substantially studied and analyzed
a protected mask work to produce its own chip, i.e. valid reverse engineering,
that chip does not infringe even if it is substantially similar to the mask
owner’s. However, if the competitor’s design incorporates identical parts of the
protected design, infringement may yet be found.IW

As with the S.C.P.A., the drafters of the I.C.T.A. expected that the work involved
in reverse engineering a chip would be substantial and would justify reverse engineer-
ing as an alternative to continually “re-inventing the wheel”.’ 8′ Both the Canadian and
American Acts rely on the concept of a “paper trail” as proof that analysis has been
performed on a chip, which is an activity allowed as a fair use under subsection 6(2) of
the I.C.T.A. ‘ Proof that “sweat of the brow” has been expended may involve the use

“6 A.Z. Hertz, Semiconductor Chip Protection in Canada: Proposals for Legislation (Ottawa: De-
partment of Consumer and Corporate Affairs, 1987) at 46. This report specifically recommended,
ibid at 49: “Canada’s chip-protection law should contain a reverse-engineering exception allowing
the unauthorized copying of a protected topography in a process of analysis and redesign leading to
the creation of a substantially similar chip topography.”

‘ Ibid at 46.
‘7 lbidL at45.
‘ Ibid at 43-44.
‘” S.P. Kasch, “The Semiconductor Chip Protection Act: Past, Present and Future” (1992) 7 High

Tech. L.J. 72 at 77.

“‘ It is estimated that reverse engineering a semiconductor chip “requires thousands of person-hours

and about a quarter of the money needed to create the original chip” (Hertz, supra note 175 at 47).

“8

The end product of the reverse engineering process is not an infringement, and itself
qualifies for protection under the [S.C.P.A.], if it is an original mask work, as con-
trasted with a substantial copy. If the resulting semiconductor chip product is not sub-

1995]

S. HANDA – REVERSE ENGINEERING

of “ordinary business documents and technical materials [such as] invoices, employ-
ment and payroll records, logic and circuit diagrams, trial layouts and computer simu-
lations of the chip.” . 3

D. Conclusion

Without further legislative guidance as to the extent of the fair dealing exception to
copyright infringement, copyright law prohibits the reverse engineering of computer
programs. Aside from copyright, no other intellectual property regime disallows the
reverse engineering of computer programs in a blanket fashion. In fact, semiconductor
chip and patent legislation either expressly allow reverse engineering or provide ade-
quate disclosure of the ideas and processes underlying the protected work. Trade secret
laws, although prohibiting reverse engineering by those who have a “special relation-
ship” with the work’s owner, allow persons not in such a relationship to reverse engi-
neer a computer program. If it is determined that the reverse engineering of computer
programs falls within the exception of fair dealing, attempts to enhance protection,
other than through trade secret laws, may be considered anti-competitive and may re-
sult in the suspension of copyright enforcement privileges.

III. Existing Approaches to the Reverse Engineering Problem

Although the legal problems associated with reverse engineering have not yet
appeared in a Canadian context, they have received attention in the United States
and in the European Union (E.U.). The American experience indicates that court
challenges based on reverse engineering claims are imminent in Canada and that a
legislative response should precede such court challenges, thus avoiding the uncer-
tainty of a judicial decision based on statutory language that was not enacted in
contemplation of computer technology. The European Community’s legislative re-
sponse to reverse engineering serves as a good example of a potentially workable
regulatory scheme.

stantially identical to the original, and its design involved significant toil and invest-
ment so that it is not a mere plagiarism, it does not infringe the original chip, even if the
layout of the two chips is, in substantial part, similar. … [T]he courts are not likely, as a
practical matter, to find it unduly difficult to draw the line between reverse engineering
and infringement, because the additional work required to come within the privilege
established by s. 906(a) will ordinarily leave a “paper trail” (United States House and
Senate Explanatory Memorandum, 130 Cong. Rec. 28,960 (daily ed. 3 October 1984)
(“Explanatory Memorandum – Mathias-Leahy Amendment to s. 1201”).

t Hertz, supra note 176 at 47. The idea is that a bonafide reverse engineer will leave a trail as
compared with a pirate who “is not able to produce a long paper trail because he has not done genuine
design-development work.”

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

A. United States

The issue of reverse engineering computer programs under the American Copy-
right Act was recently considered in two United States Court of Appeals cases which
mark the first time any court has dealt with the issue under copyright laws.’ s4 The
judgments in these cases were released a mere month apart, and together present a
consistent, albeit surprising, policy-oriented approach to reverse engineering. Prior to
the release of these cases the United States District Court examined the issue with re-
spect to the reverse engineering of data tables.” Although the more recent appellate
cases present more sophisticated analyses of the copyright issues involved, the lower
court case applies a rudimentary approach to reverse engineering that, in the opinion of
this author, arrives at an appropriate solution.’86 The reasoning in this case reflects an
elegant simplicity which speaks well to the issue of reverse engineering.

E.E Johnson Co. v. Uniden Corp. of America

In 1985, the United States District Court addressed a claim involving the re-
verse engineering of data tables contained in a microchip, and its permissibility un-
der the Copyright Act. In Uniden, the plaintiff moved for a preliminary injunction
restraining the defendant from publishing, selling, marketing or in any way dis-
tributing software that would allow users to access a range of frequencies in the de-
fendant’s two-way radio product. The radios in question were controlled by soft-
ware provided to users by the plaintiff along with the radios. The defendants re-
verse engineered the plaintiff’s computer program, uncovering data tables that
listed activation codes for the radios, and then produced their own program using
identical data tables. The defendants’ tables, however, also contained unnecessary
elements, such as errors and duplications made in the plaintiff’s table. The Court
granted an injunction based on the substantial likelihood of success of the plain-
tiff’s copyright claim on its merits. It was decided that alternative means to copying
the plaintiff’s software were available, by which the defendant could have placed
the required radio codes into their own program. It was further held that the defen-
dants had acted unfairly and were guilty of copying copyrighted expression.

The Court did not discuss whether the reverse engineering activities of the de-
fendant were infringements of copyright. The Court chose rather to look at the final
product and examine it for traces of infringing expression. It remains unclear
whether the Court’s silence indicates conscious implicit acceptance of reverse engi-
neering as a permissible act under copyright legislation. It is more likely that, be-
cause this was an early case involving computer programs under American copy-
right law, the Uniden Court simply missed the issue. One could, however, discern
acceptance of reverse engineering. This interpretation is supported by several pas-

“s Sega, supra note 4; Atari, supra note 4.
“‘ Uniden, supra note 4.
18 See Part IV.C, below, for recommendations designed to deal with the problem of reverse engi-

neering.

19951

S. HANDA – REVERSE ENGINEERING

sages in the judgment stating that there would have been no infringement had the
defendants only used functional parts of the data tables which could not be pro-
duced through some alternate means. 87 The Court’s decision presupposes that the
defendants .were permitted to uncover these codes, which could only be done
through reverse engineering the chip containing the data tables. As no statement
was made to the contrary, presumably the intermediate copies of the table produced
at this stage would not constitute copyright infringements.

Atari Games Corp. v. Nintendo of America Inc.

The first of the two decisions to be released concerning the reverse engineering of
computer programs was that of the Federal Circuit Court of Appeals in Atari Games
Corp. v. Nintendo of America Inc. In that case, Nintendo sued Atari for copyright
infringement of its 10NES lock-out program.’
lONES is a program embedded in
chips contained in the popular Nintendo NES game console, which intercepts the flow
of information from game cartridges.8 9 Nintendo allowed other software developers to
become licensees and write software for the NES. The licence agreements were highly
restrictive, but allowed the licensees to produce software which would be bundled with
Nintendo’s secret unlocking message.’ 90

In 1986, Atari had attempted to crack the 10NES program by both monitoring the
data flow between the console and the cartridges, and by chemically peeling microchip
layers from the 10NES chips and microscopically examining the data paths.’ 9′ Unable
to crack the 10NES program, Atari decided to become a Nintendo licensee in 1987,
and was barred under the terms of that agreement from gaining access to the 10NES
code and from uncovering
its operation. In 1988, a lawyer representing Atari
approached the American Copyright Office and applied for a copy of the 10NES
program, which had been filed for copyright registration purposes. The lawyer stated
that the code was required for litigation that had been commenced against Atari. The
code was provided under the condition that “the requested copy [would] be used only

“8 Uniden, supra note 4 at 1504.
‘ Nintendo had originally sued Atari for unfair competition, patent infringement, copyright in-
fringement and trade secret violations. Atari countersued for unfair competition, Sherman Act viola-
tions, and patent infringement. The cases were consolidated and Nintendo obtained a preliminary in-
junction from the District Court for the Northern District of California based on Atari’s unauthorized
use of Nintendo’s copyrighted expression. Atari appealed this injunction to the United States Court of
Appeals, Federal Circuit.

‘” LONES waits for the cartridge to send a coded message to the console. The coded message is
contained in a microchip and is referred to as the IONES “slave” chip. The portion of the 1ONES pro-
gram contained in the console is also in microchip form and is referred to as the 10NES master chip.
Once the coded message is detected, 10NES allows the console to operate the cartridge. If the appro-
priate message is not detected, 1ONES will not “unlock” the console, and will not allow the cartridge
to operate.

‘” In fact, under the licensing agreements, software developers would provide Nintendo with the
computer programs. Nintendo would then package the games into cartridge form with the 10NES
slave chip and resell them to the software developer.

‘.’ Atari, supra note 4 at 836.

MCGILL LAW JOURNAL/REVUE DE DROIT DE MCGILL

[Vol. 40
in connection with the specified litigation.”’92 No litigation actually existed between the
parties at that time.

Atari used the code provided by the Copyright Office to develop its own unlocking
program entitled “Rabbit”. Rabbit was a program that, although written in a different
computer language, replicated the data stream provided by the 10NES slave chips.
Nintendo sued Atari, and was granted a preliminary
injunction for copyright
infringement.

Atari’s defence to the copyright infringement claim was primarily based on the fair
use exception.’ 93 Section 107 of the American Copyright Act sets out four factors to
consider in a fair use claim:

In determining whether the use made of a work in any particular case is a fair
use the factors to be considered shall include –
(1) the purpose and character of the use, including whether such use is of a

commercial nature or is for nonprofit educational purposes;

(2) the nature of the copyrighted work;
(3) the amount and substantiality of the portion used in relation to the copy-

righted work as a whole; and

(4) the effect of the use upon the potential market for or value of the copy-

righted work.

The Atari Court did not perform a detailed analysis of these factors in its decision.
Instead, the Court looked to copyright to balance the monetary self-interest of
individual authors with society’s interest in promoting the free flow of ideas. With
respect to reverse engineering, the Court found irrelevant the fact that Nintendo’s code
was in object code form and stored on a microchip. It held that it is a fair use to reverse
engineer a computer program in order to gain a better understanding of the program’s
underlying ideas:

An author cannot acquire patent-like protection by putting an idea, process, or
method of operation in an unintelligible format and asserting copyright in-
fringement against those who try to understand that idea, process, or method of
operation…. The Copyright Act permits an individual in rightful possession of
a copy of a work to undertake necessary efforts to understand the work’s ideas,
processes, and methods of operation.

This permission appears in the fair use exception to copyright exclusivity.”‘

,9 Ibid
‘9 Fair use, like its Canadian fair dealing counterpart, is an exception to copyright infringement, set

out in the American Copyright Act at 107:

mhe fair use of a copyrighted work, including such use by reproduction in copies or
phonorecords or by any other means specified by [ 106, 106A], for purposes such as
criticism, comment, news reporting, teaching (including multiple copies for classroom
use), scholarship, or research, is not an infringement of copyright.

‘ Atari, supra note 4 at 842.

1995]

S. HANDA – REVERSE ENGINEERING

Fair use protects the intermediate stages of reverse engineering, along with the
resulting related copies from being deemed infringing. According to the Court, fair use
does not extend to reverse engineering undertaken for the purpose of profiting by
replicating protected expression, but rather only applies where reverse engineering is
necessary to discern unprotectable ideas.

In the case of Nintendo’s 10NES program, the Court determined that the
underlying unprotectable ideas were the codes required to unlock the NES console.
Anything in excess of that was presumed to consist of protectable expression.’9 5 Atari’s
copy of various parts of the 10NES program included errors and deletions made by
Nintendo in their program. This copying of unnecessary portions of the Nintendo table
resulted in a decision against Atari for copyright infringement since, as the Court
stated, “these unnecessary instructions strongly suggest that the Rabbit program is
substantially similar to the 10NES program … [which] Nintendo is likely to show …
contains protectable expression.”‘” Furthermore, Atari’s unfair appropriation of the
10NES source code by misrepresenting its intentions to the Copyright Office pre-
empted any use of the fair use defence which only applies to authorized copies of a
work.’ 97 Atari’s dishonest behaviour in obtaining the 10NES code also precluded Atari
from using a copyright misuse defence based on Nintendo’s extensive licensing
agreement.”‘ The Court failed to recognize the possibility that Atari copied elements
unnecessary to unlock the Nintendo console in order to ensure against “the possibility
that Nintendo could alter its console in the future to utilize currently unused portions of
the compatibility code.” 99

Atari lost its appeal to lift the injunction imposed by the District Court, and the
Court of Appeals also upheld the District Court’s substantive ruling concerning the fair
use exception.

Sega Enterprises Ltd. v. Accolade Inc.

Immediately following the Federal Circuit Court of Appeals’ decision in Atari,
the Ninth Circuit Court of Appeals released its decision in Sega. In that case, Acco-
lade had sought to reverse engineer cartridges containing video game programs
with the intention of uncovering how the software interacted with Sega’s Genesis
game consoles. Using the results of the reverse engineering, Accolade created a
manual for use by its programmers in developing games that would operate on the
Sega console. The information used by Accolade at this early stage was purely
functional and did not involve any literal copying of Sega program codes. During

195 Ibid. at 843-44.
‘” Ibid. at 845.
’97 Ibid. at 843; Harper & Row Publishers, Inc. v. Nation Enterprises, 471 U.S. 539 at 562-63, 85 L.

Ed. (2d) 588 (1985).

‘g’ Atari, ibid. at 846. See also text accompanying note 130, above.
‘9H.C. Moore, “Atari v. Nintendo: Super Mario Uses ‘Expressive’ Security Feature to ‘Lock’ Out
the Competition” (1992) 18 Rutgers Comp. & Tech. L.J. 919 at 933. This possibility manifested itself
in Sega.

MCGILL LAW JOURNAL/REVUE DE DROIT DE MCGILL

[Vol. 40

the period that Accolade was beginning to release its own compatible software,
Sega, in order to combat high degrees of piracy of its games, decided to license a
copy protection system for use in its newly developed Genesis III consoles. The
trademark security system (T.M.S.S.) would be placed in each Sega console and,
when a cartridge was plugged in, would search for the letters “SEGA” contained in
the program code of each cartridge. Upon finding this initialization code, the
Genesis II console would produce a display of “PRODUCED BY OR UNDER
LICENCE FROM SEGA ENTERPRISES LTD.”, and would unlock the console for
game play, much like Nintendo’s 1ONES system.

Upon finding that its game cartridges would not work with the Genesis III sys-
tem, Accolade examined the results of its reverse engineering efforts and found that
each licensed Sega cartridge contained a small unused portion of code in each
game’s “power-up” sequence, which was now being used by Sega in its T.M.S.S.
lockout system.2 In response, Accolade created a small header file (twenty to
twenty-five bytes of data) containing the SEGA initialization code, to be used in all
future game development. This information was also incorporated into its devel-
opment manual.

Sega brought an action against Accolade for copyright infringement, trademark
infringement and false designation of origin. Accolade countered with claims of
false designation of origin, unfair competition, false or misleading statements and
intentional interference with prospective economic advantage. Sega was granted a
preliminary injunction by the United States District Court based on Sega’s copy-
right and trademark infringement claims. Accolade appealed to the United States
Court of Appeals (Ninth Circuit), whereupon the injunction was lifted and the
matter was remanded to the District Court.

The decision of the Court of Appeals to lift the preliminary injunction and al-
low reverse engineering based on the fair use exception provides a detailed analysis
of the reverse engineering issue as it applies to computer programs. The Sega Court
enunciated a very useful step-by-step analysis of the four fair use factors, and dis-
cussed three related reverse engineering arguments: intermediate copying; the
idea/expression dichotomy as a possible justification for reverse engineering; and a
section 117 defence. Although Accolade ultimately succeeded in its claim that the
reverse engineering at issue constituted a fair use, the Court rejected Accolade’s
other three arguments.

With respect to intermediate copying, the Court held that the fact that the prod-

“‘The use of apparently useless code in an upgrade of the system highlights a difficulty inherent in
the idea/expression analysis performed by courts in order to determine what may fairly be reverse
engineered and used in a competing program. If computer code, and/or errors, serve no purpose, then,
as was the case in the Atari and Uniden decisions, they may not be legally copied. However, the
original program developers may subsequently use these pieces of computer code in a lockout system
which would render all other manufacturers’ products useless. This possibility was ignored by the
Court in the Atari case (see supra note 198 and accompanying text).

19951

S. HANDA – REVERSE ENGINEERING

uct is intermediate, as opposed to final in nature in no way exempts it from the
Copyright Act. Whether intermediate copying properly constitutes an infringement
of copyright must be examined, notwithstanding whether the final product itself
constitutes an infringement. 20′ According to the Court’s reasoning,

the computer file generated by the disassembly program, the printouts of the
disassembled code, and the computer files containing Accolade’s modifications
of the code that were generated during the reverse engineering process all …
[fall] squarely within the category of acts that are prohibited by the statute.?2

Concerning Accolade’s claim that the idea/expression dichotomy provides a
justification for reverse engineering, the Court ruled that the fact that ideas may be
contained within protected expression does not mean that the expression accompa-
nying those ideas may be ignored. Accolade had argued that the nature of computer
programs distinguishes them from other, more traditional copyrightable works, the
ideas of which may readily be perceived and understood by human beings. Conse-
quently, reverse engineering was perceived as simply a crutch to put computer pro-
grams on an equal footing with these other works.

The final argument disposed of by the Court in Sega’s favour concerned Acco-
lade’s claim that under section 117 of the Copyright Act the production of interme-
diate copies does not constitute a copyright infringement. The Court reviewed the
C.O.N.T.U. report responsible for the implementation of section 117, and ruled:

.203

Accolades use went far beyond that contemplated by CONTU and authorized
by section 117. Section 117 does not purport to protect a user who disassem-
bles object code, converts it from assembly into source code, and makes print-
outs and photocopies of the refined source code version.,

The Court did, however, accept Accolade’s fair use argument:

Because, in the case before us, disassembly is the only means of gaining access
to those unprotected aspects of the program, and because Accolade has a le-
gitimate interest in gaining such access (in order to determine how to make its
cartridges compatible with the Genesis console), we agree with Accolade.
Where there is good reason for studying or examining the unprotected aspects
of a copyrighted computer program, disassembly for purposes of such study or
examination constitutes a fair use.a

Of the four fair use factors mentioned in section 107 of the American Copy-
right Act, 26 the Court found in favour of Accolade on the first, second and fourth
factors, and in favour of Sega on the third factor.

“‘Sega, supra note 4 at 1518.
22 Ibid.
‘0’ See supra note 83 for the text of section 117.

Ibi at 1520.
Ibid.
For a lst of the factors, see text following note 193, above.

MCGILL LAW JOURNAL/REvuE DE DROITDE MCGILL

[Vol. 40

The first factor concerns the purpose and character of the use, including
whether such use is commercial in nature or for nonprofit educational purposes.
The Court ruled that “[t]he commercial nature of a use is a matter of degree, not an
absolute,” and in the case of Accolade, the direct purpose of reverse engineering
the Sega code was to ensure the compatibility of their cartridges with the Sega con-
sole.20
7 The commercial nature of cartridges was secondary and, at best, indirect. In
considering this factor, the Court also seemed to hint at a “sweat of the brow” ra-
tionale combined with some required measure of creativity.”‘ The Court held that
“there is no evidence in the record that Accolade sought to avoid performing its
own creative work … [as] it wrote its own procedures. … [T]hese facts indicate that
… its direct use of the copyrighted material … was simply to study the functional
requirements for Genesis compatibility.”201 In support, the Court also noted that the
public interest in promoting compatibility and competition reduced the strength of
Sega’s claim that Accolade stood to gain commercially.

With respect to the fourth factor listed in section 107, the effect of the use upon
the potential market for or value of the copyrighted work, the Court held that Acco-
lade’s development will not directly cause Sega to lose customers because the
games developed by Accolade are not copies of those developed by Sega.”0 Al-
though the Court recognized that Accolade’s entrance
into the market will
“undoubtedly affect” the sale of Genesis games, the public policy rationale underly-
ing copyright legislation is to promote creative expression and not to stifle compe-
tition in a specific market.

The final factor decided in Accolade’s favour dealt with the nature of the copy-
righted work. The basis of the Court’s analysis here was that computer programs
are essentially utilitarian in nature and contain many elements the expression of
which is often dictated by notions of efficiency, or external factors such as com-
patibility.21’ These elements are not protectable under copyright legislation. Accord-
ingly, if Accolade were unsuccessful in its fair use claim Sega would unrightfully

20 Sega, supra note 4 at 1522.

The application of “sweat of the brow” to copyright protection has recently gained much notori-
ety as a result of the United States Supreme Court’s decision in Feist Publications, Inc. v. Rural Tele-
phone Service Company, 499 U.S. 340, 113 L. Ed. (2d) 358 (1991). This case raised the issue of
whether or not copyright could exist in a telephone book. The Court held that mere “sweat of the
brow” was not sufficient in itself to attract copyright protection. Creative effort, including particular
selection or arrangement, applied to the data was also necessary for copyright to exist. Although the
context was decidedly different in Sega, the interplay between “sweat of the brow” and original-
ity/creativity is gradually being redefined with respect to utilitarian works such as databases and com-
puter programs

Sega, supra note 4 at 1522.

2,0 The Sega Court’s decision seems to indicate that in order to successfully argue this factor as a bar
to fair use, it must be shown that “the new work … supplant[s] the direct market for the particular
copied work” (D.L. Hayes, ‘The Legality of Disassembly of Computer Programs” (1993) 12 Comp.
LJ. I at 8).

211 Sega, supra note 4 at 1524.

1995]

S. HANDA – REVERSE ENGINEERING

obtain a monopoly over the functional ideas underlying its work. Sega argued that
there were methods other than disassembly by which its initialization code could be
uncovered. The Court disagreed and found that “disassembly of the object code in
Sega’s video game cartridges was necessary in order to understand the functional
requirements for Genesis compatibility. ‘212 In effect, the Court recognized that
while Accolade’s disassembly necessarily involved potentially infringing interme-
diate copying, to allow Sega to succeed on this claim would expand copyright be-
yond its intended scope and stifle competition. The Court therefore chose the lesser
of two evils and, based on public policy reasons, found that Accolade’s actions
constituted a fair use of Sega’s copyrighted material. The Court did not clarify
whether this balance would have tipped in Sega’s favour had methods other than
disassembly been available to Accolade.

The Court found in favour of Sega, however, with respect to the third factor
mentioned in section 107, the amount and substantiality of the portion used in rela-
tion to the copyrighted work as a whole. Accolade disassembled entire programs
written by Sega in order to uncover the initialization sequence. The Sega Court,
however, citing the decision in Sony, ruled that “the fact that an entire work was
copied does not … preclude a finding of fair use.”2″3 The Court here stated that the
fact that the entire work was copied was “of very little weight”.24 The Court’s rul-
ing on this point runs counter to that of the Canadian Exchequer Court in Zamacors,
which held that the fair dealing defence under the Canadian Copyright Act cannot
be invoked where an entire work is copied.2 5 Zamaco’s was decided in an era when
copyright did not contemplate the protection of computer programs the functional
nature of which may dictate a revision of existing copyright principles. This
“special functional nature” of computer programs was, however, clearly recognized
by the Sega Court in its analysis of the idea/expression dichotomy.

B. The European Union

The area of intellectual property, and more specifically copyright, is not ad-
dressed in the constitutive treaties of the European Union, and it has proved diffi-
cult to harmonize. According to one commentator, “[t]here will not be a unified
European copyright law in the near future. The national copyright laws of member
states will continue to apply.”2t6 The European Commission has, however, recog-

212 Ibid.
2″3 Ibid. at 1526. See also Sony, supra note 92 at 449-50; Hustler Magazine, Inc. v. Moral Majority,
hzc., 796 F.2d 1148 at 1155 (9th Cir. 1986): ‘TF]he copying of an entire work does not preclude fair
use per se.”

2,4 Sega, ibid at 1527.
2See supra note 94 and accompanying text.
,16 S. von Lewinski, “Copyright in the European Communities: The Proposed Harmonization
Measures” (1992) 18 Brooklyn J. Int’l Law 703. In 1990 the European Commission adopted a pro-
posal that would require all member states to accede to various international copyright and related
rights treaties including: the Berne Convention for the Protection of Literary and Artistic Works (as
amended), 30 June 1972, 828 U.N.T.S. 221; International Convention for the Protection of Perform-

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

nized that specific industries, including the software industry, cannot wait for a
general harmonized European copyright law. Due to their relative economic impor-
tance and to the disparity of treatment among various member-states, these indus-
tries require that common legal rules be applicable throughout member-states with-
out delay.

The E.E.C. Directive on Computer Programs

Accordingly, in order to harmonize existing protections and establish common
principles within the Community with respect to the intellectual property of com-
puter programs, in 1991 the European Council passed the Directive on the Legal
Protection of Computer Programs. The objectives of the Directive were two-fold:

[Tio prevent the unlawful copying of computer software, or “computer piracy”,
within the Community by ensuring an adequate level of protection for those
who create computer software; and

to promote the free circulation of computer software within the Community
and allow industry to take advantage of the single market by harmonising the
national laws of the Member States relating to the use and reproduction of
computer software.”7

The European Commission, charged with designing the Directive, opted for copy-
right as the basis for protection because of its interpretative flexibility, its ability to
balance monopoly rights with societal interests, its limited scope of protection and
most importantly because, to date, copyright had been the intellectual property
protection of choice of a number of Community member-states.”‘ The Commis-
sion’s choice of copyright was hardly revolutionary or controversial. However, the
decompilation exception to infringement was hotly debated prior to its inclusion in
the Directive.219

Early in its conception, the Commission’s Proposal for the Directive (the 1989
Proposal) contained two sections which could be interpreted as allowing the reverse
engineering of protected computer programs. Article 1.3 of the 1989 Proposal
stated that only a computer program’s expression would be afforded legal protec-
tion, and not

the ideas, principles, logic, algorithms or programming languages underlying
the program. Where the specification of interfaces constitutes ideas and prin-
ciples which underlie the program, those ideas and principles are not copyright-
able subject matter.2

ers, Producers of Phonograms and Broadcasting Organizations, 26 October 1961, 496 U.N.T.S. 43.

27 Clifford Chance, The European Softvare Directive (London: Clifford Chance, 1991) at 5.
8 Ibid. at 5-6.
2,9 The Directive employs the term “decompilation” in its broad sense as equivalent to the term
“reverse engineering”, as opposed to its literal meaning. “Decompilation” as used in the Directive
therefore includes disassembly as well (see supra notes 18-19 and accompanying text).

EC, Proposal for a Council Directive on the Legal Protection of Computer Programs, O.

In-

1995]

S. HANDA – REVERSE ENGINEERING

Article 5.1 of the Proposal allowed the reproduction and adaptation of a computer
program, potentially including the reverse engineering thereof, where such manipu-
lations were necessary for the use of the computer program. This adaptation right
would exist notwithstanding the copyright-holder’s refusal. Although these sections
would implicitly allow reverse engineering, the 1989 Proposal did not contain ex-
press wording to that effect. The European Parliament accepted the Commission’s
1989 Proposal, subject to certain amendments. The Commission responded with its
amended proposal in 1990, which contained express wording that would allow re-
verse engineering in certain circumstances. Article 5bis of the amended proposal
was eventually adopted almost verbatim into the Directive (as Article 6), and is
unique in that it expressly allows reverse engineering in the context of copyright
protection of computer programs.22

Subsection 6(1)

states

that decompilation may be used where

is
“indispensable to obtain the information necessary to achieve the interoperability of
an independently created computer program with other programs.” It goes on to list
three conditions which limit the scope of the decompilation, all of which must be
met in order for the copyright exemption to apply. First, individuals decompiling
software must be licensees or titularies of some other right to use the software; sec-
ond, the information being sought through decompilation must not be readily avail-
able; and third, the decompilation must be carried out only on those parts of the
software necessary to access the information being sought.

In order to further clarify and strengthen the meaning of subsection 6(1), sub-
section 6(2) emphasizes those purposes which do not justify the decompilation of a
computer program. Any information retrieved through the use of the decompiling
exception cannot be used to achieve something other than interoperability, or to un-
dertake any other task that violates copyright, such as developing software similar
in expression.” ‘ Clearly, in its request that the Commission include an express re-

formation (1989) No 89/C91/05 at 13. See J.D. Byrne, “Computer Programs and Reverse Engineer-
ing: Recent European Developments” (1991) 8 Can. Computer L.R. 45 at 47.

” The Commission’s 1990 proposal stated as background:

IThe Commission has been persuaded that the original proposal, which left the matter
of “reverse engineering” not explicitly regulated, lacks sufficient clarity. It is therefore
proposed that an additional Article 5bis dealing with a derogation allowing “reverse
engineering” of programs for the purpose of interoperability of the program should be
added. Nothing in this Directive should prevent however the “reverse engineering” of a
program, whether incorporated into hardware or not, under the conditions of Article
5bis for the purpose of independently creating an interoperable program, wherever it
may be incorporated (Byrne, ibid at 48).

2″ The enacted Directive contains a definition of interoperability in its preamble:

Whereas the parts of the program which provide for such interconnection and interac-
tion between elements of software and hardware are generally known as “interfaces”;
this functional interconnection and interaction is generally known as
Whereas
“interoperability”; whereas such interoperability can be defined as the ability to ex-
change information and mutually to use the information which has been exchanged
(Directive, supra note 3 at 43).

MCGILL LAW JOUR NAL/REVUE DE DROITDE MCGILL

[Vol. 40

verse engineering section, the European Council felt it necessary to make its inten-
tions, as expressed in articles 1.3 and 5 of the Commission’s 1989 Proposal, as un-
ambiguous as possible in light of the Council’s revolutionary and controversial
stance. The Council wished to ensure that any decompilation would be both limited
in scope and subject to strong barriers to abuse.

Implementation of the Directive’s Provisions in the United
Kingdom

The provisions of directives are not enforceable until they are implemented by
national legislation; however, directives are binding, and must be adopted by mem-
‘m To implement directives, member-states are
ber-states within the prescribed time.
not required to adopt the exact wording used in the directive, although they must
formulate legislation with the same effect. The deadline for implementing the soft-
ware Directive was January 1, 1993, which was met by only a handful of member-
states, including Denmark, Italy and the United Kingdom.”‘

The required changes to British legislation were implemented by the Copyright
(Computer Programs) Regulations 1992, which came into force on January 1,
1993.225 The Directive’s reverse engineering provisions have been incorporated into
section 50B of the British Copyright, Designs, and Patent Act 1988, but they do not
in any way supplant the Act’s fair dealing provisions, which may also potentially be
used to permit reverse engineering. 6 Section 50B allows the reverse engineering of
a computer program in order:

211

Critics of the Directive were quick to point out that the original proposals failed to adequately de-
fine the term “interoperability”. One interest group that had pushed hard on this point was the Soft-
ware Action Group for Europe (S.A.G.E.). S.A.G.E. vehemently argued that without an adequate
definition of “interoperability”, decompilation could be used for purposes other than those intended
by the Directive. That is, without the establishment of strict definitions of the level at which a pro-
gram may validly interface with another, software developers could decompile in a manner that vio-
lates the author’s protected expression.

On the other hand, the European Committee for Interoperable Systems (E.C.I.S.), and other
similar groups support the Directive’s decompilation provisions and favour a more standardized com-
puter environment. Although all groups, including S.A.G.E. and other decompilation opponents, fa-
vour standardization in the computer industry, the two sides prioritize their concems differently.
Groups like S.A.G.E. feel that on balance, there is more to lose than there is to gain through allowing
decompilation in this fashion, while groups like the E.C.I.S. charge that although their opponents’
concerns are not altogether without foundation, groups like S.A.G.E. are overly alarmist and unreal-
istic.2

11 Chance, supra note 217 at 9.
224 “Software Directive” (1993) 9 Clifford Chance Computer & Communications Bulletin 3.

Bainbridge, supra note 18 at 591.

226 It has been pointed out that the existence of an arguably broader fair dealing exception makes a
separate reverse engineering provision unnecessary. Fair dealing may, in fact, allow reverse engineer-
ing for a host of reasons not contemplated by the Directive, such as commercial research (ibid. at 595-
96).

19951

S. HANDA – REVERSE ENGINEERING

(a) to convert it into a version expressed in a higher level language, or
(b) incidentally, in the course of so converting the program, to copy it.

Subsection 50B(2) restricts the use of reverse engineered material, as required by
subsection 6(2) of the Directive, to situations where:

(a) it is necessary to decompile the program to obtain the information necessary
to create an independent program which can be operated with the program
decompiled or with another program (‘the permitted objective’); and

(b) the information so obtained is not used for any purpose other than the

permitted objective.

The United Kingdom’s implementation of the Directive has been criticized for
only exempting the reverse engineering of lower level language code into a higher
language, whereas the “Directive is more generous, allowing translation, adapta-
tion, arrangement or alteration.”22 7 Another potential troublespot in the British legis-
lation is paragraph 50B(2)(b), which potentially restricts a software developer who
has reverse engineered a computer program to create a compatible program, from
later deciding to use the information retrieved to create another, potentially compet-
ing program.2″‘ The use of the “permitted objective” mechanism seems to dictate a
sense of immediacy -, such that the reverse engineering must be conducted with a
given objective in mind, and the results only used in conjunction with that objec-
tive. A strict interpretation of the Act’s wording may require the developer to re-
work the reverse engineering should he or she want to create other compatible
computer programs at some later date. Another notable point, especially for North
American observers, is that under regulation 12(2) of the British Act, the ability to
reverse engineer a computer program cannot be pre-empted by any term in a licens-
ing agreement; any attempt to do so would be void.229

Notwithstanding various technical differences, the British Act has captured the
general spirit of the Directive’s reverse engineering exception, although arguably its
choice of wording serves to indicate a certain discomfort with using legislative ex-
ceptions to relax copyright restrictions. The British legislature seems to prefer to
defer the task of fine tuning copyright protection to the courts through the more
general fair dealing and public interest exceptions.230

2″ Ibid. at 594. Implicitly, the British Act seems to prohibit the translation of a work to an equivalent

level language such as from binary code to hexadecimal notation.

I bid. at 595.

229 Ibid. This applies only to licensing agreements executed after the implementation of the provi-
sions on January 1, 1993. This differs from the right found in section 50C of the Act to copy or adapt
a program as required for its use. The legislation makes no mention of the inability to contract out of
this provision.

o Whether this discomfort is peculiar to Anglo-American systems of copyright, as opposed to con-
tinental drnit d’auteur regimes, is an interesting question which unfortunately lies beyond the scope
of this article.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

C. Australia

If Canada were to look around the world for an analogous jurisdiction in terms
of facing computer copyright issues, Australia would be the natural choice. Since
1984, Australian copyright law has expressly placed computer programs within the
scope of copyright protection, but has left the proposition rather vague as did the
1988 Canadian amendments.2′
In terms of jurisprudential evolution, Canada and
Australia are also similarly situated.232 With respect to reverse engineering, how-
ever, Australian courts are a step ahead of their Canadian counterparts.

Autodesk Inc. v. Dyason

In 1992, the Australian High Court rendered its decision in Autodesk, a case
that concerned the reverse engineering of data tables found within a computer pro-
gram. Once again, the subject matter of the reverse engineering process concerned
a security lockout scheme designed to protect against software piracy. The differ-
ence between Autodesk and the abovementioned American cases, however, is the
manner in which the reverse engineering was conducted. In this case, the defen-
dants used an oscilloscope to map out the signals being sent between a hardware
key, referred to as the “AutoCAD lock”, and the computer program, known as
“Widget C”. In order to run the AutoCAD program, the AutoCAD lock must be in-
troduced into a computer port. When the AutoCAD program is executed, it runs the
Widget C program, which generates a stream of computer signals based on a data
look-up table. These signals are sent to the AutoCAD lock, manipulated, and then
sent back. Only where the returning signals match the profile contained in Widget
C will the AutoCAD program be permitted to proceed. The defendants used an
oscilloscope to detect the signals flowing both to and from the AutoCAD lock, and
figured out the “transitions” being performed by the hardware device. The defen-
dants then constructed their own “Auto Key lock” and marketed it as a substitute
for the AutoCAD lock. Presumably, a large number of the intended purchasers of
the defendant’s product would be persons with pirated copies of the AutoCAD pro-
gram who lacked the AutoCAD lock necessary to make the program run.

Since the disassembly was performed with an oscilloscope and yielded a non-
conventional translation of the original data tables, the Court ignored the issue of

2, S. McGregor, “Look & Feel – Australia” (1993) 1 Mealey’s Litigation Rep. 44.
23’ Whereas a Canadian court recently rendered the first trial decision dealing with the “look and
feel” issue and the protection of non-literal elements in a computer program, Australian courts have
not yet done so (see Delrina, supra note 5).

“Look and feel” properly refers to the manner in which a computer program interacts
with its user (the “user interface”). “Look and feel” includes the manner in which the
graphics, sounds, commands, output and design presentation exist in a given program,
although much of the case law to date has focused solely on the screen display clement
of these programs. “Look and feel”, it must be stressed, does not apply to elements of
program construction that are unrelated to the user interface (Handa & Buchan, supra
note 62 at 50-51).

19951

S. HANDA – REVERSE ENGINEERING

intermediate copying. The Court concentrated instead on the final use of the uncov-
ered expression and ruled that the defendants had reproduced a substantial part of
the Widget C program in their Auto Key lock and had therefore infringed the
plaintiff’s copyright. The Court adopted a very strict interpretation of the language
of the Act, which states that a computer program may be in any material form and
in any notation.23 It also noted that the stream of digits was not random and was
therefore worthy of copyright protection.2
It ignored the fact that the data repro-
duced was unique and purely functional in purpose, and that there was no other
way of creating a sequence that would unlock Widget C.

This case did not raise issues related to the creation of independent interoper-
able computer programs to work in conjunction with AutoCAD. Whether the
Court, in such cases, would be willing to override its finding of copyright infringe-
ment in favour of a fair dealing exception remains unclear. The Court’s decision,
however, foreclosed upon at least one compatible product, the Auto Key lock.
While it is tenable that fair dealing cannot apply to the creation of directly compet-
ing products, this argument, and consequently the scope of a fair dealing and public
interest defence, was never expressly addressed in the case. We are left with the
troublesome impression that this decision was perhaps made without a thorough
appreciation of the compatibility concerns that are potentially at issue in reverse
engineering cases.

Notwithstanding this reservation, Autodesk is sufficiently narrow and non-
specific with respect to the process of reverse engineering that future Australian
courts may allow reverse engineering through a fair dealing exception without of-
fending principles of judicial consistency. As with the decision of the United States
District Court in Uniden, the Autodesk Court did not oppose the process of reverse
engineering. The Court simply concentrated upon the defendant’s final product and
(rightly or wrongly) came to a factual conclusion that it did contain protectable ex-
pression copied from the plaintiff’s program. Whether ignoring the reverse engi-
neering question implies tacit permissibility remains unclear. The decision in Uni-
den, while similarly overlooking the reverse engineering issue, is logically struc-
tured in such a manner that a tacit acceptance of reverse engineering can more
readily be inferred than in the instant case.2* 5

D. Conclusion

Each of the judicial decisions discussed above has either avoided finding re-
verse engineering and intermediate copying to be infringements of copyright, or
has found fair use provisions sufficient to allow the process. None of the decisions
thus far has opposed reverse engineering as an intermediate process. The decisions

2″ Autodesk, supra note 4 at 174.
2Ibid. at 169.
..5For a discussion of this case, see Part I.A. 1, above.

MCGiLL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

all properly concentrate on and apply the infringement tests to the final product.
Considering that European Union member-states are presently bound to implement
reverse engineering provisions in their national legislation, there is a clear trend to-
wards allowing reverse engineering, notwithstanding that it prima facie constitutes
an infringing act.

In the jurisprudence to date, if the use has been to copy protected expression, a
finding of infringement has been made, subject to considerations of compatibility.
The determination as to whether the expression is indeed protected by copyright
has been based on standard principles of copyright law, such as whether a work is
more properly defined as an idea or expression, or whether it is only expressible in
a singular manner (doctrine of merger).2 -6 If the work is protectable, as it was in
Sega, the court may rely on issues of compatibility to hold that a defence of fair use
applies. The following section attempts to further reconcile these different ap-
proaches and to provide an analysis of reverse engineering in light of the goals of
copyright law. Several solutions designed to deal with the problem of reverse engi-
neering in a Canadian context are also recommended.

IV. Justifying the Reverse Engineering of Computer Programs

Conceptually, reverse engineering moves from a finished product to its underly-
ing ideas. Yet even where the goal of the reverse engineering is to obtain the unpro-
tectable ideas, the process typically involves creating intermediate copies, and will
be considered an infringing act. This unique and novel conundrum raises the ques-
tion whether copyright law should be modified to create an exception addressing
the problem of intermediate copying, and when such an exception should be ap-
plied. Clearly, these questions must be viewed in the context of copyright policy
and its intended goals.

A. The Economics of Intellectual Property Rights

The object of intellectual property protection is often described in the language
of economic theory as “public goods”. Public goods are those goods characterized
by non-exclusivity and non-rivalry. Non-exclusivity requires that a good, once pro-
duced, be equally available to all members of a group irrespective of their contri-
bution to its production.27 Non-rivalry occurs where the use of a good by one per-
son will not affect its use by others.2-4

The basic model for setting the optimal, or efficient, price of a non-public good
in a competitive market is to pinpoint the price at the point where it is equal to the

‘ 6 See supra note 51 for a discussion of the doctrine of merger.
37 D. Schmidtz, “Contracts and Public Goods” (1987) 10 Harv. J. L. & Public Pol. 475.
23 Ibid.

1995]

S. HANDA – REVERSE ENGINEERING

cost of producing the last unit (marginal cost).’; This model does not work for a
public good because its marginal cost of production is theoretically zero or very
close to it. Accordingly, it is expected that with a public good such as a computer
program, competitors will flood the market with copies and eventually force the
price towards zero. 24 If the price goes below the author’s cost of producing the
work, given that he or she is a rational economic actor, the work will not be pro-
duced.24′ In order to avoid this result, some form of legalized monopoly protection
is required to combat the effects of non-rivalry and non-exclusivity, thereby giving
the work a number of characteristics naturally attributable to non-public goods. 24′

Certain preconditions are necessary to promote efficiency with tangible prop-

erty:

(i)
(ii)
0H)

Universality: all scarce resources should be owned by someone.
Exclusivity: property rights should be exclusive rights.
Transferability: this is necessary to ensure that resources will be
transferred from low-valued uses to high-valued uses. 3

This list forms the basis for granting property rights protection in Western society,
and may consist of additional factors depending on a given economic perspecitve. 44
Absent any legal rules to the contrary, public goods do not generally display any of
these characteristics because of the principles of non-exclusivity and non-rivalry.

If a legal regime were to impart these characteristics to computer programs, the
other economic concepts that generally underlie modern government policy to-
wards markets would be more readily applicable to these goods. 2
4 Unfortunately,
the story does not end there, since by granting owners of “public good” works these
characteristics, the law is guaranteeing the owners extremely high returns for their

See R. Lipsey et al., Economics, 5th ed. (New York: Harper & Row, 1982) at 200.

240Landes & Posner, supra note 24 at 328.
24, The cost of producing a copyrightable work can best be thought of as comprising two compo-
nents: the cost to the author of producing the work, and the cost of copying and distributing the work.
In order for a work to be created, the difference between expected revenues and the cost of copying
and distribution must be greater than the cost to the author of creating the work (ibid at 326-27).

” Other examples of public goods include: national defence, police protection, road construction

and environmental protection (Schmidtz, supra note 237 at 475-76).

.43 F H. Stephen, The Economics of the Law (Ames: Iowa State University Press, 1988) at 14.
24 Other features identified might include: durability –

that property rights must be granted for
substantial periods and cannot be merely transitory; that inaccessible and unique resources are not
made the object of property rights protections; and that generally individuals, as opposed to groups,
should be given property rights (E. Mackaay, “Informational Goods: Property of a Mirage” (1985) 1
Computer L. & Practice 193 at 195).

24 According to Mackaay,

Information[al goods are] a peculiar commodity. Traditional commodities are captured
in law as physical goods. But information does not coincide with its physical support.
… This poses particular problems to lawyers: the law traditionally attaches itself to
material forms; yet the content, the information, which is immaterial, eludes it (ibid at
194).

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

efforts. These owners may continue to duplicate and distribute their works forever
at virtually no cost. Therefore, in order to regulate the rate of return, the law pro-
vides a time limit on intellectual property works after which the works fall into the
public domain and assume the characteristics of public goods.246

B. The Economics of Reverse Engineering

Framed in the language of the instant debate, the economic analysis of whether
to permit reverse engineering under the law of copyright must consider the benefits
of this activity in light of the goals of copyright law. 47

Software Compatibility

Reverse engineering benefits the public by creating compatible programs and a
standardized environment, which in turn “offers rewards by making programs eas-
ier to use, providing greater productivity, and offering greater networking capabili-
ties.”‘2 4 In the context of technological progress, societal wealth is maximized by
facilitating the creation of a greater number of computer programs. This maximiza-
tion will depend on authors being adequately protected, and hence remunerated for
their efforts. A blanket prohibition of reverse engineering would increase the remu-
neration payable to certain authors and thus stifle the creation and dissemination of

2″6 Copyright is protected in Canada for the life of the author plus 50 years (Copyright Act, supra
note 1, s. 6). There has been much debate as to whether the length of time should be decreased for
computer programs since, given the rate of technological progress, after 50 years computer programs
will be obsolete. Furthermore, as a result of the long term of protection and the purchasing character-
istics associated with computer programs, copyright holders stand to make supernormal returns. In
1984, facing a revision that eventually expressly placed computer programs under the jurisdiction of
the Copyright Act in Canada, the Ministry of Consumer and Corporate Affairs presented a proposal to
the Canadian government that would have limited the length of protection of computer programs un-
der the Act to five years (Consumer and Corporate Affairs Canada, From Gutenberg to Telidon: A
White Paper on Copyright (Ottawa: Department of Communications, 1984) at 79ff [hereinafter From
Gutenberg to Telidon]). This proposal was not acted upon and currently the term of protection for
computer programs under the Copyright Act is the same as for any other literary work (Sub-
committee on the Revision of Copyright, Second Report (Ottawa: House of Commons, 1985) at 46
[hereinafter Second Report]). This lengthy term of protection is consistent with the approach in other
jurisdictions.

” Copyright is not primarily concerned with individual rights per se. Rights attributed to individu-
als are merely incidental to the broader purpose of maximizing societal wealth. This is evidenced by
the wording of Article I, Section 8, of the United States Constitution which states that Congress has
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.” Clearly, the
protection of individual rights is recognized as necessary to societal progress. If copying were freely
allowed, the incentive for authors to create works would diminish to the detriment of society at large.
In such a situation, the cost of obtaining works would also increase, since “[t]here would be a shift
toward the production of works that are difficult to copy; authors would be more likely to circulate
their works privately rather than widely to lessen the risk of copying; and contractual restrictions on
copying would multiply” (Landes & Posner, supra note 24 at 332).

2″ Ignatin, supra note 17 at 2030.

1995]

S. HANDA – REVERSE ENGINEERING

compatible programs.249

A typical countervailing argument posits that banning reverse engineering
would only redistribute wealth in favour of the original program developer rather
than leading to an increase in societal wealth; however, this argument seems to ring
false. If a computer programmer has to pay a licensing fee to link his or her pro-
gram to another program, he or she may decide to omit the link because the fee
could fatally impact on the programmer’s cost/benefit analysis. A reduction in the
number of compatible programs would negatively impact on societal wealth for
several reasons. Without the ability to transfer data between programs, users would
be more likely to stay with one program, whether or not it is best suited to their
tasks. The cost of switching the data would be so high that the use of inefficiently
written programs would become pervasive.

The Costs and Benefits of a Standardized Computing
Environment

Whether the creation of a standardized computing environment should be en-
couraged is best examined from an economic perspective. Arguably the most sig-
nificant benefit to creating standards is that scarce resources are preserved by al-
lowing programmers to use their creative energies to build upon an existing stan-
dardized base. Although many software developers have their own internal stan-
dards for use by “in-house” programmers, clearly this wealth-maximizing effect
would be even more pronounced if industry-wide standards were to exist. Further-
more, scarce resources would be redirected to further technological progress, which
seems to be viewed as a normative good –

although this is debatable.20

Moreover, as development costs would decrease, more programmers would be
able to afford to enter the market and thereby increase competition, which is consis-
tent with the spirit of our society’s economic structure. Even if these new market
players did not possess an absolute advantage over existing firms, they could pos-
sess a comparative advantage through specialization and thus allow a more efficient
allocation of resources leading to higher levels of output and lower prices.2″ A
grant of monopoly protection to any one market player would defeat this effect by
creating barriers to entry and increasing trading friction amongst market partici-

2.. Encouraging “[c]ompatibility allows innovative new products to enter a monopolized market and
lowers development costs by allowing a programmer to attach his single innovative component to a
preexisting complete system,” in effect allowing the programmer to avoid reinventing the wheel
(ibid.).

m Western economies seem to thrive on and encourage technological progress, whether or not it is
a normative good. Technological progress is generally regarded as a societal benefit (Landes & Pos-
ner, supra note 24). When we talk of societal wealth maximization in the context of reverse engineer-
ing we are really talking about reducing inefficiencies and the waste of scarce and valuable resources.
Whether these resources are indeed better used in creating a greater number of products and fueling
technological progress is a purely subjective question.

“‘ Lipsey et aL, supra note 239 at 350-51.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

pants through licensing costs.

From the consumer’s point of view, a standardized computing environment
seems beneficial because it lowers prices. Software consumers also stand to benefit
from savings generated because less time will be required to learn new software
packages.

Notwithstanding the benefits mentioned above, the creation of a standardized
environment would not be without difficulties. One significant disadvantage would
be the risk of “going down the wrong path” of technology, and of only finding out
that it is wrong after having expended a great deal of resources. In such a case, the
transaction costs required to “get on the right track” could be so great that the cor-
rect path might never be chosen. 2 Unfortunately, this risk is not quantifiable and
therefore difficult to compare to the associated advantage of creating a standardized
environment.23

Related to the “down the wrong path” detraction is the possibility that where a
standard becomes pervasive, technological myopia will afflict the industry. Once
again, however, the disadvantages would not always exist in reality. The existence
of better solutions based on the available technology is as unknown as is the degree
to which this myopia deters research into other potentially valuable areas.

Notwithstanding whether the development of a standardized environment is a
benefit or a hindrance, the development of standards in the computer industry ap-
pears to be a naturally occurring phenomenon. One reason for this situation is that
users often need to share computer data files generated by various application pro-
grams. An example might be an accounting firm that does the books for various
businesses. This firm will recommend that its clients use specific software for in-

z2 Ignatin, supra note 17 at 2028.
2 Two recent examples of the negative effects associated with the adoption of a standard are in-
formative: the rejection of the BetamaxM video tape standard in favour of the VHS standard, and,
more related to the instant discussion, the development of the Personal Computer (PC) architecture by
IBM in the late 1970s.

With respect to the latter, computer engineers did not foresee the explosion in computer processing
and memory technologies which occurred as a result of rapid technological progress and the falling
prices of computer chips. The architects of the original IBM PC and its DOS operating system, which
now pervade our society through the distribution of PC “clones”, limited its memory accessing abili-
ties to 1024 kilobytes (KB) of which 640 KB were accessible to application programs. This restriction
is now widely referred to in the computer industry as the “640KB barrier”. Generations of PC tech-
nology computers and software that followed on the heels of the hugely successful PC were also
forced to incorporate the 640KB barrier in order to retain compatibility with other generations of per-
sonal computers. To many manufacturers and users, retaining compatibility with the PC standard was
a priority even though alternate, more effective, technologies were readily available soon after the
PC’s advent. The reasons for programmers’ loyalty to this inferior standard were based on cost, mar-
keting, and availability of compatible software. Anyone in the computer industry is aware that the
limitations brought about by the 640KB barrier have hampered software development over the years,
and resulted in huge inefficiencies reflected as opportunity costs (economic losses) in the computer
software industry and in society as a whole.

1995]

S. HANDA – REVERSE ENGINEERING

house recordkeeping that is compatible with the firm’s own accounting software.
Once these relationships develop, an industry standard surfaces, since users want to
retain the greatest level of compatibility with other users to ensure reduced costs
(through greater competition).

Software developers are themselves often the greatest proponents of a standard-
ized environment. One of the most effective ways of marketing a computer pro-
gram is to make it compatible with as many existing programs as possible. The
newer program will allow users of existing programs to switch easily to it by pre-
sumably incorporating new technological advantages. Once the software developer
has captured the market it seeks and developed a standard of its own, however, it
often changes its tune; it will then seek to protect its standards so that it may exer-
cise monopoly power over its users and obtain greater profits.

Another example of the industry’s tendency towards adopting standards is the
increased use of object-oriented programming. Object-oriented programming uses
standardized routines, known as “objects”, and ties them together with other objects
and with one’s own code to create a finished product. Objects are generally fine-
tuned to accomplish a specific task and are typically superior to similar program
procedures that are created each time a program is written. An object-oriented ap-
proach allows more-effective programs to be written more efficiently.

In summary, a standardized computing environment has associated economic
benefits and costs, although the costs tend to be less quantifiable. Given this state of
affairs, it would seem preferable to pursue a standardized computing environment.
Furthermore, since the phenomenon is naturally occurring, pursuing standardiza-
tion involves little or no legal intervention. Conversely, preventing the development
of standards would apparently require legislation and would defeat the benefits as-
sociated with a standardized environment. The justification for pursuing this latter
course of action should involve something more than unquantifiable probabilities
of negative consequences. Mere uncertainties are not sufficient to prompt interven-
tion into this naturally occurring economic order. The proverb that “a bird in hand
is better than two in the bush” would seem to apply to this thinking. The pursuit of
a standardized computing environment based on available information should be
considered an economic good, and therefore merits a laissez-faire approach.

Because granting monopoly protection over standardized computer environ-
ments would reduce their potential efficiencies, a strong reason should be required
for such a grant. An analysis of copyright law, the fundamental goal of which is to
maximize the goal of societal progress in the fields of arts and technology by pro-
tecting authors, does not yield any clear signals as to the possible effect on societal
progress of a prohibition of reverse engineering. At worst, the signals are mixed,
and at best, they lean in favour of allowing reverse engineering because it leads to
the development of a standardized computing environment. If reverse engineering
is to be prohibited, more convincing data and a clearer indication of the prohibi-
tion’s beneficial effects must be demonstrated.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

C. Developing a Solution to the Problem

Once it is accepted that reverse engineering should be encouraged, a solution to
the potential copyright infringement problems must be developed. The two possible
models are the creation of a statutory exception (the European approach), and ex-
pansion of the fair dealing exception to cover reverse engineering (the American
approach).

Creating a Statutory Exception

A statutory exception can be most effectively used to clearly delineate the de-
sired scope of a given legal exemption. Furthermore, the use of a statutory excep-
tion would allow the legislature to preclude overrides of copyright terms, which oc-
cur through trade secret licensing, as is the case with the British legislation.SM

Because the use of trade secrets in cases involving reverse engineering is not
altogether consistent with the breach of confidence pre-emption clause in section
63 of the Canadian Copyright Act, a statutory section drafted with the purpose of
allowing reverse engineering should correct this inconsistency.255 Accordingly, such
a section should expressly provide that, in the case of reverse engineering computer
programs, the breach of confidence exception outlined in section 63 of the Copy-
right Act shall not apply.”6 Furthermore, since we cannot foresee future challenges
to current thinking, the statutory section should be sufficiently flexible and avoid,
to the extent possible, any mention of specific technologies. A broad reverse engi-
neering right to create an authorized (legitimate) copy of a computer program
should result. The section should expressly provide that it creates an exception to
intermediate copying only for the purposes of reverse engineering, and not for use
of the materials once the reverse engineering is complete. If at the end of the re-
verse engineering process, those reverse engineering the product wish to use the
fruits of their labour, they will still have to respect the limits of the idea/expression
dichotomy. That is, they will be able to appropriate the uncovered ideas but not the
protected expression.

The implementation of a statutory exception to reverse engineering is not an
altogether novel idea in Canada. A recommendation which implicitly favoured the
enactment of such an exception and expressly placed computer programs within the

2″ For an analysis of the British legislation, see Part 11.B.2, above.
.. A strong argument can be made that the fundamental purpose of advancing a trade secrets argu-
ment to prohibit reverse engineering is really the prevention of competition (see text accompanying
note 121, above).

..6 Implementation of such a section will undoubtedly require redrafting section 63 so that exclusive
jurisdiction of the Copyright Act and the exception for breach of confidence laws can be delineated
into sub-parts, such as paragraphs 63(a) and 63(b). If, for example, the breach of confidence excep-
tion was placed in paragraph 63(b), the reverse engineering exception would then read “section 63(b)
shall not apply to the operation of this section.”

1995]

S. HANDA – REVERSE ENGiNEERING

jurisdiction of the Copyright Act was made to Parliament in 1985. 2
7 Although the
recommendation suggested that the government further study issues concerning the
shared use of sub-programs, the government decided against doing so, and no such
statutory exemption was ever enacted.28 Since then, both technology and its legal
protection have changed. The Canadian government’s unwillingness to implement
the “shared sub-program” exception in the mid-1980s was not accompanied by any
statement that this route would be forever closed. Its refusal should not be taken as
indicating any more than a cautious approach to what were uncharted waters. Fur-
thermore, the-“shared sub-program” proposal was far more revolutionary than any
reverse engineering exception would ever be, since it proposed an exception to the
general copyright principle that prohibits the substantial copying of a work. The re-
verse engineering exception was merely necessarily incidental to the ultimate pur-
pose mandated by that proposal. A true reverse engineering exception, as proposed
in this article, would only challenge traditional copyright principles insofar as it
would allow intermediate copying; it would not exempt the use of what was uncov-
ered by the process of reverse engineering from traditional copyright principles.

Such an enactment would no longer be revolutionary, given that the nations of
the European Union will each soon have statutory exceptions permitting reverse
engineering and that recent American decisions allow reverse engineering as a fair
use. In fact, Canada’s failure to explicitly allow the reverse engineering of com-
puter programs may soon place it in the minority amongst industrialized nations.
Clearly, the time has come to reconsider the adoption of such a statutory exception.

Relying on the Fair Dealing Exception

Another approach to the problem of reverse engineering would be judicial or
legislative expansion of the fair dealing defence. 29 Legislative changes to fair

2″ Second Report, supra note 246 at 46. The Sub-Committee on the Revision of Copyright recom-
mended that the government “study the possibility of providing an exception to permit the reproduc-
tion of a substantial part of a pre-existing program as a non-substantial part of another program.” This
would have created an exception similar to the reverse engineering exception contained in the
I.C.T.A. The Sub-Committee felt that the innovation of computer programs would be accelerated
through an exception which allowed the shared use of modular constructs. The Sub-Committee also
found that this was a “normal and healthy” practice in the computer industry, and that the law should
not impose costs on the industry by preventing the sharing of program code. Although no mention
was made of reverse engineering in the copyright revision proposals, it is clear that in order to use
sub-programs interchangeably with one’s own computer programs, one would need the specifications
and parameters. Reverse engineering would be the only means by which to obtain these specifications
and parameters, if they were not provided by the manufacturer of the first program.

“‘ B. Romaniuk, “Are Computer Software and Integrated Circuitry Legally Vulnerable to Reverse

Engineering – Part One” (1986) 3 Can. Computer L.R. 177 at 178-79.

“” The recommendation discussed in this section applies equally to a public interest defence. How-
ever, since this defence is not statutory, and has rarely received judicial notice in the field of copy-
right, it has relatively little chance of being applied when compared with fair dealing (see Part
ILA.1f iv, above).

MCGILL LAW JOURNAL/REVUE DE DROITDEMCGILL

[Vol. 40

dealing have in the past been recommended2W and subsequently rejected.26′ Legis-
lative widening of fair dealing for the sole purpose of providing an exception to re-
verse engineering would presumably elicit similar hostility. Because fair dealing is
supposed to apply generally to works protected under the Act, it would make far
better sense to enact a sui generis exception, as set out above.

A judicial widening of fair dealing, on the other hand, would avoid these diffi-
culties by putting the problem entirely in the hands of judges. Moreover, this solu-
tion provides greater flexibility than would a legislative exception. Further, judicial
intervention would not run the risk of being tied to statutory language which might
not fully contemplate future challenges to our potentially limited current knowl-
edge. Given our American neighbours’ application of their fair use doctrine to re-
verse engineering cases, it would certainly be no surprise to see Canadian courts
similarly expand fair dealing. After all, as copyright is used more frequently in
situations which were not contemplated by the drafters of the Act, the fair dealing
exception, designed to provide courts with flexibility, should not remain stagnant.
In accordance with the language of section 27 of the Copyright Act, in order to al-
low reverse engineering under fair dealing, a court would be forced to qualify the
reverse engineering as occurring for the purposes of either “research” or “private
study”. Any intermediate copying required in the reverse engineering process
would be conceived as fair dealing since the ultimate goal of the process is to un-
cover underlying ideas or unprotectable expression for the purposes of achieving
compatibility and standardization.262

The other side of the argument in favour of judicial intervention is, of course,
the uncertainty of waiting for a court to apply fair dealing to any new subject mat-
ter. Given the juridical history of fair dealing, it is not clear that the courts will at-
tempt to delineate the concept any time soon. Furthermore, a court must be pre-
sented with an infringement claim concerning reverse engineering and a defence
based on fair dealing before any judicial widening is even conceivable. Until such a
case is presented, the uncertainty associated with the permissibility of reverse engi-
neering under copyright law will create an environment that could potentially dis-

260 In From Gutenberg to Telidon, supra note 246 at 39, it was recommended that fair dealing, be-
cause of its “lack of statutory definition”, be replaced with a fair use section containing a “prioritized
list of factors to be considered in determining whether a particular use of a work is a fair use.” Pre-
sumably this list would be similar to that provided in the American Act.

262 A recommendation

26′ Second Report, supra note 246 at 65-66. The Sub-Committee on the Revision of Copyright rec-
ommended, ibid. at 64-65, that fair dealing be retained and that a prioritized list of factors not be en-
acted as “the flexibility so essential to fair dealing would be destroyed by the fact that they would be
mandatory and exhaustive.” The Sub-Committee, however, was not against the further use of illustra-
tive, non-mandatory factors.

that the “research” objective, as outlined in the fair dealing section
(Copyright Act, supra note 1, s. 27), be revised to read “private research” was rejected by the Cana-
dian Parliament (Second Report, ibid. at 66). The intent of this proposal was to preclude commercial
organizations from making use of the fair dealing defence. The rejection of this proposal suggests that
commercial organizations are indeed allowed to use the fair dealing defence to justify research and
development activities.

1995]

S. HANDA – REVERSE ENGINEERiNG

suade reverse engineering, and result in the loss of potential efficiencies and socie-
tal wealth. Should reverse engineering have been expressly permitted, it is un-
known what products, technologies and standards would have developed. Finally,
as already mentioned, using fair dealing to permit reverse engineering would not
extend to cases of trade secret protection by virtue of section 63 of the Copyright
Act.”” If, however, it is determined that existing trade secret protection constitutes a
valid limit on the scope of allowable reverse engineering, section 63 would not
pose any difficulty. 64

As a result of the apparent limitations on a solution framed in the context of fair
dealing, a clear legislative statement would probably be more successful in effec-
tively resolving the problems associated with reverse engineering computer pro-
grams.

Alternatives to Copyright

A third, more revolutionary, method of dealing with reverse engineering is one
which many critics have been calling for: the redrafting of intellectual property
protections as they apply to computer programs. Ever since the protection of com-
puter programs was recognized as a problem, there have been calls for the enact-
ment of sui generis legislation designed to deal with it. As a clearer understanding
of the economic importance of computers has emerged, coupled with the law’s
relative inability to effectively deal with the subject matter through existing legisla-
tion and judicial decision making, the calls for computer program protection other
than copyright have been received less skeptically. Unfortunately, global accep-
tance of copyright protection of computer programs has now largely precluded a
complete shift away from copyright. The resulting inertia is similar to the “going
down the wrong path” difficulties, mentioned previously with respect to standardi-
zation.6

As the differences between computer programs and more traditional copyrigh-
table works become increasingly apparent, and as more legislative exceptions be-
come warranted, it would seem logical to eventually enact legislation to supplement
the Copyright Act, and cover those aspects for which copyright protection is not
-appropriate. For example, there have been many difficulties in relation to the pro-

… It is likely that under the present copyright regime, the use of a valid trade secrets argument
would be sufficient to supplant an argument of fair dealing under the Copyright Act. As previously
mentioned, however, it is unlikely that the mere use of shrink-wrap licensing will constitute a suffi-
cient relationship between the parties so as to allow the trade secrets argument to be successfully in-
voked.

.6. There is a remote possibility, however, that the courts could broadly construe compiled object
code as confidential and thus be unwilling to apply fair dealing as a result of Beloff. This argument, if
successful, would grant computer programs special additional protection because of their form. While
this type of confidentiality protection exists within the realm of trade secrets, it is beyond the intended
scope of copyright. For an analysis of this case, see text accompanying note 98, above.

“5 See text accompanying note 252, above.

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

tection of computer screens. The Canadian government has indicated that it would
not afford such protection under its copyright legislation,2 however, under certain
circumstances, the courts feel that screens are indeed the proper subject matter of
copyright.26 The legal reasoning of the decisions, largely influenced by American
case law, is tenuous and does not offer regulatory certainty to those in the industry.
This is not the fault of the judiciary, who have been given poor guidance from the
copyright legislation and are thus ill-equipped to couch their decisions in clear and
simple terms. A more sound approach would be to supplement the copyright legis-
lation with laws expressly designed to handle those issues unique to computer pro-
grams. An exception to reverse engineering may be more at home within such leg-
islation, as reverse engineering is similarly unique to computer programs. While it
is beyond the scope of this article to present such a sui generis regime of computer
program protection, one should nonetheless be aware of the possibility of supple-
menting copyright with other legislation.

Related to that possibility is the more realistic notion that patent protection will
become increasingly important as a form of intellectual property protection for
computer programs. Patent legislation is more and more frequently being used to
protect computer programs. Computer programs protected under the patent regime
must fully disclose the manner in which they operate. The trade-off for meeting the
more rigorous requirements of the patent system is the greater monopoly protec-
tion. Unlike under the copyright regime, even if a patented product is reproduced
completely independently of the original product, a royalty must still be paid to the
patent holder. Although the term for patent protection is considerably less than that
for copyright, it is likely that the term will be more than adequate for computer
programs. With patent protection, there is no need for a reverse engineering right
because full disclosure is mandatory prior to obtaining the patent, and the relevant
information is subsequently made publicly available.

D. The Scope of a Reverse Engineering Right

If an exception to copyright is to be applied, it should not be restricted to the
discovery of ideas, as was held in Atari. Similarly, permission to reverse engineer a
computer program should not be restricted to those parts of a computer program es-
sential to understand the underlying ideas, but should rather be construed broadly.
Attempting to limit the process to only those parts of a computer program that
capture the underlying ideas is conceptually difficult and sometimes impossible.
Furthermore, a specific enquiry that attempts to uncover the various ideas within a
program runs the risk of ignoring the interaction of those ideas. This interaction is

2 In their report to the House of Commons, the Sub-Committee on the Revision of Copyright rec-
ommended that “there should be no right of display [with respect to user interface screens] in the re-
vised law” (Second Report, supra note 246 at 41). At the time, the government responded by stating
that it would study this recommendation in further detail although the first round of amendments to
the Act did not address these recommendations.

267 See Delrina, supra note 5; Gemologists International Inc. v. Gem Scan Iternational Inc. (1986),

9 C.PR. (3d) 255; 7 C.I.P.R. 225 (Ont. H.CJ.).

1995]

S. HANDA – REVERSE ENGINEERING

so conceptually abstract that it too falls within the realm of ideas. Restricting re-
verse engineering in this manner is wholly consistent with copyright law and its
intended purpose of protecting expression and not ideas.

Similarly, reverse engineering which seeks to uncover expression should also
not be restricted. However, idea and expression are so intertwined that those seek-
ing to reverse engineer a computer program will always claim that they are attempt-
ing to discover an underlying unprotectable idea. The increased potential for erro-
neous judicial decision-making coupled with the waste of resources in bringing
such actions to court makes them undesirable. Furthermore, at the end of the day,
any decision to allow or prohibit reverse engineering deters from the more appro-
priate enquiry as to whether protectable expression was incorporated into the newly
developed computer program. It is extremely difficult for a court to determine a
priori whether reverse engineering was meant to uncover idea or expression, but a
court could much more easily decide whether protectable expression had been in-
corporated into a new computer program. Furthermore, there is a convincing argu-
ment to be made that any ideas, however small, underlying a computer program are
beyond the mandate of copyright protection. If one is to allow reverse engineering
to uncover “large” ideas, whatever that may mean, then surely such reasoning must
extend to all ideas. The fact that computer programs exist in a naturally encoded
state should not cause copyright law to prevent the uncovering of ideas, whatever
their relative importance.

If protectable expression is uncovered during reverse engineering and is illicitly
used in the construction of a new computer program, this use would violate copy-
right independently of the act of reverse engineering.2′ The holder of the first pro-
gram’s copyright would simply bring an action for infringement against the creator
of the infringing work. The fact that reverse engineering was performed would not
be in issue. Given the potential liability for infringement if copied expression is
used in the construction of another program, the only justification for such an ac-
tion would be that the computer program being reverse engineered contains ex-
pression which somehow impedes access to or compatibility with it, as in the

s This is akin to the first argument made by Accolade in Sega, supra note 4 at 1517, where Acco-
lade “maintain[ed] that intermediate copying does not infringe the exclusive rights granted to copy-
right owners in section 106 of the Copyright Act unless the end product of the copying is substantially
similar to the copyrighted work.” Accolade lost on this argument because the Sega Court felt bound
by its decision in Walker v. University Books, Inc., 602 F2d 859 (9th Cir. 1979) [hereinafter Walker],
which concerned the intermediate copying of fortune-telling cards. The Walker decision was framed
in broad language, which the Sega Court felt must be applied to computer programs. The Court was
therefore unwilling to provide a less restrictive interpretation of the fair use doctrine, preferring in-
stead to allow reverse engineering only where the purpose of the process was to gain an understand-
ing of ideas and purely functional concepts, not protected by copyright, which are embodied in a
computer program. Had the Court decided otherwise, it might have distinguished the Walker decision
based on the uniquely functional nature of computer programs. That is, intermediate copying for
more traditional works should not be equated to intermediate copying for computer programs since
the nature of the copying is for altogether different purposes. Oddly enough, this argument is not a far
cry from the “purposeful” analysis engaged in by the Sega Court.

MCGILL LAW JOURNAL/ REVUE DE DROIT DE MCGILL

[Vol. 40

aforementioned lock-out cases.269 Although it is entirely within the right of a pro-
grammer to create such mechanisms, it is not within the intended purpose of copy-
right law to provide additional legal protection to such devices. If the expression
used is found to be protectable independently of the reverse engineering issue, then
an action for infringement will be successful in any event.

Conclusion

Canadian courts have not yet faced a case of reverse engineering, but it is
nonetheless prudent for our legislators to consider the issue prior to such a court
challenge, where the results are bound to be unpredictable. Thus far, at least thir-
teen other Western nations have confronted the problem and have accepted that the
reverse engineering of computer programs, under certain circumstances, does
constitute a valid exception to copyright infringement. Given the increasing reli-
ance of the Canadian economy on the software industry, it is only a matter of time
before the question of reverse engineering is raised in Canada. Furthermore, by
confronting this issue earlier rather than later, the legislature has an opportunity to
accelerate technological progress in the computer software field. This reasoning
underlay the formulation of the decompilation provisions in the European Com-
munity’s Software Directive, and was similarly recognized by the Canadian Parlia-
ment in the reverse engineering provisions of the Integrated Circuit Topography
Act.

The case for creating some form of reverse engineering exception arises from
the fact that, practically speaking, the only manner in which a computer program
may be disassembled constitutes an infringement of the copyright protections
granted to computer programs under the Copyright Act. Economically, the advan-
tages to allowing reverse engineering far outweigh the disadvantages and remain
consistent with the goals of the copyright legislation. An economic perspective was
chosen as the appropriate lens through which to examine the issue because, in an
Anglo-American context, the origins and basis of copyright protection are funda-
mentally rooted in economic issues. By applying the copyright legislation as it cur-
rently stands, thereby prinafacie prohibiting reverse engineering, the serious threat
arises of a monopoly over functional standards. Copyright legislation was designed
well before the existence of computer technology, and its application to this field is
awkward at best because it was not designed to protect ideas or purely utilitarian
works. By placing computer programs under the umbrella of copyright, the copy-
right monopoly may now stifle the development of standards and slow the pace of
technological advancement. Such a result is in direct contrast to the stated purpose
of copyright law. Preventing the disclosure of a computer’s functional processes to
the public at large does not occur in any form of intellectual property protection

. According to Ignatin, supra note 17 at 2022-23, “[t]he inadvertant protection of ideas under
copyright may suggest that reverse engineering should be permitted in all instances [although it] is
particularly justified [in] making a new program compatible with existing copyrighted software.”

1995]

S. HANDA – REVERSE ENGINEERiNG

other than copyright as it applies to computer programs.

Although the American legislature has thus far remained silent on the matter,
higher American courts have faced reverse engineering issues in several cases and
have begun to carve out a reverse engineering exception for intermediate copying,
based on the defence of fair use found in American copyright legislation. The rea-
soning of the Ninth Circuit Court of Appeals in Sega is essentially correct. The
Court’s purposive analysis, however, which requires that reverse engineering be
limited to only those elements that are not protected by copyright, falls just short of
the mark. Unfortunately, the Court felt constrained by the language of an earlier
decision 270 concerning the intermediate copying of traditional literary works, and
did not recognize the special and unique nature of computer programs.

Although the scope of the reverse engineering right under fair use remains
relatively restricted as compared with the European Community’s legislative pro-
visions, the extent to which American courts will allow reverse engineering has not
been fully settled. The Canadian Copyright Act’s fair dealing defence, although not
as evolved as the fair use defence, presents Canadian jurists with the flexibility to
allow the intermediate copying of computer programs necessary for reverse engi-
neering, notwithstanding the fact that such copying is prima facie an infringement.
Although fair dealing would allow courts to circumvent the intermediate copying
problem, it would be more suitable for Parliament to implement an express excep-
tion for reverse engineering that construes the process broadly, rather than restric-
tively.

Broadly construing a statutory exception is wholly consistent with the purpose
of copyright protection, and would continue to protect the expression of a work
from being copied. This position is, however, revolutionary in light of those excep-
tions that have thus far been passed. There is a general consensus that reverse engi-
neering should be limited to cases where compatibility with a computer program is
sought; however, the difficulty and subjectivity of such a limitation renders it ef-
fectively meaningless and highly uncertain in its application. It is extremely diffi-
cult to determine if one who reverse engineers a computer program truly does so
for purposes of compatibility. Furthermore, in cases where reverse engineering oc-
curs, it is logical to assume that no specifications have been provided by the origi-
nal program designers. It is therefore very difficult for the individual performing
the reverse engineering to determine ex ante exactly what she or he is looking for,
since effectively interfacing with a computer program requires an understanding of
its general structure as well as specifics about its operational characteristics. A court
is thus in a poor position to decide at what point the reverse engineering process no
longer concerns compatibility. Even if such a determination could be made, it
would serve little purpose in the context of the reverse engineering process. Instead,
the determination of whether compatibility is being achieved should be made when
examining the allegedly infringing computer program. At this stage, if it is deter-

2’0 See supra note 268.

690

MCGILL LAW JOURNAL/REVUE DE DROITDE MCGILL

[Vol. 40

mined that expression that does not relate to compatibility has been used, a decision
of infringement can be rendered. If a program is reverse engineered and is not oth-
erwise copied, there is no loss to the original owners, save the exposure of underly-
ing ideas, expression and processes to the reverse engineer. To keep the entire un-
derlying structure of a computer program secret because of its binary form properly
falls under the law of trade secrets and is well beyond the mandate of copyright.

Reverse Engineering Computer Programs Under Canadian Copyright Law

Table of Contents

Subscribe to our Newsletter

Thank you to our sponsors