Journal of Law, Information and Science
|
|
Home
| Databases
| WorldLII
| Search
| Feedback
Journal of Law, Information and Science |
|
CRAIG ANTHONY VERNON[*]
The public interest in encouraging new developments in computer software are not advanced by the operation of Australia’s copyright laws with respect to computer programs. Under those laws, programs are treated as literature. However, unlike other works of literature, the copyright owner of programs is not under an obligation to deposit a copy of the program with the National library.
This combined with the industry practice of only selling the object code, not the source code and standard licensing agreements binding on software users, which prohibit them from decompiling the object code back into the source code, enables the owner of the software copyright to deny the public access to the content of the program as distinct from the right to use it and hence to protect the ideas embedded in the code as well as the way in which they are expressed. By protecting the ideas as well as the way in which they are expressed, the law is protecting software copyright owners from competition from software developers who may develop different programs able to perform the same functions. This is not in the public interest as it gives the owners of the copyright a monopoly and discourages new developments.
An analysis of the cases shows that the courts cannot be relied on to protect the public interest in this area by guaranteeing the public’s right to fair use of copyrighted software. Instead legislation is needed to protect the public right to fair use by requiring that copyright owners have a duty to deposit both the source and object codes in an easily readable form, to give the public rights of fair use in copyright software and to ensure that copyright owners are not given ownership of the ideas expressed in copyright software as well as the way in which those ideas are expressed.
Due to the treaty arrangements[1] entered into by the Federal Government of Australia computer programs are protected as literary works, yet the legislature has, by wit or oversight, given more protection to them than the logic of the principles of copyright protection of literary works would dictate. Even though classified as literary works there appears to have been subtle yet deliberate oversights to exclude copyright owners of computer software from duties normally associated with literary works and prevent the public from gaining the normal fair dealing rights of access. Unlike a normal literary work where the public have access, by sale or through a library, to the text, the normal situation for software is that the purchaser or lessor, will never see the text of this literary work, that is, the source code, let alone the public at large. Further, through licensing agreements, included in purchase or lease contracts, the licensee is restricted in what they can do with the work, often far beyond whatever restrictions copyright would impose. Often, this poses no problem to the licensee as their only concern is with use of the program. Through the eyes of the user this agreement is like a "virtual" labour hire contract. This example reflects the different private interests of vendors and users of software. The public interest being much broader of course goes beyond these concerns. The public interest, as will be shown, is concerned with promoting progress and through this, ensuring that the benefits from these advances flow on to all individuals.
Computer programs are made up of two parts: source code (that is, text), which contains programming language statements and comments and in essence the algorithms and ideas expressed in that language; object code, which is an adaptation[2] of the source code through the process of compilation into machine code or machine instructions which are executed by a computer. Only the object code portion of a program (often referred to as an object module) is needed to execute and therefore use the program.
The problem with the current scheme is that due to the current practice in the market place, the protection given goes beyond merely protecting the expression. Because the source code of a program, which is more easily read and understood, is most usually not being distributed along with the object code when the program is sold, the ideas, algorithms and techniques are effectively hidden. While it is still possible to read object code it will take a lot longer to understand and discern the ideas and concepts used in a program. Decompiling or disassembling are techniques to convert object code back to source code but most software license agreements forbid this. As will be seen later in the cases reviewed copyright should only protect the expression of ideas, that is, the words that express the ideas, not the ideas themselves. Providing only the object modules exploits the dual nature of programs and thus hides or protects the ideas that they contain.
Compared with other jurisdictions like the US and the UK, Australian copyright law makes no provision for public access to published computer programs or the types of reverse engineering exceptions for fair dealing purposes allowed in the UK and the rest of the European Union (EU) due to implementation of European Council (EC) directives.[3] Because of the widespread practise of distributing object code only (OCO) often the general fair dealing provisions of the Copyright Act with respect to computer software as literary works do not apply because of the specific nature this kind of product.
One aim of this paper will be to examine and compare the mandatory deposit requirements s.407 Copyright Act (US) and those of ss.201,241 Copyright Act 1968 (Cth). It is to be noted that the deposit requirements of s.407 are quite distinct from the registration requirements for copyright. In the US there is a process of registration for copyright protection, unlike in Australia where copyright arises as a consequence of the creation of the work. This article will not discuss the matter of US copyright registration requirements.
The deposit requirements of s.407 in the US apply to "all works under copyright protection that are published in the United States are subject to the mandatory deposit provision of the copyright law”.[4] By comparison ss.201,241 only relate to "library materials" which does not include computer programs. In addition to this s.135ZE Copyright Act (Cth) specifically excludes computer programs from PART VB of the Act which contains provisions concerning copying of works by educational institutions for educational purposes inter alia. For example, protected computer programs, unless made available by the copyright owner, are not available for computing science students to study, as would normally be the literary classics to literature students.
Reference will be made to other attempts at rebalancing the copyright provisions in the public interest such as the EC Directive 91/250/EEC where decompilation is to be allowed for the purpose of error correction and obtaining information for writing interoperable programs. There are several movements arguing for similar "fair use" rights which shall be referred to also. Naturally, cases of interest will be reviewed too.
The case law history is discussed and included to reveal a number of difficulties. It is important to highlight the difficulties the judiciary are having with the subject matter of computer software due to its highly technical and scientific nature. Because of this nature some misunderstandings appear to have developed. Some of the cases will show the tension created by categorising computer programs as literary works, yet because of their unique nature they appear nothing like literary works in some aspects. One example is the dual aspect of programs, there is a text part (source code) and an execution part (object code). There is no other type of literary work quite like it. Books may be adapted to become plays and visa versa, for example, but they are separate works, that is they are separate things. Programs, however are different because their source code and their object code are much more closely linked, they are different aspects of the one thing.
Due to the difficulties that the case law history reveals, legislative reform is preferable to effect changes to avoid such difficulties by providing the judicary with guidance. Unfortunately the legislature seems unwilling or incapable of grasping the problem. It seems that the only way to gain just reforms, is for interested groups like Supporters of Interoperable Systems in Australia (SISA) to bring test cases before the High Court. The problem with a case by case approach is that it is piecemeal and can take a long time before enough cases come before the courts to settle the law in this area. In the meantime there can be much uncertainty in the law. For example when exceptions are allowed how wide are they, what are their limits? The process is slow, and very expensive.
Since the first Statute of Anne in 1709, Copyright Acts have been tinged with considerations of the public interest. The title for the Statute of Anne indicated the reasons such rights ought to be protected:
An Act for the Encouragement of Learning, by vesting the Copies of printed Books in the Authors or Purchasers of such Copies, during the Times therein mentioned.[5] The same kind of guidance is provided to the US Congress by Article 1 s.8(8) US Constitution which grants powers to make laws to promote the progress of science and useful arts.[6] However there is no such guidance in the Australian Constitution, s.51(xviii) merely provides the Parliament with power to legislate for copyrights, patents, designs and trademarks. There are no words of limitation in the paragraph. The language at the beginning of the section only uses words like "...for the peace, order and good government of the Commonwealth...", which have been the subject of case law before and were found to impose virtually no limitation at all.[7] Without any constitutional guidance the Australian and US copyright statutory frameworks vary accordingly.
The Copyright Act 1976 (US) s.102(b) provides "In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work." The US Copyright Act has thereby delineated in express words the principle known as the idea/expression dichotomy. The basic tenet of this is that ideas are free. There is no such limitation expressed in the Copyright Act 1968 (Cth) yet judgements refer to this principle time and time again. This appears to be due to the Australian judiciary importing the principle through interpretation which has been influenced by US cases, or perhaps even an inherited principle from the reception of the UK laws.[8] The cases show little reference to any legislative discussion in Hansard during passage of the Copyright Bill 1968.[9] Autodesk v Dyason refers to this traditional dichotomy repeating the notion that copyright does not extend to ideas.[10] This dichotomy is an important guiding principle in balancing the public interest.
Section 31, which explains the nature of copyright, is expressed as an enumeration of rights the copyright holder has in relation to an original work. There is no attempt to describe or distinguish the work and the ideas contained in it or the extent to which these rights extend to the ideas in a work. That is, there is no express reference to the idea/expression dichotomy.
The idea/expression dichotomy arises as a consequence of notions of public interest in promoting progress of science and the arts. Even though Australian copyright laws are lacking explicit guidance of this kind, cases will be referred to that demonstrate that these notions have been imported by the judiciary.[11] This, it is submitted, is to be considered a good thing. After all the very reason Copyright sprang to existence was for reasons of public interest. This public interest is concerned with the continuous improvement to all fields of human endeavour resulting in a better society and improving living conditions. The ultimate object being better chances for survival of all individuals and the human species.
Locating the exact source or authority for public interest becoming part of the fabric of copyright now appears to be a moot question in any event, as Australia is a signatory to the International Covenant of Economic, Social and Cultural Rights.[12] This appears to obligate members to recognise such public interest. Article 15 paragraphs 1(b) and 1(c) which provide :
1. The State Parties to the present Covenant recognize the right of everyone:
...
(b) To enjoy the benefits of scientific progress and its applications;
(c) To benefit from the protection of moral and material interests resulting from any scientific, literary or artistic production of which he is author.
Paragraphs 1(b) and 1(c) contain both sides of the copyright equation, balancing the public interest and justified expectation of benefits from progress in return for certain moral and material rights granted to the author. As today's events show these rights are worth considerable money to the Author or his or her assignee. In the last decade we have witnessed the creation of one of the wealthiest individuals on the planet through exploitation of these rights, in the form of Bill Gates. In return, for what some have equated as limited monopoly rights, it is a fair expectation that the public also gain.
In essence this is the foundation of copyright - balancing the public interest with the private interests of individuals. At this point ideas belong to the public and expressions of those ideas belong to individuals, described as Authors. By allowing this kind of distinction to be made a practical application of the idea/expression dichotomy is evident. By allowing ideas to be in the public domain these ideas can be freely adopted and used by authors in their own works. In this way these ideas can be improved upon and applied in more ingenious ways thus producing advanced and valuable works. In this way the copyright regime may be seen as an efficient way of giving incentive to authors to create works,[13] or in our case write and develop programs. This will also protect the effort and investment that has been expended,[14] which can be quite considerable, in bringing computer software products to the market.
Under various nations' copyright legislation, duties are imposed on those who hold copyright to deposit a specified number of copies of a published work with prescribed institutions, which are usually national libraries and archives. As examples, s.201 of the Act specifies the National Library and s.407 of the US Act specifies the Library of Congress. This will ensure that the works are preserved for future generations and that the work is available for those who wish to study it. Possibly the most important aspect for the purposes of balancing the public interest is that works are available for public access. This helps ensure the public benefit from works that advance science and the arts. Works that further the progress of science, technology, medicine and so on should benefit everyone. Through deposit duties this allows researchers in various fields to keep abreast of developments and follow on from each others' work. Thus progress is facilitated.
Whether this takes place in a shared research effort in areas like medicine, where research is conducted on a more cooperative basis or in commercial fields where competition, rather than cooperation, is the basis on which progress is driven, the public will benefit through the public access which legal deposit brings. This is the only apparent duty for copyright holders, and is only for published works, in return for the rights they are granted which may be of considerable commercial value. Deposit duty is a most practical way to keep ideas free and ensure continuing public benefit by facilitating progress further.
Given that publication is the point at which duties of deposit are brought into existence for both s.407 Copyright Act 1976 (US) and s.201 Copyright Act 1968 (Cth), what constitutes publication becomes an important issue. In the US Copyright Office this is defined as "the distribution of copies or phonorecords of a work to the public by sale or other transfer of ownership, or by rental, lease, or vending”.[15] What constitutes publication for Australia is defined in s.29 Copyright Act 1968 (Cth) and places the same emphasis on supply of reproductions to the public. In general it would be fair to say the definitions are the same. Importantly for the publication of software, US copyright laws specifically mention rental. Australian law includes this too by the words "whether by sale or otherwise." This is important because certain licensing agreements state that rental or licensing of the subject software does not constitute publication of that software. It is important that "publication" is adequately defined to stop those software vendors who would evade deposit duties by denying publication has taken place.
Choosing publication as the moment to bring into existence a duty to deposit work is fair as it is at this moment that commercial exploitation of copyright has begun. Computer software is excluded from such a requirement in s.201 because it is not included in the definition of "library material" to which s.201 applies. This is very unfortunate as this leaves software vendors in Australia free of any obligation to deposit but grants them the advantage of copyright protection. Software has a dual nature because there are two parts to a program - source code and object code. Source code is used to develop a program and object code is the final product that actually runs on the computer. Source code is made up of written statements in a programming language which is converted to object code by the process of compilation. Object code consists of machine code also known as machine instructions, which can be executed by a computer. Only the object code of the program is necessary for execution, that is, for the computer to "run" it. Although object code can be read and understood by humans it is far easier to read source code but this is rarely available as it usual for it not to be distributed as part of the supplied software product.
Being able to "hide" the source code, through non-distribution, and still have it protected is a double bonus to the vendor, allowing the ideas as well as the expression to be protected. This obviously circumvents the practical application of the idea/expression dichotomy. It stops any public benefit being realised through progress driven by competition as the more technologically advanced competitors can protect their ideas as well as their expression. In Australia this discriminates mostly against the local software manufacturers as they are not as advanced technologically due to size and history. This makes it difficult for the local producers to catch up and compete. Evidence of this would have to be the small number of successful Australian software suppliers.[16] With legal deposit extended to software, this catching up could be facilitated.
Calls for expansion of legal deposit to include computer software are not new in Australia. The Copyright Law Review Committee (CLRC) first made such a recommendation, along with 71 others, in its report on Computer Software Protection[17] in 1994 and they have again raised this issue in a paper Legal Deposit of Copyright Material under the Copyright Act 1968.[18] In the issues paper it was stated that deposit is only concerned with preservation of Australia's cultural heritage and arises only incidentally to copyright![19] This appears patently incorrect when having regard to the direct public interest in "progress" and Australia's recognition, through its treaty obligations, of the right to all to benefit from such progress. The view subsribed to in this article is that notions of copyright are constructed and arise out of public interest. This indeed seems to be a view widely shared, when looking at the history, some of the statutory language and cases. The National Library of Australia lists as a benefit of legal deposit that it "provides the means for research into all aspects of Australian life, history, culture, and artistic, commercial, technical, and scientific endeavour”.[20] The National Library and the National Film and Sound Archive, in a joint submission to the CLRC, have recommended extension of the legal deposit requirements to include all forms of copyright material published in Australia including computer programs.[21]
The CLRC appears to have changed its impetus since its report on Computer Software Protection where all committee members, with one exception, recommended including software in legal deposit requirements, to also accept the view that it is only incidental to cultural preservation. In its report it supported the governments commitment to protection of intellectual property rights but recognised the public interest specifically. It stated "[n]evertheless, in the creation and protection of any property rights, an attempt must be made to strike the right balance between adequate protection and the need to provide the community with reasonable access to intellectual property and the benefit which it confers”.[22] No direct mention of the idea/expression dichotomy was made, which it is submitted, is vital for the understanding of the legislature on the importance of this issue. That is, deposit is required to preserve the freedom of ideas.
The shift to a view that deposit is only of cultural interest incidental to copyright[23] could indicate a reordering of priorities and importance of this issue, which favours software vendors.[24] This may explain in part why none of the recommendations in the report have materialised in the Copyright Amendment Bill 1997.[25] The complexity of some of the issues with regards to decompilation and reverse engineering, which come with a total of 72 recommendations, have made for difficult delineation and not surprisingly has led to delays within the Attorney-General's inter-departmental committee (IDC) in its consideration of all the issues. This is unfortunate, as extending legal deposit to software is one of the most important recommendations and the single most beneficial for the public interest if it can achieve the object of public access to source code which is the text of the literary work of a program.
Deposit of published works does much for directly balancing the public interest, however exceptions are still required for fair dealing purposes. This allows some copying where it is for research or study, criticism or review, or reporting news. If the deposit of programs, that is, the source code, was available to public access then the general fair dealing exceptions[26] could operate in their intended way in favour of the public interest. They would not need to be supplemented by any specific exceptions. Unfortunately, because the source code is deliberately not available, and in most cases licensing terms and conditions now prohibit any kind of reverse engineering, neither the object code nor the source code can be copied or accessed for fair dealing purposes.
The current fair dealings exceptions in the Australian Copyright Act assume that there is public access to the work. In Australia often this is not the case at all. Thus these fair dealings exceptions fail to deliver their intended benefits to the public. The current practices of the computer industry are deliberate and intended to produce this effect. In this way the public interest in literary works that are computer programs is thwarted. Thus not giving computer programs uniform treatment like other types of literary works.
In practice it is possible to satisfy the deposit requirements yet still frustrate the purpose of the requirement. Four examples are included in the next section. Because of this there is a need for exceptions for reverse engineering where they are for fair dealings purposes.
Reverse engineering includes a number of techniques that reveal the ideas embedded in a program. The principal ones are decompilation and tracing. Decompilation is the process of converting object code back into source code. Most often this is disassembly. Assembly language is a low level programming language where statements have a one for one correspondence with machine instructions. Disassemblers convert object code into assembly language source code, even though it is possible that the original souce code for the program may have been written in a "high level" programming language like COBOL, PL/1, Pascal or C, or any other number of such languages. It is still possible to convert object code directly to source code of a high level language but less often is such a technique used. Nevertheless, the term decompilation covers all these techniques. Tracing includes a number of hardware and software techniques for intercepting data and or instructions during the execution of a program. At the very basic hardware level an oscilliscope may be employed to intercept electrical impulses to trace the data "over the wires", though it would need to be able to record these signals because of the speed at which they occur. At the software level, sophisticated tracing facilities are available as part of an operating system's diagnostic facilities, which can trace certain kinds of data during certain events or generate "dumps" of memory contents on occurence of specified events, like when a particular error occurs or when a particular program is executed. A dump will contain the data and instructions stored in the computer's memory or mainstorage at that instant. Most often these techniques are used in the diagnosis of errors and error correction. Indeed these facilities are made available so diagnostic specialists can perform this task.
Even though s.407 Copyright Act 1976 (US) specifies the requirement of deposit of works within three months of the publication, there is little guidance as to the format and way in which copies are to be presented when deposited. This apparently is left to the regulation of the US Copyright Office. Because of this it seems that the intent to provide public access is being defeated. Such frustration of public access to published software works is apparent by the cases involving reverse engineering in the US. In searching the Library of Congress Information System (LOCIS)[27] three practices are apparent:
(1) works are not given meaningful titles. Entries were found with a name simply as "computer program" thus making it very difficult to search the Library of Congress for the work. No evidence was found of use of misleading names or psuedonyms but the possibility certainly exists.
(2) depositing a whole operating system as a single program. In a sense, it is, but it also is just as correct to think of an operating system as a collection of programs working together and organised by function into components. The latter is certainly more helpful when attempting diagnosis or searching for the source code for certain functions. While it may be valid to deposit as a single program it will ensure that it will take others a long time to locate elements in which they're interested. Companies like DEC[28] and IBM[29] have deposited large operating systems as single programs in printout only. The sheer size of some works alone combined with these practices makes the need for quicker methods of finding out information about things like interfaces, file structures and the like, necessary. This need is commonplace for developing compatible programs unless the interfaces are published, which can often be the case, but just as often not.
(3) depositing source code as printout only. This disallows electronic searches of the source, which again increases the difficulties when searching. Many operating systems run into millions of lines of code. As an example 30 million lines of code is around 250 boxes of printout![30] All of this without any index or method of organisation would make reading very difficult and time consuming.
(4) by depositing just the object code, again access to the source code and the ideas within is prevented. This is apparent by a requirement by the US copyright office for printouts of object code.[31] In an Australian jurisdiction it is accepted that object code is an adaptation of source code[32] so logic would dictate that the reverse is true - source code is an adaptation of object code, thus protecting the source code too by deposit of the object code. Depositing only object code really is a mockery of the mandatory deposit provisions, assuming that their object is to provide public access to the ideas in advanced technology.[33] The deposit requirement for source code of computer programs is virtually optional, all one has to do is declare the program contains trade secrets and deposit object code instead.[34]
There is a need to reverse engineer or decompile proprietory object code in these circumstances. This would be a fair dealing or fair use in accord with the public interest. Reverse engineering is just another way of obtaining the benefits to the public that fair dealing exceptions are there to provide. The complicating factor is the licensing prohibitions which clearly show the tendancy of corporations to stymie public interest.
The public interest and primary objective of copyright as expressed by US Justice, Sandra Day O'Connor is "not to reward the labor (sic) of authors, but '[t]o promote the Progress of Science and useful Arts.' To this end, copyright assures authors the right to their original expression, but encourages others to build freely upon the ideas and information conveyed by a work. This result is neither unfair or unfortunate. It is the means by which copyright advances the progress of science and art”.[35]
In Australia there are well recognised general fair dealing exceptions for works for the purposes of: research or study (s.40); criticism or review (s.41); and reporting news (s.42).[36] These provisions are for all works in general.[37] These general provisions, particularly those of s.40 for research and study would lend scope to specific fair dealing exceptions for software due to its unique nature. The justification for these exceptions could not only be based on public interest but also on the fact that it is little or no detriment to the copyright holder.[38] In theory these fair dealings provisions apply equally to software as literary works but in reality they fail because either the source is inaccessible or reverse engineering of object code is prohibited.
Decompilation and reverse engineering for fair use purposes would seem to follow a right of public access made available through deposit. They are after all only to gain access to information that theoretically would be available in the source code but because practice in the US has shown vast printouts of source code[39] are being deposited, reading it might not be the quickest way to gain this information. Decompilation or reverse engineering would be faster as the need for information would often be quite specific and modest, for example in the case of interfaces between programs that are used to pass data.
These techniques are to be limited to fair use only. For example, it would be a breach to decompile a program completely for the purpose of altering the resulting source code and then claim it as an original work. The source code constructed by the decompiler from the object code would be an unauthorised adaptation[40] and thus a breach. If for example, the object code of a program was decompiled to give a version of source code and then with some cosmetic changes the source was used as though an independently created program it would be a breach of copyright. Most of the arguments against decompilation are concerned with piracy of this type.[41] Experience shows that regardless of the law a minority will still engage in piracy,[42] so in this sense arguments against allowing decompilation are not taking account of European experiences. The European experience shows that since introducing reverse engineering exceptions the incidence of piracy has been reduced.[43] Pirates in any case would not want to waste time and resources decompiling when they can simply make copies of the object code.[44] The decision whether to adopt such exceptions is between some unquantifiable and unclear increased risk of piracy or maintaining the public interest.
The UK has enacted legislation which allows for decompilation in accordance with EC Directive 91/250/EEC. The directive itself addresses a number of exceptions to copyright and covers three parts; (1) interfaces; (2) maintenance; (3) decompilation.[45] Article 5(1) allows whatever acts are necessary for intended use of the program, including error correction. But this can be contracted away. This would obviously include the necessary transient copies in computer storage for the execution of the program.[46] Article 5(2) provides for the making of copies for backup purposes and Article 5(3) allows the user to test the program and study its behaviour but this is only in the course of its intended use. This oddly only seems to be only of use if you were to test before you "bought" the software. Article 6 gives a limited right of decompilation. The decompilation must be necessary to obtain information, otherwise not readily available, and restricted to only the necessary parts of the original program, to achieve the interoperability of an independently created program. This excludes the general right to reverse engineer computer software.[47] The provisions of Article 5 are already part of UK copyright laws.[48] The provisions of Article 6 were carried into legislation by the Copyright (Computer Programs) Regulations 1992 (UK) which inserted s.50B Copyright, Designs and Patents Act 1988 (UK).[49] Decompilation of a program will be allowed for the creation of an independent interoperable program, known as a permitted objective, provided that: the necessary information sought is not readily available; the amount is limited to that which is essential for the permitted objective; the information is not supplied to anyone not necessary to achieve the permitted objective; and the information is not used to create a program substantially similar in its expression to the target program or to do any act that would breach copyright.[50] Decompilation as permitted, cannot be restricted by contract per s.296A Copyright, Designs and Patents Act, unless such agreement was entered befored 1st January 1993.[51]
The US position with respect to such fair dealing exceptions appears similar, even though such exceptions are not specifically provided for in the legislation. Fair use exceptions have been allowed as a defence. In Sega Enterprises Ltd. v Accolade Inc. (1992) US App LEXIS 26645 reverse engineering, specifically decompilation (and more specifically, disassembly), was used to copy a piece of code that generated a code that allowed the defendant's game software to work with the Sega console and games sysem.[52] The closest case on this point in Australia is Autodesk v Dyason (No.1) [1992] HCA 2; (1992) 66 ALJR 233 where the defendant, Dyason, had reverse engineered a logic circuit to produce the Auto Key Lock. This was held to be substantially similar to the AutoCad WidgetC program's lookup table.[53] It was found that that Dyason's purpose was not a "fair dealing" as it was not for the purpose of interoperability but so that illegal copies of AutoCad could be used.
Limited rights of decompilation are not in conflict with the law in Australia as it currently stands. The CLRC made recommendations for the same kind of right of exception for decompilation on the same limited basis as enacted, or already available, in the UK. The CLRC recommended that decompilation be allowed for the permitted objectives of interoperability,[54] error correction,[55] understanding techniques as part of noncommercial activities.[56] As well a recommendation was made to expressly prohibit decompilation for the purpose of defeating program "locks".[57] Recommendations were also made to amend the Act to void contractual terms from removing such exceptions as recommended by the CLRC.[58]
Even though legal understanding of computing science has increased continually since 1984 when the Apple Computer Case first started there are still some misunderstandings and fictions that appear to persist. An Austalian case history will be presented tracing formulation of the law but also paying attention to apparent misunderstandings and misnomers. Some of these even today may still be leading to confused thinking on software copyright protection matters. By highlighting these matters it is hoped to show that legislative guidance in the shape of the amendments for reverse engineering for fair dealing purposes could help. It is also necessary to note the very piecemeal nature of case by case development of law makes clear that legislative reforms can provide better and more complete solutions to gaps and uncertainties in the law.
This was an appeal from the full Federal Court to decide to what extent the Copyright Act 1968 (before 1984 ammendments) protected computer programs. The majority of Gibbs CJ, Brennan and Deane JJ, held that source programs were clearly "literary works" because they were readable. On the question of whether object code embodied in ROMs were literary works, Gibbs CJ said "[t]hey were not visible or otherwise perceptible, and they were not, and were not intended to be, capable by themselves of conveying a meaning which could be understood by human beings”.[60] Thus it was held object code could not be a literary work and therefore could not be an adaptation of the literary work in the source code.
This reasoning was flawed on three counts of fact. While not directly visible to the eyes object code is otherwise perceptible by use of storage monitors and other tools. The contents of the ROM may be viewed on screen. This can be seen when viewing object code in storage.[61] Various devices such as "eyecatchers" are employed so that various software specialists can see them during fault diagnosis - they are used to mark the beginning of modules and programs and even data areas, they make the job easier. It does show there is intention or awareness that a small group of humans will read object code. Finally, system programmers, software engineers and the like do understand object code.[62]
Contents of storage can be viewed like this using special types of storage monitors and other kinds of hardware and software tools for looking at "live" storage. In the earliest computers , like ENIAC and ILLIAC of the 1950s vintage, they were programmed directly in binary form by toggling the bit switches in memory. In this way programs were "hand loaded". Humans, albeit not the general population, understand object code, after all they must have, they had to design it when they invented and designed the computer.
Mason and Wilson JJ, in their dissent, were on the right track in their understanding of object code being simply the hexadecimal representation of assembly language instructions (Motorola 6502 in this case) and ROM being used as a storage device was no different to any other medium.[63] The legislative response to the majority decision were the 1984 amendments, which actually preempted the decision.
Overall, the effect of this case was to say that source code is a literary work but not object code because object code runs on computers, something which literary works do not. This is something of a contradiction - a computer program is not a protected computer program when it runs on a computer. The result of this case would have been catastrophic for the computing industry in Australia had the parliament not responded quickly. It would have meant the end of useful copyright protection of computer software. This first case shows the tension between software and the paradigm of literary work.
The misunderstandings in this case were that object code is unreadable or imperceptible, object code is only electrical impulses, and was incapable of being understood by humans and there was no intention that it should.
This case involved reverse engineering. The AutoCAD product was a computer program that required some electronic hardware called the AutoCad lock in order for it to function. The purpose of this lock was to make pirated copies of AutoCAD useless. The respondent, Dyason, marketed a similar device called the Auto Key lock, which had been designed be reverse engineering. By tracing signals produced by AutoCAD and the responses by the AutoCAD lock with an oscilliscope, it was possible to design the Auto Key lock to imitate these electronic responses.
Fortunately with a change on the bench of the High Court there was sufficient understanding to rescue the appellant, Autodesk, from an appeal which had overturned the decision by Northrop J of the Federal Court because he decided the matter using the wrong principles. Northrop had found that the Auto Key lock performed the same function as the AutoCAD lock and therefore infringed copyright! This of course crossed the traditional boundary of the idea/expression dichotomy. On appeal to the full Federal Court the decision was reversed finding neither the AutoCAD lock nor the Auto Key lock were computer programs nor did they incorporate one and therefore there was no infringement. They held, the locks were simply devices that issued automatic responses,[65] which is odd reasoning to my mind, as computers and their programs essentially automate processes.
In any case the reasoning used by the majority on appeal to the High Court correctly pointed out that a substantial part of the WidgetC program, in the AutoCAD product, had been copied by the indirect copying, through reverse engineering, of a lookup table it used. Even though as the table may have only been a small part part as Gaudron J pointed out it was the "linchpin" of the program.[66]
Hence, in my opinion they correctly found an infringement had occurred using sound reasoning. This was a moral victory too. The same victory which Northrop J had originally tried to give, albeit for the wrong reasons. The respondent through the manufacture and sale of the Auto Key lock device was encouraging pirate copying of the AutoCAD product (apparently in their advertising they had asked why pay $5000 for AutoCAD when for $500 you could use a copy of it!).
In this case the literary work paradigm, that is, infringement by copying a substantial part of the work, protected Autodesk from "predatory" and unfair competition. Northrop J had misunderstood the principle to be applied and had reasoned using a principle of "functional infringement", which goes beyond the originality requirements of copyright which are simply that the work is not copied nor a substantial part. The notion of functional infringement would effectively punish people who had not copied a program but created a program created for the same purpose as one already in existence. The full court had made errors of fact in finding that the locks themselves did not contain programs, but were just devices that gave automatic responses. If there were provisions for reverse engineering, the case could have been originally decided on the basis of whether the reverse engineering was for the purpose of a fair dealing. Instead the issues of what if anything was copied and if it was a substantial part or not had to be decided.
There has been much critcism of this case. There was an appeal which led to Autodesk v Dyason (No. 2) in which both Mason CJ and Deane J were even critical of the decision in Autodesk v Dyason (No. 1). Mason was now indicating the 127 bit lookup table was merely data and therfore not a program or part of the Widget-C program.[67] While the 127 bit lookup table is definitely data it can still be part of a program. Programs always contain some amounts of data, often it is literal data (also known as constants), which is often used for comparison.
It was mentioned above there there is tension sometimes fitting computer programs into the category of literary works. The next cases, Sega Enterprises v Galaxy Electronics and Sega Enterprises v Gottlieb Electronics demonstrate that the paradigm of literary work does not completely describe the properties or behaviour a computer program can take on. In this case it was held that a computer program which was a video game which produced moving pictures could be protected as a "cinematograph film". This finding is due in part to the legal strategy adopted by Sega knowing that they were barred from claiming infringements under ss. 37,38 because the programs were embodied in intergrated chips made under license to Sega offshore. S.24(2) of the Circuits Layout Act precluded the operation of ss.37,38 of the Copyright Act.[68]
The definition in s.10(1) Copyright Act contained a broad definition of cinematograph film and was not intended to tie it to any kind of technology. The main focus of the definition was "moving pictures”.[69] This allowed the court to hold that where a computer program produced moving pictures it could be protected as a cinematograph film, which really demonstrates the multifacted thing which computer programs are. Luckily for Sega in this case the court could find through legislation that a computer program was not just a literary work.
This decision was to protect copyright holders of "moving pictures" from copying infringements being avoided because a different method was used to produce the "moving pictures" and thereby the infringer claiming it to be a different work. This case stands out as perhaps an example where a piecemeal change was required to cover an unforeseen gap in the laws. This is entirely consistent with type of change where a piecemeal approach is probably best. It is where new doctrines and principles are to be developed that this piecmeal approach is inferior to a unified and cohesive approach which is best achieved using legislative changes.
Data Access v Powerflex is another recent example where the nature of a computer program is being misunderstood and confused decisions are the result. Powerflex a local company wrote PFXplus to be compatible with Data Access's Dataflex program. PFXplus provided the same functions as Dataflex but provided many other enhancements too. The code had been developed independently and no copying of source code or object code took place. PFXplus used the same command names, file structure and had similar macros. Jenkinson J, held that copyright was infringed on these 3 counts: (1) the command names were copyright protected; (2) the macros being of similar function were an infringement (like Northrop J); (3) by using the same file structure copyright was infringed.
This decision is bad as it imports notions of functional infringement back into copyright. The command names "save", "name", "clear", and "loop" are general words and common words that would be used in the computing field. The (originality) threshold here is way too low. These words themselves are not literary works nor could they be considered a substantial part of one. I think Jenkinson J has based the decision here on an incorrect notion of functional infringement, though he doesn't say this expressly. His finding of infringement in the macros appears more obviously to be incorrectly based on functional infringement. If this were allowed to stand this would be detrimental to the process of standardisation, which makes use of software easier for users by standardising on things like command names and other types of interfaces with the user (like the macros in this case). Standardisation is a good thing for the consumers and the industry.
The third finding of infringement due to the use of the same file structure is indeed grave cause for concern. I believe again on balancing public policy interests using the same file structure does not involve copying, directly or indirectly, a substantial part of Dataflex. File structures, unless it can be shown are significantly unique and new and novel, are inconsequential in terms of creative importance and are not the "linchpin" of a program, to use the language of Gaudron J (in Autodesk v Dyason). Many third party software products operate as an adjunct or complementary to existing software products and by necessity duplicate the file structures already in use so they can access the files already in use. In patent law where the requirement for originality and novelty is higher than copyright use can be made of existing patented devices so long as the new device or application adds something novel and original in the final product.[71]
This decision would deny the ability to follow on someone's work and improve upon it. For these same reasons under the rubric of copying a substantial part of a work, the threshold should not be set too low as there should be encouragement to improve upon that which exists. This means being able to use what already exists and add to it.
This is the appeal to the full Federal Court from Jenkinson J's decision in Data Access v Powerflex. It is significant because it specifically interprets "adaptation" as being restricted with respect to software as being a "translation". That is, adaptation is narrowly construed to avoid the notion of "functional infringement" through claims that a program that performs the same function as another is an "adaptation" of the other. Adaptation is defined in s.10 of the Act as "...a version of the work..not being a reproduction of the work;" The construction was based upon the second meaning of "version" in the Macquarie Dictionary being "translation". The choice was made by referring to the Explanatory Memorandum for the Bill which was enacted as the Copyright Amendment Act.[73]
Macros are not themselves "computer programs" they are but a part of a computer program and likewise the functions performed by function keys are not of themselves a "computer program". Macros are used to perform standard sequences of instructions for a standard function of some kind that is used often by programmers. An example would be the group of instructions to open a file, instead of constantly repeating this sequence of instructions they are replaced by the macro. The macro is then later expanded out into the sequence of instructions it represents during the process of compilation or assemby.
The decision of Jenkinson J was all but overturned except on the score of copying the Huffman Code compression table, which was necessary for PFXplus to be able to read Dataflex compressed files. Possibly the Court was confused by Autodesk v Dyason where a lookup table had been indirectly copied through reverse engineering which was held to be an infringement because it was a substantial reproduction because the table was the "linchpin" of lock program. The court acknowledged in this case that it was not the linchpin but ordered that Powerflex had infringed by copying it and ordered it not to use it.
The question here is if it wasn't the linchpin then why did the table meet the test of copying a substantial part? It did not. However the Court found that the compression table was protected as a compilation. The data or the facts in a compilation cannot be protected.[74] The Huffman Compression code algorithm is taught at all computing science faculties and really is very easy to reproduce. The only thing that may have been unique in this case may have been the order in which the letters were defined, which depends upon their frequency, usually taken from a sample of the data to be compressed. So the uniqueness, if there be any, is the result of data used as a sample to represent the typical data that was to be compressed. This is again too low a threshold to set. This table needs to be viewed as a required interface allowing interoperability with compressed Dataflex files. Alternatively, the table could have been viewed as simply containing facts and facts like ideas are not protected by copyright.
The distinction between the compression table here and the 127 bit lookup table in Dyason is that the lookup table was to facilitate the use of pirate copies of AutoCad. The compression table was only to provide compatible file support. It could perhaps be said the lookup table contained unique data or data unique for the purpose, which gave it a quality that made it a substantial part. Amending the Copyright Act for decompilation in accordance with EC Directive 91/250/EEC would clarify all of this and circumvent the need to develop new doctrines or interpretations of copyright principles to apply to the facts that arise with computer software.
Huffman compression relies on the fact that in certain data some characters are used more often than others. For example, in text the letter "e", in English at least, is one of, if not, the most frequently occurring letter. In this case it would be given the shortest bit[75] string to represent it instead of a byte (8 bits), as normally used by ASCII and EBCDIC, to represent a character. Some characters may actually end up being represented by bit strings longer that 8 bits but because of their low frequency of occurrence a net saving of bits is made thus compressing data. The Huffman algorithm is used to construct a weighted tree, where the weights are the frequencies, and the tree is used to derive the bit values representing each letter or character to be compressed.[76] By means of an algorithm data is translated into compressed data, thus the compression values are derived from sampled original data. If there is no copyright in the original data samples how can there be in the compression values? The data was probably randomly chosen from samples that were thought to be representative of the typical data that Dataflex would compress.
The practical implications of this decision are that interoperability of the program PFXplus with compressed Dataflex files could not be achieved. This is because of a decision based on the existing copyright laws and an incomplete understanding of computing science. The laws can be easily applied to literary works that are well understood but where software is concerned this is not always the case. The judiciary has no way of knowing that Huffman compression tables are common and they are based on random samples of data rather than unique data, unless expert evidence of such is put before it. No consideration of such evidence is apparent. So the result is errors of fact leading to misapplication of the law. Exceptions for reverse engineering would avoid a lot of these types of errors by confining the issue to whether the reverse engineering was for a fair dealing purpose.
If the CLRC reverse engineering exceptions had been enacted then the judiciary would have been guided to look to the purpose of the decompilation. By finding that its purpose was for interoperability it would fall within the reverse engineering exceptions proposed by the CLRC.[77] The court would not have needed to look further to see if the Huffman Code compression table could or should be protected. In any event the data or facts of a compilation cannot be protected.[78] So there is still hope for appeal on this aspect.
This case advances the prohibition on the notion of functional infringement and says that pseudocode[80] comparison cannot be used to show the source code of one program infringes the copyright of another. The claim was made that because a pseudocode constructed from one program's source code was substantially similar to that constructed from a second program that performed the same function, it showed the second program was an adaptation of the first. The court rejected this saying that since Powerflex v Data Access adaptation meant translation. For there to be an infringing adaptation of the first source code the second source code would have to be a translation of it. A "version" would mean the same thing.
Fortunately, this case was decided this way otherwise the notion of functional infringement would have been reincarnated as psuedocode is often expressed in terms of functional logic. It would be an easy task for any analyst to abstract the logic to a functional level to show two programs with the same purpose had the same or similar pseudocode. This would effectively give a monopoly position in the market to vendors who could write the first program for a particular function by giving them a method by which all subsequent vendors of programs for the same purpose could be made out to be in breach of copyright. This would obviously be too much protection and would not facilitate competition or advancement and be completely contrary to the purpose of the fair dealing exceptions.
The belief that object code is unreadable by humans, is a fiction. I still see references to this today.[81] It is more than settled now that object code is protected so it seems this misnomer can not cause any further mischief. Nevertheless it is worrying that such misunderstandings can still persist today in a specialised field such as copyright protection of software.
The threshold of the test of whether a substantial part of a work has been copied and therefore infringes copyright, does not appear to include an explicit consideration of how much originality is involved. That is, the threshold appears too low. Even if it did it would probably not have made a difference to Powerflex v Data Access as it appears the data was thought to be a compilation of unique data which was of sufficient originality to qualify for protection. The compression table was really a compilation of facts, derived from some other data, and therefore should not have been protected.
Some things should be copyable because like cliches no one can claim them to be their own expression. For example, a sorting algorithm coded in the same programming language by different programmers could well be identical segments of source code. Also expressions built on observable facts like the "brown cow" will be repeated as a natural consequence of shared human experience and language. In the same way this can happen when programming to solve the same problem in the same programming language. The same code is often used because programmers are taught using the same or similar examples. Source code for basic functions, once proven to work correctly, is often reused,[82] in the same way lawyers will reuse formal language in contracts. Such identicality may leave a judge with little apparent choice but to conclude the second program has copied a substantial part of the first without understanding that there was little choice in the matter.
There is also a basic tension between the paradigm of copies, which is at the heart of copyright, and the paradigm of use which is an understanding prevalent within the computing field. Through advanced computing network technolgies like LAN and distributed computing environments and even older virtual memory techniques one copy of a program can be used by many users. Copyright is used to stop pirating of software but license agreements must be used to control usage.
None of these problems and misunderstandings really add anything to the basic argument for legal deposit requirements for source code and limited decompilation. They do serve to illustrate the complexity and difficulty in applying the law to the facts in this area. Because of this some weight is added to the basic argument for limited decompilation as this will simplify the matter and likely limit such misunderstandings.
By now it should be obvious that Australia has some catching up to do with both the US and the UK. Judicial development may evolve principles of exceptions for reverse engineering like the Sega v Accolade case in the US but they can not order deposit of source code for all software for public access. For better or worse, we are stuck with a scheme of protecting computer software as literary works, through international treaty obligations, and also now because of the fact that economic stability and some legal certainty have developed around the scheme. This no doubt is intentional so that our market place practices will be similar to those of our trading partners thus making free trade possible. If Australia were to employ a different scheme now, disregarding the local situation, it could act as a trade barrier and thus have protectionist style effects. This would not endear us to our trade partners at all.
One of the single most important reforms would be to extend our legal deposit requirements to include software. This would bring Australia's Copyright Act further into unison with the US Act. This could in no way be perceived as a weakening of our copyright laws. This should be closely followed by the recommendations for limited decompilation exception for fair purposes such as interoperability and error correction which would also apparently bring the law into line with the US copyright law. The Sega v Accolade Case in 1992 where the defendant, Accolade was allowed the defence of an exception for the purpose of a fair use for decompiling Sega's game programs. In this Accolade went further than just limiting the decompilation to that necessary to achieve interoperability. So the CLRC recommendations are more limited. A limited decompilation exception for error correction, where no correction is readily available, seems only fair from a consumer point of view and entirely in accord with the principle of public interest espoused regularly in US case law.
Since these reforms have been widely called for and would apparently bring Australian copyright further into line with US copyright, the question must be asked why have these recommendations not found their way into the Copyright Amendment Bill 1997 (Cth)? There has been concern expressed concerning decompilation by some manufacturers as this may lead to piracy. For reasons already expressed in this article that can not be a serious concern when the recommendations are looked at in detail and given the tendancy to follow the US in copyright anyway. Further the European experience has been that such limited decompilation rights have not facilitated piracy and in fact software piracy has been reduced since introduction of the EC directive across European Union countries. The EC directive recommendations were not dissimilar to the CLRC recommendations. Under this recommended scheme if the reverse engineering is for a permitted purpose it is a defence to a claim of infringement. Therefore this puts the onus to prove that the purpose of the decompilation is legitmate on the decompiler.[83] Far better to have the limits clearly defined in the statute than have unsettled limits that a judicial development could yield because of the case by case operation of the legal system. This is precisely what has happened in Sega v Accolade in the US. This is the first case where a right to decompile has been recognised but no limitation was placed on how much could be decompiled. More cases will need to come before the Court to settle the limits on decompiling.
It would seem that opposition to the CLRC limited decompilation recommendations due to piracy concerns are tenuous and it must be questioned if there is not an ulterior motive in continually putting forward such arguments. This may be the only means at the disposal of such opponents to oppose the CLRC recommended changes, as an argument to simply maintain the status quo and its incidental inequities, would fail for lack of any merit. Other diversionary tactics may also be being used as it appears software reform is on the back burner and now changes for the "information super highway" and associated broadcast technology are on the agenda as evidenced by the Copyright Convergence Group's report Highways to Change[84] as well as the Attorney-General's Department discussion paper, Copyright Reform and the Digital Agenda.[85]
By maintaining the status quo with respect to software protection foreign competitors are given a clear advantage over local software producers as they do not have to provide public access to their source code. Whereas Australian competitors wishing to compete in the US must deposit their source code and hence share their ideas with US competition[86] but when US software suppliers compete here there is no need to deposit. The end result is that we have to share any advances we make but they get to keep theirs to themselves. Unless an Australian software producer has an office in the US it may be quite expensive and impractical to visit the Library of Congress to gain such public access and perhaps only to be surrounded by a large number of boxes of printout. While Australian software developers can do the same, it should not be so easy to frustrate the public interest in deposits. Soft copy should also be requested that can be searched electronically.
Legal deposit and limited decompilation are necessary for balancing the public interest and also putting our local software manufacturers on an equal footing as foreign competitors. Unfortunately the opportunity to make the necessary legislative amendments has been missed for the present as the current bill before the parliament contains no such amendments. Possibly the amendments could be added during the course of debate through the houses but this seems unlikely. The final hope would be an appeal from Powerflex v Data Access to the High Court. This is something interested groups like SISA should be allocating funds and resources for.
Just as importantly, lobbying for legislative reforms needs to be maintained. What needs to happen is that the large players in the field like IBM, DEC and Microsoft, be convinced that they are better off with carefully limited fair dealing exceptions through legislative amendment. Case by case development has the disadvantage of being piecemeal and until enough cases have come before the courts the level of uncertainty can increase. Cases like Sega v Accolade create exceptions for decompilation for interoperability but to what extent, at this point it seems unlimited as Sega's programs were decompiled completely. This kind of unlimited scope could allow collateral purposes to fair dealing purposes to be satisfied too. The same kind of decision could be made in Australia as cases like Powerflex v Data Access demonstrate a judicial willingness to allow the use of standardised interfaces like macros, command names and file structures and by clearly rejecting notions of functional infringement (ideas are free). Sega v Accolade seems the next logical step given the current position of interpretation of copyright law and the large influence US cases have in this field of law. With widespread support and the backing of large computer companies the chances of getting the reforms this article proposes wil be handsomely improved. The tradeoff here for the large computer companies might be that the time within which to deposit can be extended from three months to to something like twelve. For a fee this time could perhaps even be extended for perhaps another six months. This should give adequate time to exploit the competitve advantage of being first into the market with a new product. If deposit is not made within the stipulated time period then fines should be imposed, copyright protection should still be maintained, as that is a duty the state owes the holder. Deposit of course is the duty the copyright holder owes the state in order to serve the public interest.
Because of inaccessibility to ideas in computer programs because either, the source code is made unavailable, or through license conditions object code may not be reverse engineered, the public interest policy behind the Copyright Act is defeated. The best road to reform is legislative amendment. Some of the reasons are that it offers a more unified and cohesive approach rather than the piecemeal case by case approach necessitated by the legal system. However this path has become blocked due to the complexity of the issues to be considered and the process has stalled. The reason a unified approach is required is that reforms are required for legal deposit of source code and reverse engineering exceptions to work in unison by complementing each other. This is due to some of the practices revealed by research that show despite deposit requirements public access to the ideas in programs is denied. This leaves only the case by case approach available through the legal system. Recent history of cases shows judicial misunderstandings may cause meanderings along the way but there is a direction towards making exceptions for fair dealings for computer software as Sega v Accolade shows. In Australia, there is room for this kind of exception too, as the Copyright Act provides for fair dealing exceptions. For the time being this is the path that must be taken even though it will be slow and expensive.
If a better appreciation by the opponents of the proposed amendmentments of some of the dangers of not taking the legislative path compared with the advantages it has to offer could result in an increased solidarity in lobbying, the priority of these amendments could be increased.
1. Ricketson, Intellectual Property: Cases, Material and Commentary, Butterworths, Sydney, 1994
2. Lane, Lane's Commentary on the Australian Constitution, Law Book Company, Sydney, 1986
3. Lane, Fifth Cumulative Supplement to [2], Law Book Company, Sydney, 1993
4. Hanks, Australian Constitutional Law: Materials and Commentary, 5th Ed, Butterworths, Sydney, 1994
5. Copyright Law Review Committee, Report on Computer Software Protection, Attorney-General's Department, Canberra, 1994
6. Christie and Hunter, Recent Developments in Computer Law, Longman BLEC, Melbourne, 1994
7. Beth Gaze, Copyright Protection of Computer Programs, Federation Press, Sydney, 1989
8. Carr, Arnold, Computer Software: Legal Protection in the United Kingdom, Sweet & Maxwell, London, 1992
9. Bainbridge, Software Copyright Law, 2nd ed, Butterworths, London, 1994
10. Jensen, Tonies, Software Engineering, Prentice-Hall, Englewood Cliffs NJ, 1979
11. Knuth, The Art of Computer Programming, 2nd ed, Volume 1: Fundamental Algorithms, Addison-Wesley, Reading MA, 1973
12. Bowrey, LAW423 Seminars in Copyright Law: Course Materials, Vol 1, Macquarie University, Sydney, 1997
13. CLRC, "Legal Deposit of Copyright Material under the Copyright Act 1968", Attorney-General's Department, Canberra, 22 August 1997, see internet http://www.agps.gov.au/customer/agd/clrc/Legal.html
14. US Copyright Office, Circular 7d, "Mandatory Deposit of Copies or Phonorecords for the Library of Congress (Under Section 407 of the Copyright Act)", see internet URL
http://lcweb.loc.gov/copyright/circ07d.html
15. US Copyright Office, Circular 61,"Computer Programs", see internet URL
http://lcweb.loc.gov/copyright/circs/circ61.html
16. International Covenant on Economic, Social and Cultural Rights, Australian Treaty Series 1976 No.5, Australian Government Publishing Service, Canberra, 1995
17. Covenant establishing the World Intellectual Property Organization [WIPO], Australian Treaty Series 1972 No.15, Australian Government Publishing Service, Canberra, 1995
18. Universal Copyright Convention, as revised, Protocols 1 and 2, Australian Treaty Series 1978 No. 2, Australian Government Publishing Service, Canberra, 1995
19. SISA, The CLRC Recommendations on Reverse Engineering and Decompilation: Giving Local Developers an Equal Right to Compete, see internet URL www.sisa.org.au
20. "Highways to Change", Copyright Bulletin, Australian Council of Libraries and Information Services, Issue No.10, September 1996
21. National Library of Australia, Legal Deposit in Australia, 4th ed, NLA, Canberra, 12 May 1997 see internet URL www.nla.gov.au/1/services/ldeposit.html
22. Rohan Mishra, "Reverse Engineering in Japan and the Global Trend Towards Interoperability", E Law - Murdoch University Electronic Journal of Law, Vol 4, No.2, June 1997. See internet URL www.murdoch.edu.au/elaw/issues/v4n2/mishra42.html .
23. Owen Trembath, The Copyright Amendment Bill 1997, NSW Young Lawyers, Sydney, 1997
24. Copyright Act 1968 (Cth), Reprinted as at 1 July 1995, Reprint No.7
25. Copyright Act 1976 (US)
26. Copyright, Designs and Patents Act 1988 (UK)
Examples of clauses in software licensing agreements prohibiting decompiling or disassembling the software.
Lotus
You may not, without written permission from Lotus:
Alter, modify or adopt the LOTUS Software or Documentation, including, but not limited to translating, decompiling, disassembling, or creating derivative works.
WPCorp
You may not decompile, disassemble, reverse engineer, copy, create a derivative work, or otherwise use the software except as stated in this agreement.
IBM
You may not:
2) reverse assemble, reverse compile, or otherwise translate the program except as specifically permitted by law without the possibility of contractual waiver.
Microsoft for Windows 95
Limitations on Reverse Engineering, Decompilation and Disassembly. You may not reverse engineer, decompile or disassemble the SOFTWARE PRODUCT except and only to the extent that such activity is expressly permitted by applicable law notwithstanding this limitation.
[*] B.App.Sc(Comp Sc) (UTS) LLB (Macq).
[1] Copyright Law Review Committee, Report on Computer Software Protection, Attorney-General's Department, Canberra, 1994, pp. 26,27,33.
Article 10(1) TRIPS Agreement provides: "Computer programs, whether in source or object code, shall be protected as literary works under the Berne Convention (1971)." This was agreed to at the Uruguay Round of GATT, 15 December 1993 and signed 15 April 1994.
[2] Since the 1984 amendment of s.10(1) Copyright Act to include the term "material form" there is little doubt that the object code of a computer program is an adaptation of the source code despite the finding in Computer Edge Pty Ltd v Apple Computer Inc [1986] HCA 19; (1986) 65 ALR 33 which was decided on the law before those amendments. See Ricketson, Intellectual Property: Cases, Material and Commentary, Butterworths, Sydney, 1994, p.234.
[3] EC Directive 91/250/EEC, covered in further detail below.
[4] US Copyright Office, Circular 7d.
[5] Ricketson, Intellectual Property: Cases, Material and Commentary, Butterworths, Sydney, 1994, p.69.
[6] The actual words of the provision are "[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."
[7] "the utmost discretion of enactment", importing amplititude rather than restriction, see Lane, Lane's Commentary on the Australian Constitution, Law Book Company, Sydney, 1986, p.80, N5, N6: Chenard v Joachim Arissol [1949] AC 127, 132; Cobb & Co v Kropp [1967] AC 141, 154; found the words to be interpreted in the same way as for state constitutions. R v Sharkey [1949] HCA 46; (1949) 79 CLR 121, 152-3; and see Lane, Fifth Cumulative Supplement to Lane’s Commentary on the Austrlian Constitution, Law Book Company, Sydney, 1993, p.31 N6 Union SS Co v King [1988] HCA 55; (1988) 166 CLR 1, 10; "peace, order and good government" are not words of limitation.
[8] Hanks, Australian Constitutional Law: Materials and Commentary, 5th ed, Butterworths, Sydney, 1994, p.188. The view was held that the laws of England were imported into Australia by the first colonists, the Australian Courts Act 1828 (UK) was enacted to clarify and settle this view. This view has remained undisturbed even after the doctrine of terra nullius was rejected in Mabo v Queensland (No.2) [1992] HCA 23; (1992) 175 CLR 1.
[9] Sega v Galaxy ; Sega v Gottlieb (1996) 69 FCR 268 at 274 does refer to a speech by the Attorney-General but this was in relation to what was meant to be included in the category of cinematograph film and not the dichotomy, nevertheless there appear to be other good sources of law for this principle.
[10] Autodesk v Dyason [1992] HCA 2; (1992) 66 ALJR 233 at 238, Lindley LJ's description of this dichotomy in Hollinrake v Trusswell [1894] UKLawRpCh 158; [1894] 3 Ch 420 at 427 was quoted at length.
[11] Ibidem.
[12] Australian Treaty Series 1976 No. 5, Australian Government Publishing Service, Canberra, 1995.
[13] Breyer,"The Uneasy Case For Copyright: A Study Of Copyright In Books, Photocopies, And Computer Programs", (1970) 84:2 Harvard Law Review 152-194, extracted in Bowrey, LAW423 Seminars in Copyright Law Course Materials, volume 1, Macquarie University, Sydney, 1997, pp.24-39, at p.26.
[14] Ibidem, at p.27, this is to really protect the expected return on investment.
[15] US Copyright Office, Circular 7d.
[16] In the Australian mainframe market I can only think of one example, Software Developments Incorporated. Their product NetMaster dominated the Australian market particulary in the finance sector, and had success in the global market. Today its success continues but the company has been bought by a US company.
[17] See note 1 above, p.19,20 recommendation 2.66.
[18] CLRC, Legal Deposit of Copyright Material under the Copyright Act 1968, Attorney-General's Department, Canberra, 22 August 1997.
[19] Idem, at para.11 p.3.
[20] National Library of Australia, Legal Deposit in Australia, 4th ed, NLA, Canberra, 12 May 1997.
[21] See note 18 above, point 6.
[22] See note 1 above, Summary of Recommendations, para 2.01, p.4.
[23] See note 19 above.
[24] Oracle Systems actually made a submission opposing mandatory deposit requirements. It is surprising there were not more, although some submissions were made in confidence, see note 1 above para.16.08.
[25] Owen Trembath, The Copyright Amendment Bill 1997, NSW Young Lawyers, Sydney, 1997, the only major change was the introduction of moral rights.
[26] ss.40-42 Copyright Act 1968 (Cth).
[27] Electronic public access to LOCIS is available by telnet://locis.loc.giv or tn3270://locis.loc.gov via the internet.
[28] VAX/VMS version 4.2, LOCIS call number TXu-286-800.
[29] OS/390 release 2, LOCIS call number TXu-769-716.
[30] Based on 60 lines per page and 2000 pages per box, for an idea of size a box of 2000 pages is 220mm deep x 310mm long x 240mm wide.
[31] See email forwarded by Dr. Kathy Bowrey, from Mark Lewis regarding a requirement for a printout of object code (see appendix C).
[32] See note 2 above, and see also Admar v Ezy Systems case notes below.
[33] An assumption which no longer appears as certain as first thought. Cultural preservation might have been the original motive but, the public interest in access, it is posited, has overtaken this motive.
[34] US Copyright Office, Circular 61.
[35] Feist Publications v Rural Telephone Service Co [1991] USSC 50; (1991) 499 US 340, 349 extracted in CETUS (Consortium for Educational Technology in University Systems) Working Document 1/18/95 see internet http://arl.cni.org/scomm/copyright/uses.html, CETUS consists of the State University System of New York, California State University System and the City University of New York.
[36] See note 5 above, pp.302-5.
[37] The sections begin with "A fair dealing with a literary, dramatic, musical or artistic work, or with an adaptation of a literary, dramatic or musical work,..".
[38] See note 5 above, p.299.
[39] I have reason to believe that some of these printouts may actually be object code given the lack of presentation guidelines and a complaint on an email note that he was required by the US Copyright Office to lodge a printout of the object code!
[40] Adaptation is to be interpreted as a translation when dealing with computer software, decided in Powerflex v Data Access, see case notes below.
[41] See note 1 above, paragraphs 10.80 and 10.81, p.170. IBM and Oracle were concerned about the increased risk of piracy. Lotus too expressed concern that the purpose of reverse engineering was to create clones, see also paragraphs 10.38 and 10.50.
[42] SISA, The CLRC Recommendations on Reverse Engineering and Decompilation: Giving Local Developers an Equal Right to Compete, Section 7, pp.10,11.
[43] Ibidem.
[44] Ibidem.
[45] Carr, Arnold, Computer Software: Legal Protection in the United Kingdom, Sweet & Maxwell, London, 1992, p.16.
[46] Idem, pp.17-18.
[47] Idem, pp.18-19.
[48] Ibidem.
[49] Bainbridge, Software Copyright Law, 2nd ed, Butterworths, London, 1994, pp.160-2.
[50] Ibidem.
[51] Ibidem.
[52] Idem, pp.158,9.
[53] See note 5 above, pp.251-61.
[54] See note 1 above, para.2.23, p.10.
[55] Idem, para.2.27, p.11.
[56] Idem, para.2.28, p.12.
[57] Idem, para.2.29.
[58] Idem, para.2.31.
[59] See note 5 above, pp.132-37.
[60] [1986] HCA 19; 65 ALR 33 at 39, the underlines are mine.
[61] See figure 1 in Appendix A.
[62] I am a software engineer and I regularly read object code.
[63] [1986] HCA 19; 65 ALR 33 at 44,45.
[64] See note 5 above, pp.251-61.
[65] [1992] HCA 2; 66 ALJR 233 at 239, Dawson J's judgement with whom Mason CJ, Brennan, Dean and Gaudron JJ agreed.
[66] Autodesk v Dyason (No. 2) [1993] HCA 6; (1993) 25 IPR 33 at 57, see note 5 above, p.257.
[67] Idem.
[68] Sega v Galaxy; Sega v Gottlieb [1996] 761 FCA 1 extracted in LAW423 Seminars in Copyright Law: Course Materials, pp.465-72, p.465.
[69] 69 FCR 268 at 274,275.
[70] Unreported, viewable on Australian Legal Information Institute, home page: www.austlii.edu.com
[71] See note 5 above, p.641. s.18(1)(b) (i) and (ii) Patents Act 1990 (Cth) provides that a patentable invention is one that when compared to prior art it is novel and involves an inventive step.
[72] Ureported, as viewed on www.asutlii.edu.au
[73] Idem, at17th page of the judgement.
[74] Justice Sarah Day O'Connor in Feist v Rural Telephone Service [1991] USSC 50; (1991) 499 US 340 see Price and Pollack, "The Author in Copyright. Notes for the Literary Critic", Cardozo Arts & Entertainment Law Journal (1992), Vol 10 pp.703-730 extracted in LAW423 materials p.77.
[75] Bit is short for binary digit which is either 0 or 1 ie. base 2 numbering.
[76] D E Knuth, The Art of Computer Programming, 2nd ed, Volume 1: Fundamental Algorithms, Addison-Wesley, Reading Massachusetts,1973, pp.402-5.
[77] See note 1 above, paragraph 2.23 p.10.
[78] See note 32 above, and also note the CLRC view that it would be unnecessary to have such explicit guidance like s.102(b) Copyright Act 1976 (US) because the fundamental nature of copyright does not protect ideas (see paras.6.29 and 6.30) see note 1 above, p.62.
[79] Unreported, judgement viewed on www.austlii.edu.au
[80] Jensen,Tonies, Software Engineering, Prentice-Hall, Engelwood Cliffs, 1979, pp.264-7. "Pseudocode is an extension of the Thinking Process." It is actually a freeform way of expressing the logic required to perform functions that a program is to achieve in English language. It is used as a design and analysis tool in the specification of programs in the same way flowcharts are. The logic expressed can be very detailed or abstracted to high level function.
[81] SISA, The CLRC Recommendations on Reverse Engineering and Decompilation: Giving Local Developers an Equal Right to Compete, section 3 on reverse engineering, pp.5,6.
[82] This concept is generaly known as "copybooks".
[83] SISA, The CLRC Recommendations on Reverse Engineering and Decompilation: Giving Local Developers an Equal Right to Compete, point 7.
[84] "Highways to Change", Copyright Bulletin, Australian Council of Libraries and Information Services, Issue No.10, September 1996, p.1
[85] See internet site http://law.gov.au/publications/digital.htm
[86] Although they can use the same evasive tactics apparently in practice that were discussed earlier.
AustLII:
Copyright Policy
|
Disclaimers
|
Privacy Policy
|
Feedback
URL: http://www.austlii.edu.au/au/journals/JlLawInfoSci/1998/15.html