Judgments

Decision Information

Apple Computer, Inc. v. Mackintosh Computers Ltd.

Collection Judgments
Date 1987-10-13
Report citation [1988] 1 F.C. 673
File numbers A-275-86
Subjects Copyright

Decision Content

A-275-86
Mackintosh Computers Ltd., Compagnie d'Élec- tronique Repco Ltée/Repco Electronics Ltd., Maison des Semiconducteurs Ltée/House of Semiconductors Ltd., Chico Levy and Nat Levy (Appellants) (Defendants)
v.
Apple Computer, Inc. and Apple Canada Inc. (Respondents) (Plaintiffs)
A-276-86
115778 Canada Inc., carrying on business under the firm name and style of Microcom, James Begg and 131375 Canada Inc. (Appellants) (Defen- dants)
v.
Apple Computer, Inc. and Apple Canada Inc. (Respondents) (Plaintiffs)
INDEXED AS: APPLE COMPUTER, INC. v. MACKINTOSH COM PUTERS LTD.
Court of Appeal, Mahoney, Hugessen and Mac- Guigan JJ.—Ottawa, September 28, 29, 30 and October 1 and 13, 1987.
Copyright — Infringement — Duplication of computer chip infringing copyright in computer program embodied therein Conversion of program from assembly language to machine language not "translation" within Act s. 3(1)(a) — Embodi ment of program in chip reproduction of work in material form, covered by copyright — For copyright purposes, pro grammed chip to be assessed in material static state.
The appellants duplicated two computer programs (Auto- start ROM and Applesoft) used by the respondents in their Apple II Computer. The appellants did not copy the programs as written in assembly code. Instead, they reproduced, appar ently by mechanical means, the silicon chip in which the programs are embodied.
These are appeals from the Trial Judge's finding that copy right in the computer programs had been infringed.
Held, the appeals should be dismissed but the judgment of the Trial Division should be varied by awarding relief to Apple Computer, Inc., only.
Per Mahoney J.: The argument that programs in machine language merely constituted specifications for the manufacture of the chips is totally irrelevant. Copyright in the machine language version of the programs is not in issue. Furthermore,
the chips were not made according to instructions of the machine language.
The Trial Judge erred in finding that the conversion of the programs from the original assembly language to the hexadeci mal machine language was translation within the meaning of paragraph 3(1)(a) of the Act. It is rather a reproduction of the original, protected, like the original, by copyright.
The programs embodied in the chips are reproductions of the programs. The Trial Judge was correct in approaching this case on the basis that the reproduction in which copyright was claimed and infringement alleged consisted of the static circuit ry of the chips, not the dynamic sequence of electrical impulses dictated or permitted by that circuitry. The opening words of section 3 of the Act, according the copyright owner "the sole right ... to reproduce the work ... in any material form whatever", were drafted broadly enough to encompass the embodiment of the respondents' programs in a chip. Nothing in the Act required that a reproduction be in a humanly readable form to be subject of copyright or to infringe copyright. So when the appellants copied the programs as embodied in the chips, they infringed the respondents' copyright in the programs.
It is unnecessary to decide whether the chips are contrivances by means of which the work may be delivered within the meaning of paragraph 3(1)(d) of the Act.
Per Hugessen J.: The anthropomorphic character of every thing that is said or written about computers—words with cognitive connotation like "language", "memory", "read"— creates a real difficulty. Thus, even if the programs were transferred from one language to another this was not a "translation" within the meaning of the Act—the turning of something from one human language to another. Nor were the appellants' chips contrivances within the meaning of paragraph 3(1)(d). The programs are not works which may be mechani cally "performed" or "delivered".
However, the Trial Judge was correct in finding that the chips embodying the programs were reproductions of the assembly code programs in which the respondents held the copyright. For the purposes of the Act, when a literary work is produced or reproduced, the result must necessarily be a liter ary work (expressed in print or writing). In this case, since the chips contain no print or writing, their production would not, at first glance, constitute an infringement of copyright. However, "the sole right to produce or reproduce in any material form" includes by necessary implication the sole right to produce the means of reproduction of the work. This interpretation is reinforced by sections 10 and 21 of the Act which deal with ownership of the means of reproduction of works protected by copyright. And even if the means of reproduction does not itself constitute a reproduction within the meaning of subsection 3(1), the implied authorization to reproduce—present in this case given that the chips in the appellant's computer were
capable of being used to reproduce the respondents' programs in their hexadecimal code version—is an infringement of the copyright.
Per MacGuigan J.: In Computer Edge, an identical case, the High Court of Australia decided that there was no infringe ment of copyright. However, the Australian legislation inter preted in that case was based on a different English statute than that on which the Canadian legislation is based. Further more, the point of view on which that decision is based, that the program was in reality a series of electrical impulses stored in the chips, is inconsistent with the findings of the Trial Judge, which were not successfully challenged. And it has been decid ed, in a United States Court of Appeals decision, Apple Com puter, Inc. v. Franklin Computer Corporation, that there is no reason to afford less protection to an operating system program, as in this case, than to an application program.
STATUTES AND REGULATIONS JUDICIALLY CONSIDERED
Copyright Act, R.S.C. 1970, c. C-30, ss. 2, 3, 4, 10, 17(2)(b), 21.
Copyright Act, 1911, 1 & 2 Geo. V, c. 46 (U.K.). Copyright Act, 1956, 4 & 5 Eliz. II, c. 74 (U.K.). Copyright Act, 1968 (Cth), No. 63 (Austl.).
Federal Court Act, R.S.C. 1970 (2nd Supp.), c. 10, s. 52(b)(i).
CASES JUDICIALLY CONSIDERED
DISTINGUISHED:
Computer Edge Pty Ltd v Apple Computer Inc (1986), 65 ALR 33 (H.C.).
CONSIDERED:
Compo Company Ltd. v. Blue Crest Music Inc. et al., [1980] 1 S.C.R. 357; Moreau, Alfred v. St. Vincent, Roland, [1950] Ex.C.R. 198; Thrustcode Ltd. v. W.W. Computing Ltd., [1983] F.S.R. 502 (Ch.D.); University of London Press v. University Tutorial Press, [1916] 2 Ch. 601; Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240 (3rd Cir. 1983).
REFERRED TO:
Cuisenaire v. South West Imports Limited, [ 1969] S.C.R. 208; Boosey v. Whight, [1899] 1 Ch. 836, affirmed [1900] 1 Ch. 122 (C.A.); Ladbroke (Football) Ltd. v. William Hill (Football) Ltd., [1964] 1 W.L.R. 273 (H.L.).
COUNSEL:
Robert H. C. MacFarlane for appellants. Alfred S. Schorr, Ivor M. Hughes and Joseph I. Etigson for respondents.
SOLICITORS:
Fitzsimmons, MacFarlane, Toronto, for appellants.
Ivor M. Hughes, Concord, Ontario, for respondents.
The following are the reasons for judgment rendered in English by
MAHONEY J.: I have had the opportunity to peruse the reasons in draft of my brother Mac- Guigan herein. While I agree with him in the result, I do so on what I consider significantly narrower grounds.
These appeals are from judgments [[1987] 1 F.C. 173; 10 C.P.R. (3d) 1 (T.D.)] in two actions tried together on what, for all purposes of the issues we have been asked to deal with, was common evidence. The appellant, James Begg, abandoned his appeal against the finding of liabili ty in his personal capacity as such. The other individual appellants had not raised that ground of appeal. The respondents abandoned their allega tion of infringement by reason only of the alleged making of copies of three-dimensional objects themselves made according to designs or specifica tion in which copyright subsisted. It was acknowl edged that the respondent, Apple Computer, Inc., was the only plaintiff entitled to the remedies granted and that, therefore, in any event of the appeals the judgments of the Trial Division should be varied accordingly.
To paraphrase Mr. Justice Estey in Compo Company Ltd. v. Blue Crest Music Inc. et al., [1980] 1 S.C.R. 357: arising out of a rather com plex set of facts is the simple issue whether a person who duplicates a computer chip thereby infringes the exclusive rights granted by section 3 of the Copyright Act, R.S.C. 1970, c. C-30, to the holder of the copyright in the computer program embodied in that chip. There are two programs in issue here. Reasons for the judgments under appeal have been fully reported in 10 C.P.R. (3d) 1, to which I shall refer. The report at [1987] 1
F.C. 173 is an abridgment which does not recite many essential findings of fact.
The provisions of the Act primarily in play are paragraphs 3(1)(a) and (d).
3. (1) For the purposes of this Act, "copyright" means the sole right to produce or reproduce the work or any substantial part thereof in any material form whatever, to perform, or in the case of a lecture to deliver, the work or any substantial part thereof in public; if the work is unpublished, to publish the work or any substantial part thereof; and includes the sole right
(a) to produce, reproduce, perform or publish any translation of the work;
(d) in the case of a literary, dramatic, or musical work, to make any record, perforated roll, cinematograph film, or other contrivance by means of which the work may be mechanically performed or delivered;
The appellants concede that the original computer programs, written in assembly language, were lit erary works within the meaning of the Act and were subject of copyright owned by the respon dent, Apple Computer, Inc. They further concede that, if the programs as embodied in the chips were subject of that copyright, they were substan tially copied and, thereby, infringed.
The meaning of "translation", as Parliament intended it in paragraph 3(1)(a) of the Act, is very much in issue. I shall therefore avoid using it in the following recitation of facts and adopt the term "conversion" which I intend to be entirely neutral.
THE FACTS
The learned Trial Judge found, at page 7, C.P.R. the following:
A computer language, of which there are many, is a code for writing a program. A language is said to be "higher" or "lower" depending upon the ease with which it can be read. A high level language has symbols and rules that correspond closely enough to ordinary mathematics and English (or other common language) that it may be read and understood with relative ease. Examples are languages such as BASIC, COBOL, PASCAL and FORTRAN.
A second level of language, which can be referred to as an intermediate level, consists of mnemonics which correspond more explicitly to the operations the computer must perform. For example, in the excerpt set out [below], LDY stands for "load index Y with memory", and STA stands for "store accumulator in memory". This intermediate level is referred to as assembly language.
A third level of language, the lowest, is sometimes referred to as machine language or object code. There are two versions of machine language relevant for the purposes of this case: a system of hexadecimal notation and a system of binary nota tion. I do not think it necessary to describe the two in detail except to say that hexadecimal notation is based on a number system having a base 16, while binary notation is based on a number system having a base two. The hexadecimal system of notation is merely a shorthand way of writing the binary code.
Binary notation utilizes only the symbols "1" and "0". Hexadecimal notation utilizes the ten numer als "0" to "9" and the first six letters of the alphabet "A" to "F".
The programs in issue were originally written in assembly language. A section of one of them, recited by the learned Trial Judge at page 6 C.P.R., read:
LDY #$2F
BNE CLRSC2
LDY #$27
STY V2
LDY #$27
LDA #$00
STA COLOUR
JSR VLINE
The program was then converted to machine lan guage. In the instant case, hexadecimal notation was chosen. The hexadecimal format of the above excerpt is [at page 8 C.P.R.]:
AO 2F
DO 02
AO 27
84 2D
A9 00
85 30
20 28 F8
The learned Trial Judge found, at page 9 C.P.R., that:
In fact, it is the binary form of machine language which the computer "understands". The string of ones and zeros can be
represented in the computer as the two states "electrically on" and "electrically off' (or high and low voltage levels).
She found, as an example [at pages 8 and 9 C.P.R.], that:
... the assembly code instruction LDY set out above, which became in hexadecimal notation A0, would in binary become 1010 0000.
In order for a program to be used by a computer it must, then, be converted into a series of "electrically on" and "electri- cally off' states, an electrical code which corresponds to the binary notation of the program. This can be accomplished in a variety of ways since a program once written can be embodied in a variety of material forms. All that is required is that the material be capable of reflecting two discrete states in order to correspond to the ones and zeros of the binary code.
The learned Trial Judge then discussed a number of ways in which the conversion to an electrical code may be achieved, e.g., punch tape or cards where a hole represents one and its absence repre sents zero, or vice versa; magnetic tape or discs in which the presence of a magnetic charge repre sents one and its absence represents zero, or vice versa. All can be machine read by sensors detect ing the presence or absence of light or magnetic charge. She went on, at page 10 C.P.R.:
The evidence was that conceptually the internal memory of the computer could be thought of as a series of mail-boxes (pigeon holes) arranged in a rectangular array into which information in the form of bytes is stored (eight ones and zeros per box). At the electrical level, the computer memory is an integrated circuit which is capable of holding, because of its circuitry, a pattern of high and low voltage states. That electri cal code will correspond on a one-to-one basis with the program as written in binary notation:
So you can imagine 8 transistors, then, being all arranged in a row, and those transistors somehow would be employed in the construction of the device to exhibit the information that is to be stored ...
(Transcript Vol. I, page 49.)
The internal memory of the computer is of two types, RAM and ROM. RAM (random access memory) is volatile. When the power is turned off whatever has been placed in RAM. is erased .... ROM (read only memory), on the other hand, is permanent in nature and whatever is stored therein is not erased when the power is turned off. This follows from the fact that while the eight transistors in RAM have the property of switching on or off (i.e., of representing either a one or a zero depending upon what program or data is loaded into it), the transistors of the memory cells of the ROM chips have been "blown" and thus permanently structured to carry only one set
of instructions. Thus ROM can be "read" only; it cannot have information written into it as is the case with RAM.
The chips in issue are ROM chips. The learned Trial Judge continued, at pages 11 and 12 C.P.R.:
When a programmer or a computer manufacturer wishes to have a program embodied in ROM, he or she sends a copy of the program in machine language (hexadecimal or binary) to a ROM manufacturer. The manufacturer of the Apple II ROMs also required that it be submitted in at least two different input media (e.g., punched paper tape and a floppy disk) to facilitate verification. A pattern is then etched into the memory cells of the chip by a photo-lithographic process which it is not neces sary to describe in detail for the purpose of these reasons. The pattern created in the ROM chip corresponds to the binary pattern of the written program. Thus, the chip is structured so that it can replicate the program or part thereof in its electrical code form as required whenever an electric current is passed through the appropriate [mailbox] of the ROM.
After dealing with other computer components and rejecting the argument that the programs in issue were merely specifications for a machine part, the learned Trial Judge found, at page 16 C.P.R., that:
... the program retains its character as such in the computer. It can be "read out" of the ROM as and when required. A "disassembler" program is often used for this purpose. Thus the hexadecimal code version of the programs Autostart ROM and Applesoft can easily be displayed on the screen of the monitor or produced in a print-out version by "reading" such directly from the respective chips. In addition this hexadecimal form of the program can be converted subsequently to its original source code assembly language version without great difficulty.
Finally, as to the facts, while substantial copying was admitted, it is worth noting that the programs were not copied in their written forms; rather they were copied directly from the chips by a process which produced substantially identical chips.
THE ISSUES
The appellants object to the judgment below on the following grounds:
1. That the rejection of the submission that the programs in machine language merely constituted specifications for the manufacture of the chips was contrary to all the evidence and to an earlier finding of fact by the learned Trial Judge.
2. That the learned Trial Judge erred in law in holding that machine language, whether binary or hexadecimal, is a "translation" of assembly lan guage within paragraph 3(1)(a).
3. That the learned Trial Judge erred in law in holding that the programs embodied in the chips are reproductions of the written computer pro grams within the meaning of subsection 3(1) of the Copyright Act and, in particular, in holding that subsection 3(1) does not require a reproduction to be in humanly readable form.
4. That the learned Trial Judge erred in law in holding that the appellants' chips were infringing contrivances within paragraph 3(1)(d).
1. MACHINE CODE ONLY A SPECIFICATION FOR MAKING CHIP
I really do not see that this objection advances the appellants' cause. In fact, I do not see that the submission was rejected.
The previous contrary finding relied upon is that which I have recited from page 11 of the reported judgment, which describes how a copy of a pro gram in machine language is used in the manufac ture of a ROM chip. I have been unable to locate, in the judgment below, a finding that the machine language version of a computer program serves any purpose other than to facilitate a physical embodiment of the program in a chip, disc, tape or whatever. That is essential to the program's utili zation in a computing machine, which, as I under-
stand it, was the only purpose of writing the program in the first place.
The issue as precisely defined by the learned
Trial Judge, at pages 180 F.C.; 19 C.P.R., correct ly in my view, is:
... whether a computer program which originates as a written text, in the normal and usual sense of those terms, but which has a dimension (as appears from the facts set out above) which it is not traditional to associate with such texts, continues to be covered by copyright when it is ... embodied in a device designed to replicate that code.
The relevance of this objection to that issue escapes me. It is not copyright in the machine language version of the programs that is said to have been infringed. The appellants did not make their chips according to the instructions of the machine language. This is not analogous to Mrs. Beaton's recipe for rabbit pie.
2. MACHINE LANGUAGE NOT A TRANSLATION
The learned Trial Judge, correctly, dismissed as irrelevant the fact that in the jargon of computer programming conversion of a program from one computer language or code to another is termed "translation". She did, however, find that the con version of the programs in issue from the assembly language in which they were originally written to the hexadecimal machine language was translation within the contemplation of paragraph 3(1)(a).
3. (1) For the purposes of this Act, "copyright" means the sole right to produce or reproduce the work or any substantial part thereof in any material form whatever ... and includes the sole right
(a) to produce, reproduce, perform or publish any translation of the work;
In so finding, she was attracted by the analogy of the conversion of a text into Morse Code or short hand and the dictionary definition of "transla- tion". In the latter behalf, she said, at pages 181 F.C.; 20 C.P.R.:
... I note that "translation" in The Concise Oxford Dictionary (6th ed., 1976) is defined as:
Express the sense of (word, sentence, speech, book, poem, ...) in or into another language; in or to another form of representation .... [Underlining added.]
The conversion from one code to another clearly falls within that definition.
It is axiomatic that copyright exists in the expression of the idea, not in the idea expressed. The principle was stated by Thorson P., in Moreau, Alfred v. St. Vincent, Roland, [1950] Ex.C.R. 198, at page 203.
It is, I think, an elementary principle of copyright law that an author has no copyright in ideas but only in his expression of them. The law of copyright does not give him any monopoly in the use of the ideas with which he deals or any property in them, even if they are original. His copyright is confined to the literary work in which he has expressed them. The ideas are public property, the literary work is his own. Every one may freely adopt and use the ideas but no one may copy his literary work without his consent.
That statement was expressly approved by the Supreme Court of Canada in Cuisenaire v. South West Imports Limited, [1969] S.C.R. 208, at pages 211 ff.
The sense of a word, sentence, speech, book or poem, to use the examples of the foregoing diction ary definition, is the idea expressed by it. It follows that, for purposes of the Copyright Act, "transla- tion" cannot be the expression of that sense in "another form of representation".
As to the analogies, the learned Trial Judge continued, at pages 182 F.C.; 20 and 21 C.P.R.
An analogy can be drawn to the conversion of a text into morse code. If a person were to sit down and convert a text into the series of dots and dashes of which morse code is comprised, one might argue that the resultant notations were really instructions to the telegraph operator on how to send the message. But the message written in morse code, in my view, still retains the character of the original work. It is not a different literary work. Similarly, a text written in shorthand might be said to constitute a description of the oral sounds of the text if it were spoken aloud (shorthand being phonetically based), but that would not make it a different literary work from the long hand version.
I agree with the learned Trial Judge that the conversion of a text into Morse Code or shorthand does not result in a different literary work and that the text, so converted, does retain the character of the original. That, however, does not lead to the conclusion that such conversion is translation for purposes of the Act. A person knowledgeable of Morse Code or the particular shorthand system could read the converted version and what would be heard would be the original text verbatim. Such a conversion is not, in my opinion a translation within the contemplation of the Copyright Act. It is rather a reproduction of the original, the making of which was equally the exclusive right of the owner of the copyright in that original.
Success in this objection does not advance the appellants' cause.
3. CHIPS NOT REPRODUCTIONS
The argument here is that the programs embod ied in the computer chips are not reproductions at all of any literary work in which the respondents' copyright subsists and that, in the alternative, even if they are reproductions in an ordinarily sense, they are not reproductions in the context of the Copyright Act because they are not in humanly readable form. The basic argument appears to be founded largely on the judgments of a majority of the High Court of Australia in Computer Edge Pty Ltd y Apple Computer Inc (1986), 65 ALR 33, a decision published subsequent to the judg ment under appeal. My brother MacGuigan has distinguished that decision in his reasons for judg ment herein and I am in full agreement with him on that point. While I do not agree that embodi ment of the programs in the ROM chips can, within the contemplation of the Copyright Act, be considered translations, they can be considered reproductions of the original programs and, if they are, the result is the same: the exclusive right to reproduce, as to translate, was the respondents'.
It is clear that the learned Trial Judge approached this case on the basis that the repro duction in which copyright was claimed and infringement alleged consisted of the static circuit ry of the chips, not the dynamic sequence of electrical impulses dictated or permitted by that circuitry. I find logical suppport for that approach in the reasons herein of my brother MacGuigan who points out, albeit in another context, that the original programs for which copyright registration was obtained were "programs at rest, as it were, not in operation" and that "the activity which occurs in an operating computer originates from the [central processing unit], not from the memory chips, and by looking into the latter one cannot see the operation of the CPU". I think the learned Trial Judge was correct in that approach and, with respect, I am of the view that the distinction has legal significance. While I do not have to pursue the matter further, I would observe that, had the claimed subject of copyright been found to extend to or to be the dynamic sequence of electrical impulses, there might be a very real issue as to whether copyright was claimed only in the recipe or in the rabbit pie itself.
Section 3 of the Copyright Act accords to the copyright owner "the sole right ... to reproduce the work ... in any material form whatever". The learned Trial Judge reviewed pertinent authorities and the legislative history of that language at pages 190 ff. F.C.; 27 ff. C.P.R. of her reasons. I see no useful purpose in repeating that exercise. Suffice it to say, she discussed Boosey v. Whight, [1899] 1 Ch. 836, affirmed [1900] 1 Ch. 122 (C.A.), which found that perforated piano rolls could not be found to be infringing reproductions of the sheet music from which they were derived under the existing Imperial copyright law and the remedial amendments of the 1911 Imperial Act [Copyright Act, 1911, 1 & 2 Geo. V, c. 46 (U.K.)], which were imported into sections 3 and 4 of the present Canadian Act in, for all practical purposes, identical verbiage. I am in complete
agreement with her conclusion, at pages 193 F.C.; 29 C.P.R., that:
... the opening words of section 1(2), now section 3 of the Act, were purposely drafted broadly enough to encompass new technologies which had not been thought of when the Act was drafted.
and, at pages 194 F.C.; 30 C.P.R., that:
the opening words of section 3 clearly cover the plaintiffs' program as embodied in a ROM chip. Such embodiment is surely the production or reproduction of the work in a material form....
Nothing in the Act requires expressly that a reproduction be in humanly readable form to be subject of copyright or to infringe copyright. To the extent the matter has arisen in the jurispru dence since the advent of the 1911 Imperial Act, it appears to have been dealt with in an evidentiary context. Like the learned Trial Judge, I am much impressed by the observation of Megarry, V.C., in Thrustcode Ltd. v. W.W. Computing Ltd., [1983] F.S.R. 502 (Ch.D.), at page 505:
For computers, as for other things, what must be compared are the thing said to have been copied and the thing said to be an infringing copy. If these two things are invisible, then normally they must be reproduced in visible form, or in a form that in some way is perceptible, before it can be determined whether one infringes the other.
In summary, in my opinion, the respondents' computer programs, entitled Autostart ROM and Applesoft, embodied in their ROM chips were reproductions of the computer programs written in assembly language in which their copyright is admitted to have subsisted. The appellants copied those reproductions and they thereby infringed the respondents' copyright in the programs.
4. CHIPS INFRINGING CONTRIVANCES
The learned Trial Judge held, at pages 198 F.C.; 34 C.P.R.:
In addition, I think there is merit in the argument that the ROM chip, whatever the interpretation of the opening words of section 3, can be said to fall within paragraph 3(1)(d) as a
contrivance by means of which the work may be delivered. The program (as noted above) can be delivered to the screen of the monitor, or as a print-out to be read by human beings. I do not find it necessary to decide whether "delivery" to the CPU satisfies paragraph 3(1)(d).
With respect, I am unable to construe that as finding it unnecessary to decide whether the ROM chips fell within paragraph 3(1)(d). I should have thought it to be a finding that they did, since the programs embodied therein can be delivered by print-out or display, reserving only the question whether delivery to the CPU is sufficient to bring them within the paragraph. That said, I do agree with my brother MacGuigan that it is not neces sary to decide that issue in view of the disposition I propose on the third ground of objection.
CONCLUSION
Pursuant to subparagraph 52(b)(î) of the Fed eral Court Act [R.S.C. 1970 (2nd Supp.), c. 10], I would vary the judgments of the Trial Division by awarding the remedies granted to Apple Comput er, Inc., only and I would otherwise dismiss the appeals with costs.
* * *
The following are the reasons for judgment rendered in English by
HUGESSEN J.: The appellants have, in popular parlance, pirated two computer programs used by the respondents in their Apple II+ computer. It is common ground that the programs in question, when written, as they originally were, in letters, symbols and figures known as 6502 assembly code, were original literary works subject to copyright and that the copyright therein vests in respondent Apple Computer, Inc. But the appellants have not copied the programs written in assembly code. They have not copied any writing at all. What they have done is reproduce, apparently by mechanical means, the electrical circuitry of a silicon chip in which, by the magic of computer science, the programs are embodied. The question at issue is to know whether the appellants have infringed the
respondents' copyright in the original assembly code programs.
The question was answered by Reed J., of the Trial Division, in the affirmative. Her reasons for judgment are both full and thorough. I am in general agreement with her conclusions but, with due respect for the scholarship of the reasons, find that I have arrived there by a somewhat different route.
The principal difficulty which this case has given me arises from the anthropomorphic charac ter of virtually everything that is thought or said or written about computers. Words like "language", "memory", "understand", "instruction", "read", "write", "command", and many others are in con stant use. They are words which, in their primary meaning, have reference to cognitive beings. Com puters are not cognitive. The metaphors and analo gies which we use to describe their functions remain just that.
A computer is a highly complex miniaturized interconnected collection of electrical circuits. The genius of the invention flows from the fact that since an electrical circuit can only have two states, "on" and "off", and since those states can be made to correspond to the figures 0 and 1 of binary notation, any mathematical function and any code which is capable of binary expression can be per formed and stored. The only limitation is a physi cal one, namely, the number of separate circuits which can be connected together and accom modated within the machine.
A computer program therefore, whatever be its original form and whatever its purpose, must in the final analysis be capable of being reduced to a code which can be expressed in binary notation. In simple terms, when one sits at the keyboard of a computer and strikes the letter "a", one does not, as with a typewriter, mechanically activate (with or without electrical assistance) a bar of type;
rather one engages a series of electrical circuits whose condition corresponds to whatever binary number has been arbitrarily assigned to the letter "a" and which, in their turn, will result ultimately in "a" being displayed on the monitor or printed.
The programs here in issue, called respectively Autostart ROM and Applesoft, are what is known as operating system programs. Their role is largely to control and manage the internal operations of the computer itself. It is this fact which explains the appellants' interest in copying them precisely. There is a virtually unlimited number of ways in which both appellants' and respondents' computers can be programmed to carry out the functions which are the subject matter of the programs here in issue. But the purchaser of appellants' comput ers will be unable to use most of the voluminous software available and specifically designed for use with the respondents' machines unless the operat ing system programs are substantially identical.
The two programs were, as I have indicated, originally written in assembly code. Such a code normally bears a discernible relationship to common language by the use of mnemonics. In an example given by the Trial Judge, the assembly code indication LDY stands for "load index y with memory". The programs were then converted from assembly code into machine code. Since this is a code which can be "read" by the machine, it must, as explained above, be expressed in binary notation.
The conversion from the ordinary language description of the function into assembly code and from the latter into machine code are purely arbi trary functions of much the same order as the conversion of the alphabet into Morse Code: one starts with the source and a code key and by applying the latter to the former a coded version is produced.
By contrast a further conversion is often effect ed for convenience purposes: the machine code is transposed from binary to hexadecimal notation (base 16). Here the conversion is purely arithmetic from one system of notation to the other. It is made simply because it is easier and less tedious for humans to read and write hexadecimal nota tion. It is in their binary form, however, that the programs were and are transformed into a series of on-and-off electrical circuits embodied in a silicon chip. The issue then may be restated as being to know whether, by copying the respondent's chips, the appellants infringed the copyright in the two programs.
Copyright is a statutory right. It exists only in virtue of the Copyright Act.' As was stated by Estey J. for the Court in Compo Company Ltd. v. Blue Crest Music Inc. et al., [1980] 1 S.C.R. 357, at pages 372 and 373:
... copyright law is neither tort law nor property law in classification, but is statutory law. It neither cuts across exist ing rights in property or conduct nor falls between rights and obligations heretofore existing in the common law. Copyright legislation simply creates rights and obligations upon the terms and in the circumstances set out in the statute.
The rights of the owner of a copyright are compendiously set out in subsection 3(1) of the Copyright Act.
3. (1) For the purposes of this Act, "copyright" means the sole right to produce or reproduce the work or any substantial part thereof in any material form whatever, to perform, or in the case of a lecture to deliver, the work or any substantial part thereof in public; if the work is unpublished, to publish the work or any substantial part thereof; and includes the sole right
(a) to produce, reproduce, perform or publish any transla tion of the work;
(b) in the case of a dramatic work, to convert it into a novel or other non-dramatic work;
(c) in the case of a novel or other non-dramatic work, or of an artistic work, to convert it into a dramatic work, by way of performance in public or otherwise;
(d) in the case of a literary, dramatic, or musical work, to make any record, perforated roll, cinematograph film, or
' R.S.C. 1970, c. C-30.
other contrivance by means of which the work may be mechanically performed or delivered;
(e) in the case of any literary, dramatic, musical or artistic work, to reproduce, adapt and publicly present such work by cinematograph, if the author has given such work an original character; but if such original character is absent the cinematographic production shall be protected as a photograph;
(J) in case of any literary, dramatic, musical or artistic work, to communicate such work by radio communication;
and to authorize any such acts as aforesaid.
The learned Trial Judge found in this text sever al alternate routes to her conclusion that the appel lants' chips constituted an infringement of the respondents' copyright in the assembly code pro grams. All of them give me difficulty although some to a greater extent than others.
In the first place, I cannot accept, as the Trial Judge seems to have done, that the appellants' chips were a "translation" of the respondents' programs, contrary to paragraph 3(1)(a). In my view, "translation" is used here in its primary sense of the turning of something from one human language into another. To give it its extended meaning of an expression in another medium or form of representation seems to me to be at vari ance with the fundamental principle that copyright subsists not in the idea expressed but in the form of its expression. That principle in turn finds application in the well-known rule that a transla tion is itself an original literary work and the subject of copyright, and this regardless of wheth er the work translated is itself subject to copyright. The rendering of a literary work in code, for example, Morse or Braille, is, I think, properly characterized as a reproduction of the work, not as a translation. Indeed, since the respondents' pro grams originated in code, I do not see how one can properly speak of their translation at all. The fact that both machine and assembly codes are called "languages" is simply an example of the anthropomorphic phenomenon to which I referred at the outset.
Secondly, the Trial Judge seems to view with favour the argument that the appellants' chips were contrivances by means of which the work may be mechanically performed or delivered within the meaning of paragraph 3(1)(d). Here again I cannot follow her. The copyright work, the assembly code version of the programs, is a liter ary work. "Performance" is defined in section 2 of the Act to mean:
2....
"performance" means any acoustic representation of a work or any visual representation of any dramatic action in a work, including a representation made by means of any mechanical instrument or by radio communication.
There is no indication that the subject programs can be acoustically represented nor that they con tain any dramatic action; indeed the contrary seems almost self-evident. With respect to the possibility that the chips may be said to be capable of "delivery" of the programs, the Trial Judge, rightly, did not view herself as being bound by the inclusive definition of "delivery" in section 2 2 to hold that only a lecture could be "delivered". In the context of paragraph 3(1)(d), however, as well as of the definition of "delivery", I cannot escape the conclusion that the word "delivered" is used here solely with reference to communication by audible means. To give "delivery" its more common meaning of handing over would lead to the somewhat startling conclusion that the trolley (a "contrivance") used to deliver books from the library to my office could only be made by the owner of the copyright in the books.
Finally, and it is fair to say that this was the principal foundation for her decision, the Trial Judge held that the chips embodying the programs were reproductions of the assembly code programs in which the respondents held the copyright. As I read this part of her judgment, it seems to me to turn primarily on the use of the words "in any material form whatever" in the introductory part
2 "delivery," in relation to a lecture, includes delivery by means of any mechanical instrument.
of subsection 3(1). I agree that the appellants' chips are such a reproduction but I find it neces sary to expand somewhat on the process by which I arrive at that conclusion.
At first blush, it would seem obvious that when one produces or reproduces a work the result of that operation must necessarily also be a work. The only kinds of works in which copyright can subsist in accordance with section 4 of the Act are literary, dramatic, musical and artistic works. The only category into which the respondents' pro grams in assembly code could possibly fit is that of literary works and it is indeed common ground that they are such works.
The sole distinguishing characteristic of a liter ary work is not its quality as literature or art but simply that it be in print or writing. As early as 1916, within five years of the passage of the U.K. Copyright Act, 1911, upon which our act is mod eled, it was said in University of London Press v. University Tutorial Press, [1916] 2 Ch. 601:
Although a literary work is not defined in the Act, s. 35 [in Canada, s. 2] states what the phrase includes; the definition is not a completely comprehensive one, but the section is intended to show what, amongst other things, is included in the descrip tion "literary work," and the words are " `Literary work' includes maps, charts, plans, tables, and compilations." It may be difficult to define "literary work" as used in this Act, but it seems to be plain that it is not confined to "literary work" in the sense in which that phrase is applied, for instance, to Meredith's novels and the writings of Robert Louis Stevenson. In speaking of such writings as literary works, one thinks of the quality, the style, and the literary finish which they exhibit. Under the Act of 1842, which protected "books," many things which had no pretensions to literary style acquired copyright; for example, a list of registered bills of sale, a list of foxhounds and hunting days, and trade catalogues; and I see no ground for coming to the conclusion that the present Act was intended to curtail the rights of authors. In my view the words "literary work" cover work which is expressed in print or writing, irrespective of the question whether the quality or style is high. The word "literary" seems to be used in a sense somewhat similar to the use of the word "literature" in political or electioneering literature and refers to written or printed matter. Papers set by examiners are, in my opinion, "literary work" within the meaning of the present Act. (Per Peterson J., at page 608.) That statement found approval in the House of Lords in Ladbroke (Football) Ltd. v. William Hill (Football) Ltd., [1964] 1 W.L.R. 273.
It seems to me to follow from the foregoing that when a literary work is produced or reproduced the result must necessarily be a literary work, i.e., expressed in print or writing.
The difficulty which this raises in the present case is obvious. The respondents' copyright is in the assembly code version of the programs. Clearly that version is expressed in print or writing. When the program is reproduced in either binary or hexadecimal notation, the result being in figures and letters is also in print or writing and hence a literary work. When, however, the binary code version is transposed into open-and-closed electri cal circuits embodied in a silicon chip, the latter contains no print or writing and hence cannot be said to be a literary work.
Does this mean that the production of the chip embodying programs does not constitute an infringement of the copyright? I think not. We have already seen that the statute defines copy right as being, amongst other things, the sole right to produce or reproduce in any material form. It is, in my opinion, possible to read those words as including by necessary implication the sole right to produce the means of reproduction of the work or, to put the matter another way, the sole right to produce anything used or intended to be used to reproduce the work. When the opening words of subsection 3(1) are read in the context of the remainder of that subsection and of other sections of the Copyright Act, it is my view that such interpretation is not only possible but is required.
I find section 21 of the Act particularly instruc tive in this regard:
21. All infringing copies of any work in which copyright subsists, or of any substantial part thereof, and all plates used or intended to be used for the production of such infringing copies, shall be deemed to be the property of the owner of the copyright, who accordingly may take proceedings for the recov ery of the possession thereof or in respect of the conversion thereof.
The owner of the copyright is thus deemed to be the owner not only of all infringing copies but also of all plates used or intended to be used for their production. As such owner he is given the right to take proceedings to recover their possession or to claim for their conversion. Surely the only theoretical justification for this provision must be that the making of the plates, like the making of the infringing copies themselves, constitutes the doing of something which is the sole right of the owner of the copyright.
The definition of "plate" in section 2 is also helpful:
2....
"plate" includes any stereotype or other plate, stone, block, mould, matrix, transfer, or negative used or intended to be used for printing or reproducing copies of any work, and any matrix or other appliance by which records, perforated rolls, or other contrivances for the acoustic representation of the work, are or are intended to be made.
The Trial Judge found as a fact that the appel lants' chip could be used to produce the hexadeci mal version of the respondents' programs on the monitor or in the printer of a computer. Whether or not the chip falls within the very extended definition of "plate", that definition itself serves to strengthen the argument that the sole right to produce or reproduce a work includes the sole right to produce the means of reproduction.
Reference may also be made to section 10 of the Act:
10. The term for which copyright shall subsist in records, perforated rolls and other contrivances by means of which sounds may be mechanically reproduced shall be fifty years from the making of the original plate from which the contriv ance was directly or indirectly derived, and the person who was the owner of such original plate at the time when such plate was made shall be deemed to be the author of such contrivance, and where such owner is a body corporate, the body corporate shall be deemed for the purposes of this Act to reside within Her Majesty's Realms and Territories if it has established a place of business therein.
Thus, the term of copyright in a contrivance dates from the making of the "original plate" and the owner of such plate is deemed the author of the contrivance in which copyright subsists. Here
again the only theoretical justification for this provision is that the making of the original plate (which, it may be noted, is distinguished from the contrivance itself) is something which is the sole right of the owner of the copyright, if any, in the work which may be mechanically performed or delivered by the contrivance. This was made clear in Compo Company Ltd. v. Blue Crest Music Inc. et al., supra, where the appellant Compo had produced records using a matrix which had been supplied to it by Canusa which the latter had, in turn, had made without the authorization of the owner of the musical work. Estey J. said, at page 374:
Thus Canusa, by making or causing a third party to make the matrix from which the stamper and thereafter the records were made, might, if authorised by the owner of the recorded musical work, be the owner of the copyright in the record. However, the recording of the work was not authorised. Compo, of course, as the agency which, using the master acetate (which I take to be in these circumstances the plate or matrix described in ss. 10 and 2 of the Act) actually stamped out the records, would in law have obtained no copyright in the record whether or not the owner of the copyright in the musical work had granted a licence to record that work. The ownership of the copyright in the record would, had the recording been authorised, reside in the owner of the matrix at the time of its making, here Canusa. [Emphasis added.]
I also find some help in the text of paragraph 17(2)(b) of the Act:
17....
(2) The following acts do not constitute an infringement of copyright:
(b) where the author of an artistic work is not the owner of the copyright therein, the use by the author of any mould, cast, sketch, plan, model, or study made by him for the purpose of the work, if he does not thereby repeat or imitate the main design of that work.
Here again, as with so many other provisions of the Act, we appear to be dealing with an ad hoc text designed to meet some particular perceived problem. Nonetheless the provision is only expli cable if one is of the view that absent the saving words the means for the reproduction of a work belongs solely to the owner of the copyright in the work and may be used only by him.
Finally I refer in this connection to the closing words of subsection 3(1) following the enumera tion in paragraphs (a) to (f). These words make it plain that only the owner of the copyright may authorize any of the acts listed in the subsection. Anyone who makes and sells to the public any thing designed or intended to reproduce a copy righted work gives an implied authorization to the purchaser to effect such reproduction. Here, as found by the Trial Judge, the ROM chips in the appellants' computer were capable of being used to reproduce the respondents' programs in their hex adecimal code version; in those circumstances, even if, contrary to what I have suggested, the means of reproduction does not itself constitute a reproduction within the meaning of subsection 3(1), the implied authorization to reproduce is an infringement of the copyright.
I would dispose of the appeal as proposed by Mr. Justice Mahoney.
* * *
The following are the reasons for judgment rendered in English by
MACGUIGAN J.: These cases, which were argued together, are concerned with copyright in computer programs as embodied in silicon chips.
It is admitted that the respondents hold a regis tered copyright in two programs, Autostart ROM (reg. no. 319465) and Applesoft (reg. no. 319468), both registered, on October 8, 1982 and that these programs, as registered, are copyrightable literary works. It is also admitted that the appellant copied the two programs. It is, however, argued that such programs encoded within silicon chips cannot be the subject matter of copyright. The issue, then, is as to the existence of copyright in computer pro grams in that state.
A computer is a complex system of interconnected, integrated electrical circuits. It consists of a circuit
board onto which have been pinned or soldered a number of electronic components, which communi cate by means of traces (called "buses" or "wires") etched into the board. The main electron ic components of the system are the input/output devices, the microprocessor or central processing unit (CPU) and the memory.
The memory is now usually provided by memory chips, approximately one quarter inch square pieces of silicon. The ROM ("read only memory") type of memory, which is in question in the cases at bar, is permanently structured, through having been permanently etched with electric circuits, to carry a particular program, which can subsequent ly be read (by way of contrast with RAM or random access memory, which is erased when the power is turned off). ROM circuits consist of interconnected transistors built in and of the sili con. Their pattern cannot be discerned by the human eye except with the aid of an electron microscope. Since electricity can have only two states, "on" and "off'', the transistors are operated as switches in these two states. (All computer circuits in fact carry out arithmetic and logical functions based only on binary arithmetic and Boolean algebra.)
Since the circuitry of computers is thus limited, the only form of program a computer can utilize is one written in binary notation (using the two digits "1" and "0"), or some derivation of that such as hexadecimal notation (using the digits "0-9" and "A-F").
In these cases the programs in question were written in assembly language to be encoded on chips to be mounted in the respondents' Apple II+ computer as the operating instructions for that computer. It was these assembly codes which were registered for copyright on October 8, 1982, by reference to their titles.
Assembly codes may be said to be an intermedi ate level of language. A higher-level language would have corresponded to ordinary mathematics or ordinary English. In order to be encoded in a chip, however, the assembly language had to be transposed into a still lower level of language, machine language or object code (here hexadeci mal notation).
The programs as registered for copyright were machine specific, that is, they were in 6502 assembly language for use with a 6502 CPU, which also consists of a silicon chip. (Each of the dozen or more kinds of CPU's commercially avail able contain different circuitry and so the codes to which each type responds are different.) Autostart ROM and Applesoft are both operating system programs, the former consisting of initiating procedures, the latter of procedures allowing for user-communication with the computer in a high- level or easily read language.
* * *
The Copyright Act, R.S.C. 1970, c. 30 ("the Act") provides in subsection 4(1) that "Subject to this Act, copyright shall subsist in Canada ... in every original literary, dramatic, musical and artistic work ..." The part of the Act principally in ques tion is subsection 3(1), which reads as follows:
3. (1) For the purposes of this Act, "copyright" means the sole right to produce or reproduce the work or any substantial part thereof in any material form whatever, to perform, or in the case of a lecture to deliver, the work or any substantial part thereof in public; if the work is unpublished, to publish the work or any substantial part thereof; and includes the sole right
(a) to produce, reproduce, perform or publish any transla tion of the work;
(d) in the ease of a literary, dramatic, or musical work, to make any record, perforated roll, cinematograph film, or other contrivance by means of which the work may be mechanically performed or delivered;
In an admirably reasoned decision reported at [1987] 1 F.C. 173; 10 C.P.R. (3d) 1, Madam Justice Reed held at trial that the respondents' computer programs as encoded were covered by copyright. She set out the issue as follows, at pages 180 F.C.; 19 and 20 C.P.R.:
The issue is whether a computer program which originates as a written text, in the normal and usual sense of those terms, but which has a dimension which it is not traditional to associate with such texts, continues to be covered by copyright when it is converted into its electrical code version, or more precisely in this case when it is embodied in a device designed to replicate that code.
The defendants' argument that copyright protection does not so extend, has several facets: (1) the hexadecimal code version of the program is not a translation of the source code version; (2) since there is a one to one relationship between the source code program and its embodiment in the chip there is a merger of the idea and the expression of the idea which copyright law does not cover; (3) the text of the Copyright Act [R.S.C. 1970, c. C-30] does not cover computer programs in their chip form; (4) there are compelling policy arguments for refusing to extend copyright protection in the present case such as poten tial restrictions on commerce and a possible overlap with patent law.
She then proceeded to deal with each of the defendants' (appellants') arguments in turn. On the argument as to translation she held, at pages 181-182 F.C.; 20 and 21 C.P.R.:
I do not find this argument convincing. In the first place I note that "translation" in The Concise Oxford Dictionary (6th ed., 1976) is defined as:
Express the sense of (word, sentence, speech, book, poem, ...) in or into another language; in or to another form of representation .... [Underlining added.]
The conversion from one code to another clearly falls within that definition.
An analogy can be drawn to the conversion of a text into morse code. If a person were to sit down and convert a text into the series of dots and dashes of which morse code is comprised, one might argue that the resultant notations were really instructions to the telegraph operator on how to send the message. But the message written in morse code, in my view, still retains the character of the original work. It is not a different literary work. Similarly, a text written in shorthand might be said to constitute a description of the oral sounds of the text if it were spoken aloud (shorthand being phonetically based), but that would not make it a different literary work from the long hand version.
In my view the conversion of a work into a code, or the conversion of a work originally written in one code into another code constitutes a translation for the purposes of the Act. In addition, as noted above, a programmer, in creating a program, is not thinking in terms of the specifications of the ROM chip, either in writing the assembly code version or when the hex adecimal notation is prepared. What media [sic] is finally chosen for embodiment of the program is irrelevant to the programmer. Accordingly, I find it difficult to accept counsel for the defendants' argument that the hexadecimal form of the program is a different literary work and not a translation of the original.
With respect to the argument that copyright does not extend to computer programs because copyright protects the expression of an idea rather than the idea itself and that a computer program embodied in a ROM chip exhibits a merger of the two, she held at pages 185-190 F.C.; 23-27 C.P.R.:
The program as originally written is copyrightable subject- matter. There has in fact been copying—the code read from the defendants' chips is the same as that which can be read from the plaintiffs' chips .... The computer program when written is clearly a literary work. What is more, its embodiment in a silicon chip retains the form of expression of the original work. The program in its source code version can be retrieved (read) by a process of translation or translations from the ROM chip.
Counsel argues that in copying the ROM the defendants are doing no more than following the recipe prescribed by the program, ie: making Mrs. Beeton's apple pie. I think a closer analogy is that what they are doing is copying the recipe book.
What then of counsel's argument that a computer program exhibits a merger of the idea and the expression of the idea and therefore is not copyrightable. I have considerable difficulty with this argument from a number of points of view. In the first place, if this argument is valid then it is difficult to understand why the concession was made that the assembly code version of the program is copyrightable. Surely if there is a merger of idea and expression, it exists not only in the machine code version of the program but also in the written assembly code version as well. Secondly, the exact scope of the legal rule (if one exists) that when there is a merger of the idea with its expression, copyright protection does not operate, is not at all clear. And, thirdly, the evidence that there is a multitude of forms of expression in which any given program can be written seems to me to demonstrate that there is no merger of the idea and the expression of the idea with respect to the programs in issue.
Examples given in argument of the application of the merger principle are formulations such as E = mc 2 , the pythagorean theorem, or various algebraic proofs. It may be that these
examples demonstrate nothing more than that the particular idea being expressed is really information about the external world (recognizing that E = mc 2 is in fact a hypothesis) and that factual information is something which copyright does not protect. I see an analogy, for example, between a statement such as "the Prime Minister of Canada met with the President of the United States on March 17" and "the square on the hypotenuse of a right-angled triangle is equal to the sum of the squares on the other two sides". Neither of these phrases as such, apart from a larger text in which they might appear, would be subject of copyright. The example of algebraic proofs leads, it seems to me, to an almost philosophical discussion of the nature of intellectual thought. It is a discussion which a higher Court than this may find it necessary to consider but which I consider it sufficient to deal with by noting that a computer program is not similar to any of the examples cited. It is not a statement reporting a fact. It is not comparable to an algebraic proof. It is a creation in the same way that an instruction book is a creation. While copyright would not prevent someone making Mrs. Beeton's rabbit pie (indeed it was to encourage people to do so that Mrs. Beeton's Book of Household Management was first published), it does prevent someone copying the book itself. The order in which the recipes are listed, the form and expression in which they are couched are properly the subject of copyright. This order, form, pattern of expression of the plaintiffs' program is retained in the ROM and is copied when the defendants copy the ROM.
There is another branch of the merger doctrine which has prevailed in the United States and which it is necessary to consider. It would appear to have originated with the decision in Baker v. Selden, 101 U.S. 99 (1879), a case referred to in several "commonwealth" decisions, but without reliance on the full scope of the decision given therein. Baker v. Selden dealt with copyright claimed in a book which described a new system of accounting and particularly accounting forms (consisting of certain ruled lines and headings) contained in the book. The Court found that the defendant's use of the book and develop ment of his own forms did not breach the plaintiff's copy- right—i.e.: there was no substantial copying. But the Court went further and drew a distinction between works of science or instruction and other types of works. It held that where a useful art could only be employed by using the forms or diagrams by which it was explained there would be no copyright in such forms and diagrams.
The Court thus proceeded to lay down a much broader rule than was necessary for the purpose of the case; it held that the forms in the book were not protected by copyright.
This development of copyright law is not one that has been adopted in this country, as far as I know. Indeed rejection thereof would seem to be implicit in the Court of Appeal's decision in Bulman Group (The) Ltd. v. Alpha One-Write Systems B.C. Ltd. et al. (1981), 54 C.P.R. (2d) 179, and in the subsequent decision of the Trial Division Bulman Group Ltd. (The) v. "One Write" Accounting Systems Ltd., [1982] 2 F.C. 327; 62 C.P.R. (2d) 149. At issue in those cases was the copyrightability of accounting forms. Initially the claim for copyright in such forms was rejected by the Trial Division on
an application for an interlocutory injunction. The Court of Appeal overturned that decision indicating that there was a seriously arguable case. Following full trial of the issue copy right was held to exist (this decision was not appealed to the Court of Appeal).
In any event, I have not been persuaded that there is a merger of the idea and the expression thereof in a computer program. The fact that a program can be written in a variety of different forms, that the same programmer would not write a program the same way if he or she were to start anew a second time, that the programmer is indifferent to the medium in which the program is embodied, all indicate that computer programs do not fall within the merger exception to copyright- able subject matter (if such exception exists).
On the statutory argument she held, at pages 194-198 F.C.; 30-34 C.P.R.:
In my view the opening words of section 3 clearly cover the plaintiffs' program as embodied in a ROM chip. Such embodi ment is surely the production or reproduction of the work in a material form, just as a record or a cassette tape is a production or reproduction of a work in a material form. (I have not overlooked the fact that there is a separate section in the Act which covers copyright in records.)
In my view the requirement of "readability" or "appearance to the eye" found in the jurisprudence requires no more than that there be a method by which the work in which copyright is claimed and the work which is alleged to infringe can be visually compared for the purpose of determining whether copying has occurred. Since in this case the programs can be "read" out of ROM and so compared, this requirement is met. I adopt in this regard the reasoning of Mr. Justice Megarry in Thrustcode Ltd. v. W.W. Computing Ltd., [1983] F.S.R. 502 (Ch.D.), at page 505:
For computers, as for other things, what must be compared are the thing said to have been copied and the thing said to be an infringing copy. If these two things are invisible, then normally they must be reproduced in visible form, or in a form that in some way is perceptible, before it can be determined whether one infringes the other.
Normally ... what will be needed is a print-out or other documentary evidence of the program alleged to have been copied, and of the alleged infringing program, or sufficient parts of each.
As I understand counsel's argument it is that even if the reproduction itself does not have to be in human readable form there is a requirement that the purpose for which that repro duction is made must ultimately be to communicate the work to human beings. A record or cassette when used with a machine produces sounds for human listening while the ROM chip does not so communicate to humans as its primary function. As noted above, it can be used to so communicate and for certain limited purposes is so used, but that is not its primary purpose.
In my view counsel's arguments based on the context of the Act and on the specific wording of the sections thereof are conclusively answered by section 3 itself. Section 3 provides that "`copyright' means the sole right to produce or reproduce the work ... in any material form whatever". In my view that clearly covers the program as embodied in the ROM chip. To find otherwise, it seems to me, would require reading words into section 3 of the Copyright Act which are not there.
The argument based on paragraph 3(1)(d) I find unconvinc ing. It has a two-edged quality. It can equally be asked why, if Parliament only intended to cover records or contrivances which utter sounds or communicate to humans, the opening words of section 3 were drafted so broadly. The references to the definition sections are not persuasive. They only purport to define one particular aspect of delivery as it relates to lectures; no all inclusive definition is given of that word. The definition of performance is irrelevant since no one has suggested that it applies in the present case. Also, I have difficulty with the distinction counsel makes between programs which result in something being displayed on the screen and those (such as ROM) which do not. When "something" is displayed on the screen it is not the program (i.e.: it is not the original or a hexadecimal version thereof which is displayed). The program remains invisible unless disassembled by the translation pro cesses similar to that used to "read" the ROM.
On the fourth argument, she found no compel ling policy reasons for refusing to extend copyright protection in this case, at pages 200 and 201 F.C.; 35 and 36 C.P.R.:
With respect to the first argument, the purpose of the Copyright Act is and always has been to grant a monopoly. No distinction is made therein as to the purpose of the work created—for entertainment, instruction or other purposes. The legislation historically, in my view had two purposes: to encour age disclosure of works for the "advancement of learning", and to protect and reward the intellectual effort of the author (for a limited period of time) in the work. A book is an article of commerce, as is a map or a chart. The interpretation of the legislation which the defendants urge, based on a view that the Act was not intended to interfere with commerce, is both not accurate and would add a gloss to the statute which its wording does not bear.
With respect to the second argument, as I read the authori ties and references to which I have been referred by counsel they indicate that computer programs are not per se patentable but that an apparatus or process that meets the standards of novelty and unobviousness required by the Patent Act, R.S.C. 1970, c. P-4, will not be disqualified from patent protection merely because a computer is used to operate the apparatus or implement the process. In addition, I do not see that overlap ping areas of the law are particularly unusual (reference need only be made to the fields of tort and contract). Nor do I see it as the role of the courts to frame decisions to avoid such results.
In addition, I would note that the present case does not deal with a special purpose program. It concerns the operation of a general purpose computer, although I must admit that I do not, at the moment, see any rationale for distinguishing between the two. I note that Mr. Justice Megarry in the Thrustcode case, supra, page 194, did not think it necessary to draw such a distinction.
Finally, she adverted to the trend both abroad and in Canada to extend copyright protection in such cases, at pages 203 F.C.; 38 C.P.R.:
There are a growing number of cases in jurisdictions where the copyright law is not too dissimilar from our own which have held that computer programs in their machine code version are protected by the copyright law applicable existing in that jurisdiction. (Sega Enterprises Limited v. Richards and Another, [1983] F.S.R. 73 (Ch.D.); Thrustcode Ltd. v. W.W. Computing Ltd., supra, pp. 194 F.C.; 41 C.P.R.; Northern Office Microcomputers (Pty) Ltd. v. Rosenstein, [1982] F.S.R. 124 (S.C. S. Afr.); Apple Computer Inc v Computer Edge Pty Ltd, [(1984), 53 A.L.R. 225 (F.C.)].) Not only is there such a trend abroad, but a similar phenomenon can also be discerned in this country. Apple Computer Inc. v. Computermat Inc. (1983), 1 C.I.P.R. 1 (Ont. H.C); Apple Computer, Inc. v. Minitronics of Canada Ltd. et al. (1985), 7 C.P.R. (3d) 104 (F.C.T.D.), affirmed (1985), 8 C.P.R. (3d) 431 (F.C.A.); Société (La) d'Informatique R.D.G. Inc. v. Dynabec Ltée et al. (1984), 6 C.P.R. (3d) 299 (Que. S.C.); F & I Retail Systems Ltd. v. Thermo Guard Automotive Products Canada Ltd. et al. (unreported decision of the Supreme Court of Ontario of June 26, 1984); Logo Computer Systems Inc. c. 115778 Canada Inc. et al. (unreported decision of the Quebec Superior Court of October 25, 1983); Nintendo of America, Inc. v. Coinex Video Games Inc., [1983] 2 F.C. 189 (C.A.), which deals with the issue in a peripheral way; Spacefile Ltd. v. Smart Computing Systems Ltd. et al. (1983), 75 C.P.R. (2d) 281 (Ont. H.C.).)
The learned Trial Judge was quite right in her analysis of the trend in the case law up to that time, and had it not been for the decision of the High Court of Australia in Computer Edge Pty Ltd y Apple Computer Inc (1986), 65 ALR 33, released the week after her decision, it might not have been necessary for this Court to add to her reasons for judgment at all. However, in the Com puter Edge case, the Australian Court held by a 3-2 margin that there was no infringement of copyright in an identical case. The appellants in argument before this Court relied heavily on the
three majority judgments in that case, particularly
that of Deane J.
* * *
All four of the Australian High Court judges who decided the question held that the source or assembly programs were copyrightable. In the cases at bar that point is admitted by the appellants.
Both Gibbs C.J. and Brennan J. in the majority held that under the Australian Copyright Act, 1968 (Cth), No. 63 the object programs embodied in the ROMs as well as the source programs were required to be literary works, and that they were not because they were not "written". Deane J. agreed with respect to the object programs. Mason and Wilson JJ. in dissent agreed that the object programs had to be themselves literary works but disagreed that they had to be in writing.
This Court must certainly accept the interpreta tion of the 1968 Australian legislation, which fol lows the English Act of 1956 [Copyright Act, 1956, 4 & 5 Eliz. II, c. 74 (U.K.)], by the highest Court in that country as final, but that is not to say that the Canadian Act, which follows rather the English legislation of 1911, should be inter preted in the same way. The Australian Act differs from ours in establishing a fragmented series of copyright rights rather than a simple comprehen sive statement.
However, to my mind the principal distinguish ing feature of the cases at bar from the Computer Edge decision is expressed in the following com ment by Mason and Wilson JJ., at page 45:
Whether it is the series of electrical impulses stored in the ROMs or the written description in binary or hexadecimal notation that is truly the object program, is not entirely clear. However, the case has been conducted on the footing that it is the former rather than the latter .... We propose to deal with the case on this footing as no issue is raised with respect to it in this court. [Emphasis added.]
It is true that Gibbs C.J. applies his conclusion that the object programs were not literary works both to the dynamic sequence of electrical impulses and to the static pattern of circuits. But Brennan J. seems to approach the matter only in terms of a computer in operation. He says, for example, at pages 54 and 55:
The electrical charges which constitute the object programs cannot be seen or touched or heard or, if they can, they do not communicate the letters of the original literary work, the source programs. Nor, for that matter, do the electrical charges communicate the letters or figures by which an object program may be represented. 6502 Assembly Language is not a lan guage; it is a code. Even if the code were thought to be susceptible of "translation" into another language, it was not so translated; the electrical charges which constitute the object programs are clearly not a language. To describe such electrical charges—no doubt helpfully enough for the purposes of com puter science—as "machine readable language" is to make metaphor serve as reality. The machine has no comprehension of thought which it is the essential purpose of language to convey, and the fact that a microprocessor is activated by a sequence of electrical charges in a predicatable way does not mean that it has understood and executed some command.
Deane J. was even more radical. He said, at page 60:
The programming of a ROM is done electrically. Its effect is essentially functional rather than visual and the program state of a ROM is ascertained not by visual observation but by actual electrical operation.
And again, at page 63:
The re-arrangement of electrons in a programmed ROM is not visible to the human eye. The programmed ROM may be used as a switching device in a functioning computer to produce a print out or visual display of something which could, for relevant purposes, properly be called a "literary work". Of itself however, and regardless of how widely one construes the phrase, the arrangement (or series) of electrons or electrical charges in the silicon chip does not constitute a "literary work". It is not written. It is not in a comprehensible language. It cannot be read. It cannot even be seen. Nor is it designed or produced to be read or seen. It is, and was designed and produced to be, an attribute of a functioning part of an operating machine.
The closest that Brennan and Deane B. came to considering the static pattern of circuits in the chips as a language was in relation to the visual display or printout which they arè capable of producing. Deane J. described as "an inversion of
logic" (at page 63) the contention that such a product could be a translation or reproduction protected by copyright.
Deane J.'s approach appears to me to be the more radical because he refuses to view pro grammed chips as anything more than "an attrib ute of a functioning part of an operating machine." For him "the program state of a ROM is ascertained ... by actual electrical operation." There is therefore no static point of view possible, only the dynamic viewpoint of an operating machine.
The appellants accepted this approach and added to it a further refinement. It is common ground that computer programs can be catego rized by function as either application programs or operating system programs. Application programs perform a specific task for the user, such as word processing, checkbook balancing, or playing a game, whereas operating system programs manage the internal functions of the computer or facilitate use of application programs. It is also common ground that the two computer programs in ques tion in the cases at bar are operating system programs.
The appellants' ultimate contention was that, regardless of the law with respect to application programs, operating systems programs are not a proper subject of copyright since they are entirely a process or method of operation.
* * *
In my view it is not necessary for this Court to decide whether, seen in a dynamic or process approach, the ROM chips are, the subject of copy right. All that is necessary in the cases 'at bar is to note that such an approach is inconsistent with the findings of the Trial Judge, which were not suc cessfully challenged.
The evidence of the respondents' witnesses, which was accepted by the Trial Judge, was that the programs in question are permanently encoded on chips that are permanent storage devices within the Apple II+. Dean J. W. Graham put it this way in his affidavit, Appeal Book, Common Appendix I, pages 171 and 172:
(a) I am personally familiar with the Apple II+ computer having used it and studied it for years. The programs APPLESOFT and AUTOSTART ROM are contained in six ROM chips which fit into sockets in the "mother board" (printed circuit board) of the Apple II+ computer. ROM chips are permanent storage devices designed to plug into printed circuit boards within computers. The devices are generic in the sense that they are manufactured in a "raw state". The raw state includes circuitry within the chip known as decoders designed to locate areas of memory storage therein and deliver the contents thereof to the micro-processor. The ROM manu facturer encodes the customer's computer programs into the storage area of the ROMs. It is not necessary for the purposes of this affidavit to deal with the technologies involved in ROM manufacturing. All that is important to state at this juncture is that ROMs by their very name (read only memories) are specialized chips designed to act as storage media for programs or data. They are permanent in the sense that any programs and data encoded therein reside therein whether or not power is turned on or off. The programs and data contained therein are readable by computers and the contents thereof can be dis played or printed in various languages by computer. They are also readable under microscope. There are numerous other permanent storage devices in which computer programs are often contained, some of which are designed to reside in the computer itself (plug into the printed circuit board) and others of which are designed to be held in external media which can be read into the computer. In the Apple II+ computer, these programs, however, are contained in six ROM chips which plug into the mother board.
The Trial Judge adverted to the fact that the appellants' expert evidence focussed on the opera tion of a computer from a hardware point of view, a perspective which she rejected, pages 14-17 C.P.R.:
Consonant with this emphasis on the computer as an electri cal machine, counsel for the defendants invited me to charac terize a program as merely specifications for a machine part. I do not accept that characterization. It is clear that when a programmer writes a program, he is thinking of instructions to the computer in terms of moving information between certain registers and of performing certain operations thereon. In this
sense it is not remiss to think of a program as being an instruction manual, but addressed to the CPU of a computer, rather than to another human being. A programmer writing a program, such as those in issue in this case, is not thinking about voltage levels, or logic gates, or conditioning the circuitry of the CPU, or of providing specifications to the manufacturer of the ROM chip for the construction of that chip. Indeed the medium in which the program eventually finds itself (punch cards, magnetic tape or silicon chip) is not of concern to him. He is thinking of the computer as having a certain number of registers and of being able to perform a certain number of limited operations and of moving information between those locations.
Not only is the program when written a sort of instruction manual, the program retains its character as such in the computer. It can be "read out" of the ROM as and when required. A "disassembler" program is often used for this purpose. Thus the hexadecimal code version of the programs Autostart ROM and Applesoft can easily be displayed on the screen of the monitor or produced in a print-out version by "reading" such directly from the respective chips. In addition this hexadecimal form of the program can be converted subse quently to its original source code assembly language version without great difficulty. Indeed it is this ability to read out of ROM the programs written therein which allows a comparison to be made between the programs encoded in the Microcom, Mackintosh and Apple II+ ROMs in this case—for the pur pose of determining whether there has been copying. (The chips in the Microcom and Mackintosh are technically not ROMs but EPROMs but this is irrelevant for present purposes.)
While the programs can be "read out" of ROM, and they are occasionally so read by human beings for diagnostic purposes when the computer is malfunctioning, this is not their primary purpose. As is evident from the above, computer programs are designed primarily to be "communications" to computers. What is more they are used to cause the computer (the CPU) to perform certain operations (to perform calculations; create and retrieve information from data banks; facilitate word proc essing). No other form of written text operates in quite this way and indeed it stretches ordinary conceptual notions of the nature of a written text to conceive of it operating in this fashion.
In her conclusions, which I set forth earlier, the Trial Judge adopted the view that the permanent etching of the programs in the computer chips was a genuine translation of the programs.
The appellants did not succeed before us in casting doubt on the software perspective adopted
by the Trial Judge. Indeed, the point of view that the programmed chips are only attributes of a functioning part of an operating machine seems to fly in the face of common sense. The original programs registered were works at rest, as it were, not in operation. When a comparison is to be made for purposes of establishing whether the memory chips are either reproductions or translations of the original works, it is only reasonable that they, also, should be taken in their static state. More over, the activity which occurs in an operating computer originates from the CPU, not from the memory chips, and by looking into the latter one cannot see the operation of the CPU. One can, however, under examination by an elecron micro scope, discover the pattern of the electric circuits, and so read the code, which turns out to be an exact replica of the assembly code, though in a different material form.
I am strengthened in my approach by the fact that the programs of the appellants which give rise to the allegations of infringement of copyright are themselves produced and sold in a static state and should therefore be compared on the same basis.
There remains the variation on the process per spective urged by the appellants with respect to operating systems programs. This approach was fully considered by the Third Circuit of the U.S. Court of Appeals in Apple Computer, Inc. v. Franklin Computer Corp., 714 F.2d 1240 (1983), a recent decision apparently not brought to the attention of the Trial Judge. Some of the Court's views are specifically related to U.S. law, but I believe that the following statement by Circuit Judge Sloviter for the Court expresses generally valid reaction to the argument (at page 1251):
Since it is only the instructions which are protected, a "proc- ess" is no more involved because the instructions in an operat ing system program may be used to activate the operation of the computer than it would be if instructions were written in ordinary English in a manual which described the necessary steps to activate an intricate complicated machine. There is, therefore, no reason to afford any less copyright protection to
the instructions in an operating system program than to the instructions in an application program.
Franklin's argument, receptively treated by the district court, that an operating system program is part of a machine mistak enly focuses on the physical characteristics of the instructions. But the medium is not the message. We have already con sidered and rejected aspects of this contention in the discussion of object code and ROM. The mere fact that the operating system program may be etched on a ROM does not make the program either a machine, part of a machine or its equivalent. Furthermore, as one of Franklin's witnesses testified, an operat ing system does not have to be permanently in the machine in ROM, but it may be on some other medium, such as a diskette or magnetic tape, where it could be readily transferred into the temporary memory space of the computer.
In the final analysis there is no reason for any kind of programmed chip to be assessed for copyright purposes in other than its material static state. In that state all such chips are the subject of copy right either as translations or as exact reproduc tions of the assembly language.
The appellants argued that the memory chips cannot be translations because in its proper sense a translation is always an interpretation rather than a one-for-one copy, and that, on the respondents' theory of them, the ROMs were exact replicas. (In fact, in the words of the respondents' counsel before us, they were described as "brailled morse code" versions of the originals.)
It is certainly true that, given the normal ambiguity of language, translations are normally no more than interpretations of the original texts. In this sense, translating may be thought of as an art rather than a science. But I am far from convinced that an exact correspondence with the original makes for less rather than more of a translation. However, I do not believe it is neces sary for this Court to make a final decision on this point, since the ROM chips are either translations or reproductions and in both cases meet the requirements of subsection 3(1). In all other
respects the reasons for decisions of the Trial Judge should be affirmed.
I do not find it necessary to decide, any more than did the Trial Judge (at page 198), the respondents' argument that the ROM chips fall within paragraph 3(1)(d) of the Act as contriv ances by means of which the works may be delivered.
In the result, I would dispose of the appeal as proposed by my brother Mr. Justice Mahoney.
 You are being directed to the most recent version of the statute which may not be the version considered at the time of the judgment.