For centuries intellectual (or industrial) property rights have been regarded as a somewhat arcane area of the law. Times are changing. With the increasing importance of the service sector to national economies, attention increasingly requires to be directed to the means of protecting investment in intellectual rather than tangible property. The twin pillars of the intellectual property system are the concepts of patents and copyright. Traditionally, patents have provided the means for protecting and rewarding the practical work of inventors whilst copyright functions in more aesthetic areas of human activity such as music, art and literature.
The first recorded patent statute was enacted in Venice in 1474. In England, the first statutory intervention occurred with the enactment of the Statute of Monopolies in 1624 which sought to put an end to the Crown's issuance of monopolies. For Scotland, patents continued to be issued under the Royal prerogative until the Act of Union had the effect of extending the provisions of the Statute of Monopolies to Scotland. Today, United Kingdom patents are issued under the terms of the Patents Act 1977. An inventor may also apply for a European patent subject to the provisions of the European Patent Convention. The extent to which software related inventions may be awarded a patent is a matter of some controversy. Both the Act and the Convention state that a patent may not be awarded for a "program for a computer". It has been recognised however, that an invention is not to be barred from patentability because it contains computer programs and even that a claim will be considered where the novelty resides only in the program related aspects. Although the topic of software patents has raised considerable passions especially in the United States, the exacting requirements of novelty and inventiveness which constitute a feature of the patent system seem likely to ensure that the role of patents remains on the periphery of the topic. Primacy, at least for the present, rests with the copyright system which will constitute the subject of the remainder of this chapter.
The law of copyright developed from the 18th century with the enactment of the Copyright Act of 1709. As the name suggests, copyright is concerned with the right to make copies of a work. Such a right only assumed practical importance with the development of technology in the form of the printing press which permitted multiple copies to be made of a work. Initially, the new technology was seen by both religious and secular authorities as posing a threat to their own ability to control the dissemination of information and knowledge and its exploitation was controlled by requiring that a licence be obtained in respect of every publication. Under such a regime, questions concerning the extent of any property rights in works were of very limited significance. A party in receipt of a licence could make copies in accordance with its terms. Persons who did not hold a licence would face sanctions imposed by the licensing authorities were they to venture to print copies.
Government control over publishing was abolished towards the end of the seventeenth century. From this point, the question who would possess the right to produce copies of a work began to assume real economic significance. The same technology which would allow an author (or a publisher) to produce multiple copies of a work for profitable distribution could equally well be used by others to produce competing works. Almost for the first time therefore, the question required to be resolved whether the author of a work (or in practice the publisher to whom any rights would be assigned) possesses a property related interest in the work sufficient to allow them to prevent or seek a remedy for any unauthorised reproduction. In 1709 the first copyright statute was passed which conferred an exclusive right upon an author or his or her assignee in respect of a work which was registered with the Stationers' Company. The right would subsist for a period of 28 years.
The copyright system has been subject to periodic amendment in the intervening centuries. The requirement to register a work as a pre-requisite for its protection was limited by the Copyright Act of 1842 and abolished by the Copyright Act of 1911 which also extended the duration of protection to a maximum of 50 years after the death of the author[1]. Copyright contains many of the attributes of property, indeed the Copyright, Designs and Patents Act of 1988 states expressly that "(c)opyright is a property right which subsists in the following categories of work ...". Copyright can be bought and sold and any copyrights will form part of an author's estate upon his or her death. Copyright, however, remains a special form of property and the passage of two and a half centuries have done nothing to challenge the validity of Lord Kames observation that: "it is a statutory property, and not a property in any just sense to be attended with any of the effects of property at common law" (Kames, Rem. Dec Mor. 8298).
As copyright has developed in the United Kingdom, it has come to protect virtually every form of recorded information. Over the years, the scope of copyright has expanded to keep pace with technical developments. Thus the Copyright Act of 1911 brought the products of the fledgling film industry into the categories of protected works. The invention of the phonograph also saw sound recordings becoming a subject for protection whilst the Copyright Act of 1956 extended the scope of protection to include television broadcasts.
These events indicate that copyright has proved a flexible concept capable of being applied to the output of new forms of technology. In the early days of computerisation, there was little concern for the legal protection of the creative and intellectual efforts which went into the development of computer programs. The principal reason for this was that equipment was supplied as a unit with the same supplier being responsible for both hardware and software components. As software could only be used in conjunction with the hardware for which it was designed, there was little incentive to attempt to make unauthorised copies.
From the 1960's this system began to break down and issues of intellectual property rights began to surface. The first wave of cases concerned the question whether computer programs could be regarded as protected under the law of copyright. Although the matter was never determined by the higher courts, the view was favoured in a number of interlocutory decisions that computer programs should be protected as a species of literary work. Although the relationship between the somewhat impenetrable (at least to the layperson) lines of code which make up any program and a novel or other form of literary work appears somewhat obscure, the fact that items such as football fixture lists and even tide tables and tables of logarithms had been regarded as qualifying for protection lent considerable support to such an approach.
Any question whether computer programs were entitled to copyright protection was resolved with the enactment of the Copyright (Computer Software) (Amendment) Act of 1985. The terms of this measure were restated in the Copyright, Designs and Patents Act which, subject to amendments made in 1992 to incorporate the provisions of the EEC Directive on the Legal Protection of Computer Programs (Council Directive 91/250/EEC. OJ 1991 No L122/42), constitutes the current copyright regime. Subject to the most minor quibbles concerning extremely small computer programs, there is no doubt that the act of copying a program will constitute an infringement of copyright. Computer programs possess an additional dimension which makes their location in the category of literary works somewhat uneasy. Certainly, a program's code can be so regarded but any program will also have an on-screen appearance. This can be replicated even though there may be no similarity at the level of the code. One of the most controversial issues in the intellectual property field concerns the extent to which non-literal copying, replication of what used to be called the "look and feel" of a computer program will be permitted.
Much is written and spoken today on the subject of software piracy - the producing and supplying of computer programs which infringe copyright in an original work. Estimates have put the world-wide losses caused by such activities at some $12.8 billion with almost 40% of this loss being incurred within Europe. In terms of numbers, it has been suggested that 13 unauthorised copies are made of every computer game. For business related software the ratio is calculated at 4 to 1[2]. In Russia it has even been calculated that some 98-99% of copies are unauthorised. As with estimates of the scale of computer crime, these calculations are based on based on particular sets of definitions and assumptions which may be open to challenge. In particular, if software piracy could be stopped overnight, it seems unlikely that the sales of computer games would rise thirteen-fold. Indeed, it might be argued that, deprived of the possibility of obtaining cheap software, some potential users would determine not to purchase a computer. Elimination of piracy might therefore damage the sales of hardware producers.
In many instances, software piracy will raise no significant legal issues. As stated above, the unadorned act of copying the contents of a computer program will constitute an infringement of copyright. In this situation, a conflict of interests between equipment producers and copyright owners has been a notable feature of recent years. The problem, of course, is not unique to computers. In almost every area of copyright, technological developments are making it easier for copyright infringement to occur. The ubiquitous photocopier makes us all potential copyright infringers to an extent undreamed of twenty years ago. The development of the cassette recorder and in particular its incorporation into musical systems including a record player and radio created a new generation of domestic infringers. With the advent of systems containing two cassette recorders allowing copies to be made of tapes (often in less than the normal playing time of the original), the attempt was made by representatives of copyright owners to seek recourse from those producers who were putting technology which was almost certainly going to be used for unlawful purposes onto the market. In the case of CBS Songs Ltd v. Amstrad Consumer Electronics ([1988] AC 1013) however, the House of Lords rejected all arguments to this effect. Even though the producer's promotion of the equipment was criticised as "deplorable" and "cynical", the fact that it was capable of being put to legitimate uses sufficed to exclude the producer from any form of liability.
In the software field, although it is possible for a software producer to utilise techniques to make copying difficult the use of these has met with considerable customer resistance, and most of the anti-copying devices can be overcome by a reasonably skilled user. In general, copying a computer program is a very simple task. The contents of a disk can be copied in seconds using standard equipment. In one significant respect, the producers of computer programs are more vulnerable to those who wish to copy their works than most other copyright owners. Where devices such as photocopiers or cassette recorders are used to copy a protected work, the copy will be of inferior quality to the original. If the process is repeated, each generation of copies will show a further reduction in quality, soon reaching the point where it may be unusable for any practical purposes. Because of the digital nature of computer programs every copy of a program will be identical to the original. A fiftieth generation copy will be identical to the original. Much the same applies to the musical contents of compact discs and strenuous efforts were made by copyright bodies to prevent the marketing - or at least require the incorporation of anti-copying devices - of digital tape recorders.
Although it may be stated as a general rule that the reproduction of a protected work constitutes infringement, problems arise in reconciling the claims of copyright owner and legitimate user. Two issues may be cited as illustrative of this topic. First, because of the manner in which a computer functions, on every occasion when a program is used, its contents are copied from a permanent storage location - which may be on a floppy disk or a segment of a computer's hard disk - to a temporary working location. Use of a computer program, therefore, calls for its reproduction, a situation which is not found with more traditional literary works. A second issue concerns the entitlement of a user to make a second copy of a computer program to be brought into use in the event that the original copy is damaged or destroyed. Again, whilst it would be difficult to argue that the purchaser of a book would be entitled to copy its contents against the eventuality that the original were destroyed, computer programs and the devices upon which they are stored are much more fragile creatures. A coffee stain or ink spot will not prevent the book being used for its intended purpose. An electronic equivalent is likely to render both the program and the user's investment worthless.
The 1988 Act remained silent in respect of both of these issues. This created a potentially absurd scenario whereby users would pay considerable sums to acquire a copy of a computer program only to be faced with a claim alleging copyright infringement were they to be bold enough to use it. Although the point was never tested before the courts, it is likely that such a result could have been avoided by the implication of a term into the contract of supply permitting normal use of the program - although the question what might be considered normal use would itself often admit of no easy answer. Implying a term permitting the making of a back-up copy might be more difficult to justify. The practical solution adopted by most suppliers was to issue users with a licence permitting both use of the software and the making of a back-up copy. The downside of such licences was that they would normally also exclude any form of liability on the part of the producer and/or supplier. As a matter of contract, the enforceability of such licences is open to question. In the vast majority of cases, a contract for the supply of software is concluded before the customer becomes aware of the licence which is normally contained inside the program's packaging. Coupled with the fact that the licence is offered by a producer whilst the program may well have been supplied by a retailer, it must be likely that the terms of the licence will be regarded as constituting a further offer which may be either accepted or rejected by the customer. The difficulty, of course, was that the customer might require the positive aspects of the licence in order to lawfully put the program to its intended and reasonable use. Subject to the possible intervention of the Unfair Contract Terms Act 1977, the price would be the denial of all remedies in the event that the program should prove defective.
To an extent these issues were addressed in the European Directive. As implemented by the Copyright (Computer Programs) Regulations 1992 (SI 1992 No 3233), it is now provided that although the copyright owner retains the exclusive right to reproduce the work, such copying as occurs in the course of the normal use of the program is not to constitute infringement when it is conducted by a lawful user (Regulation 8 inserting a new section 50C into the 1988 Act). Perhaps surprisingly, this right may be excluded by the terms of an agreement. For the first time, therefore, a use right forms a part of the United Kingdom copyright regime. Provision is also made in the Regulations allowing a lawful user to make a back-up copy of a program where this is "necessary" for its operation (Section 50A). The use of the word "necessary" (which is also found in the Directive) may rob this provision of any real meaning. It is always advisable to take a back-up copy of a program (many producers advise users to copy the program and to use the copy leaving the original as back-up) but it is implicit in the term "back-up" that the second copy is not necessary for the program to function. It is only in the event that the original copy should suffer damage that the back-up will be required.
Whilst conferring an absolute right to make a back-up copy would have the excessive side effect of depriving a producer of the possibility of utilising anti-copying devices, the conferring of a meaningless right must be considered equally objectionable and leaves users in the situation where they may have little alternative to accepting the terms of a licence.
It is received wisdom that every computer program contains errors or "bugs". In accordance with the requirements of the European Union Directive, it is provided that an authorised user may copy or adapt a program "for the purpose of correcting errors in it" (Section 50C). This provision might appear to give a user carte blanche to copy a program in the quest to discover errors. An alternative, and perhaps preferable view, is that the right will extend only in respect of particular errors which have been discovered by the user in the course of running the program in a normal manner. Even on this basis, uncertainties remain as to the extent of the user's rights. Computer programs are not like other literary works. A typing or grammatical error occurring in a book may be corrected without the act having any impact upon the remainder of the work. The relationship between the various elements of a computer program is much more complex. If an error is discovered in the course of running a program, its cause may lie almost anywhere in the program. If the source of a particular error is detected and a correction made, it cannot be certain that the effects of the change will not manifest themselves in an unexpected and undesirable fashion elsewhere in the program. There is indeed a school of thought in software engineering that suggests that when errors are detected, rather than amending the program, operating procedures should be changed to avoid the conditions which it is known cause the specific error to occur.
When software is supplied to a customer, it will be in a form known as object or machine-readable code. If this were to be viewed by a user it would appear as a series, a very long series of zero's and one's. Obtaining sight of these digits will give little indication as to the manner in which the program is structured. Although it is possible for a program to be written in object code, much more programmer friendly techniques are available and almost universally utilised. A number of what are referred to as "high level" languages exist - examples are BASIC and FORTRAN. These allow programmers to write their instructions in a language which more closely resembles English although the functional nature of computer programs limits the variations in expression which are a hallmark of more traditional literary works.
Most users, of course, will be concerned only with what a program does rather than the manner in which this is accomplished. Some, however, may have different motives. The practice of reverse engineering has a lengthy history in more traditional industries and typically involves the purchase and dismantling of the products of a competitor. In the computer context, reverse engineering may involve study of the operation of a computer program in order to discover its specifications. This is essentially a process of testing and observation and might involve pressing various keys or combinations of keys in order to discover their effects. The technique known as decompilation may be used as part of this process. Normally involving the use of other computer programs to analyse the object code, the technique seeks to reproduce the original source code.
Computer programs can be divided into two broad categories - operating systems and application programs. An operating system, the best known examples are perhaps MSDOS or Microsoft Windows contains the basic instructions necessary for a computer to operate. A very simple analogy might be made with a railway system. The gauge of the track and the height and width of tunnels and bridges might be regarded as equivalent to an operating system. They set down basic parameters which must be respected by anyone wishing to build a train to operate on the system. If the track gauge is 4ft 8ins, no matter how technologically advanced an engine might be, it will be quite useless if its wheels are set seven foot apart. In the computer field, programs such as word processing and spreadsheet packages constitute the equivalents of railway engines. They work with the operating system to perform specific applications and must respect its particular requirements.
A producer intending to develop an applications package for use on a particular operating system must be aware of its functional requirements. In most instances, the information necessary will be made available by the producer of the operating system whose own commercial interests will be best served by the widest possible availability of applications to run on the system. In the event that the information is not readily available - or that it is suspected that only partial information has been made available - the attempt may be made to reverse engineer the operating system.
A second occasion for the use of reverse engineering occurs at the level of applications packages. Programs such as word processors and spreadsheets store data in a particular format. In the case of basic text, a widely used standard exists ASCII (American Standard Code for Information Interchange) The text of most word processed documents is a much more complex creature. Particular fonts, type size and line spacing will be used. Portions of the text may be printed in italics or may be emboldened or underlined. These matters are not standardised. A producer intent on developing a new word processing program, may wish to discover the codes used by rival producers so that conversion facilities may be built into the new product. From a commercial perspective, existing users are more likely to change to a new program if they can still use documents created using their existing program.
The final form of reverse engineering is the most controversial. Here, the object of the reverse engineering is to discover information about the user interface of an applications package which may then be used as the basis for the attempt to produce a substantially similar package. In early court cases on the point in the United States it was often asserted that the intent was to reproduce the "look and feel" of the original package.
Given that a lawful user cannot be prevented from using a program for its normal purpose, some aspects of reverse engineering must be considered legitimate. A user who operates the program in a normal fashion in order to study its various aspects will not infringe copyright. Subject to strict conditions, a user will also be given the right to attempt to decompile a program's object code when this is done in order to produce a further program which will be interoperable with the copyright owner's. This would apply with respect to the first and second forms of reverse engineering discussed above The right cannot be excluded by contract but will apply only where the information required has not been made "readily available" by the copyright owner. The term "readily available" appears imprecise. It would not seem to require that the information be supplied free of charge. The levying of excessive charges would obviously be incompatible with the provision but the question will arise what level is to be so considered. In most cases where interchange information is used in, for example, the word processing programs referred to above, it would appear that this is done under the terms of cross-licensing agreements between the parties involved. A second issue raises more technical questions. Producers of operating systems will normally find it in their own commercial interest to make the information available to those who wish to produce applications to run on the system. In some case, the producer of an operating system will also produce applications packages. The best known example is Microsoft. Although sufficient information concerning its operating system is made available to other producers, the systems have a number of what are referred to as "undocumented calls" and it is frequently asserted that these are used by Microsoft's own applications packages. The situation might be compared to producing a road map of the British Isles which omitted all reference to motorways. A motorist who relied totally on the map would certainly be able find a route between Glasgow and London although the journey might take considerably longer than one making use of the motorway network. Returning to the computer context, it may be queried whether the provision of incomplete information will resurrect the decompilation right. Against this, it may be noted that the legislation makes no mention of the quality of the interconnection which is to be enabled. If comparison is made with the patent system which requires that an inventor disclose details of the manner in which the invention functions, the duty here is to disclose an effective manner of performing the invention and not necessarily the optimum method. Any claim relating to the sufficiency of disclosure above and beyond that necessary to achieve interoperability might more reasonably lie under the heading of competition law.
The provisions described above apply also where the program under investigation is an applications package. Here the usage of the information acquired will be restricted significantly by the statutory prohibition against using the information to produce a program which will infringe copyright in the original. This is not a prohibition against the production of a competing program. The example was cited above of a word processing program which can import or export to or from other word processing programs. Where legal problems arise is in respect of the third form of reverse engineering described above where the attempt is made to produce a program which appears to a user to be substantially similar to the original. Copyright issues abound here irrespective of the question whether any decompilation has occurred. These will receive detailed consideration in the following section.
The purpose of copyright is to protect the expressions which are used by an author (or anyone else responsible for the creation of a protected work). The underlying ideas and concepts are not protected. The distinction between unprotected ideas and protected expressions has long been explicitly recognised in the United States. For the United Kingdom, although judicial decisions are replete with dicta to the effect that "it is trite law that there is no copyright in ideas" there is no explicit statutory recognition of this situation although the European Union Directive does provide that "... ideas and principles which underlie any element of a computer program ... are not protected under this Directive" (Article 1(2)). The only reference to this provision in the implementing regulations is a provision voiding any contractual attempt to exclude an authorised user's right to "study or test the functioning of the program in order to understand the ideas and principles which underlie any element " (Section 296A(1)(c)).
Although it may be "trite law" that ideas are not protected under the law of copyright, the task of distinguishing between ideas and the manner in which they are expressed is often a difficult one. Indeed, few works will be based upon a single idea. Rather, there will be a number of ideas ranging from the extremely general to the extremely specific at which point they may become inextricably linked with the forms of expression used. Thus the idea may occur to a person that they will write a book. Certainly, this aspect of the work cannot be protected under the law of copyright. The thought process may move on to the notion that the work is to be a legal tome and that it is to be concerned with the legal response to aspects of computer-related behaviour. Here it seems certain that we are still in the realm of unprotected ideas. The next idea may concern the main categories of material to be included and the order in which it is to be presented. The creative process may move on with ideas being expressed as to the format of particular sentences and phrases until a complete text is produced the reproduction of which will constitute a breach of copyright. On this analysis, every expression represents ideas and ideas are nothing until they are expressed. The task for a court is to determine where along the creative spectrum, the protection afforded by the law of copyright should begin.
In one of the few United Kingdom cases where copyright has been refused to a written work, the Court of Appeal held in the case of Exxon Corp. v. Exxon Insurance Consultants Ltd. ([1982] Ch 119) that the single invented word Exxon[3], could not be regarded as a literary work and that another party making use of it was not guilty of infringement of copyright. Beyond this, it is generally recognised that character names and details of the plot of a literary work may obtain copyright protection. Much publicity was recently given to the publication of a sequel to the novel "Gone With the Wind". It is suggested that one of the reasons for the appearance of the work was that the period of copyright protection in the original novel was about to expire. At that time anyone could write such a sequel making use of the character names and any other elements of the original storyline[4]. By commissioning the sequel, the estate of the original author would be able to create what was effectively a new copyright.
Although character names and storylines may be regarded as the subject of copyright, the case of Green v. Broadcasting Corp. of New Zealand ([1988] 2 NZLR 490) shows that this approach has its limits. The case concerned a television programme devised by the appellant called "Opportunity Knocks". The programme was effectively a talent contest offering new performers an opportunity to appear on television. The programme was presented by the appellant. Each show followed a particular format with acts being introduced by sponsors, the appellant who presented the show using particular catch phrases. The comparative popularity of the acts with the studio audience was measured using a device referred to as the "Clapometer" whilst a postal ballot was conducted to determine the preferences of the viewing audience. The show ran for a number of years in the United Kingdom. Some time later the appellant discovered that a show of the same name, was being produced and broadcast in New Zealand utilising much the same format as the original. An action alleging breach of copyright was raised only to be dismissed by the New Zealand courts, a decision upheld on appeal to the Privy Council. In part the decision was based upon evidential failings on the part of the appellant who was unable to present any detailed scripts pertaining to the programme. Such evidence as was submitted referred to the general format of the show. As Somers J commented in the New Zealand Court of Appeals:
The scripts, as I understand them, could not constitute a dramatic work. They could not themselves be acted or performed which I take to be the essence of such a work for they were no more than a general scheme for a proposed entertainment. (p 497)
Dismissing the appeal, the Judge concluded:
Not surprisingly he (Mr Green) feels his ideas have been appropriated. But that, I am afraid, is all that has happened. Whether taken item by item or as a whole I am of opinion that the scripts as they are inferred to be from the description given in evidence did not themselves do more than express a general idea or concept for a talent quest and hence were not the subject matter of copyright. (p 498)
Concurring, Casey J made reference to an article by an American commentator which suggested that:
Formats are thus an unusual sort of literary creation. Unlike books, they are not meant for reading. Unlike plays, they are not capable of being performed. Unlike synopses, their use entails more than the expansion of a story outline into a script. Their unique function is to provide the unifying element which makes a series attractive - if not addictive - to its viewers.
Again reference was made to the sense of grievance which might by felt by a person in Mr Green's position who considered that there had been "piracy of a good entertainment idea." (p 504) "Nevertheless", he concluded:
the overall interests of society in maintaining the free exchange and ability to develop ideas must also be considered. Any extension of the defined breadth of copyright protection must be approached with caution and is probably best left to the informed decision of the legislature. ( p 504)
It is a moot point whether the presence of a detailed script making reference to the catch phrases and features of the show might have prompted a different conclusion. In the case of Harman Pictures v. Osborne ([1967] 2 All ER 324) an interlocutory injunction was sought to prevent the making of a film on the Charge of the Light Brigade whose script was alleged to infringe copyright in a book written on the subject. Given the nature of the proceedings, the precedential value of the decision is limited but the absence of any attempt by the defendant to identify sufficient alternative sources for their work to rebut the evidence of similarity was held sufficient to justify the grant of an injunction.
A relevant aspect of this case for the present work is that the nature of both the book and the film was constrained to some extent by external factors. The charge of the light brigade was a historic reality. Reference would be expected in any work on the topic to the key characters involved and to the major events. This situation may occur in many other contexts. The information in a telephone or trade directory is factual information and, at least in respect of alphabetical listings, virtually no discretion is afforded to a person seeking to compile such a work. In the United States, the Supreme Court has held in the case of Feist v Rural (111 Sup Ct Rep 1282 (1991)) that the functional constraints identified above were such that no copyright subsisted in a telephone directory as there was insufficient originality involved in its compilation. It may be doubted whether the same decision would be reached in proceedings before a United Kingdom court. Here, the requirement that a work be original has been interpreted as requiring nothing more than that it is not copied from an original source. Indeed, in the case of Waterlow Directories Ltd. v. Reed Information Services Ltd. (The Times, 11 October 1990) copyright was held to subsist in alphabetic listings of English solicitors arranged by town.
Incidents such as that at issue in Harman Pictures indicate that the exact reproduction of a protected work is not necessary in order to establish infringement. The Copyright Designs and Patents Act, restating the provisions of earlier statutes, states specifically that the making of an adaptation of a protected work, in circumstances where there may be few if any signs of literal similarities between the two works, will constitute infringement. The basic infringing act is referred to as that of reproducing the protected work or a substantial part thereof. The concept of reproduction is broader than that of copying. By moving away from the situation where copying of a work was the sole criteria for infringement, the basis of the system rests more in the element of unfair exploitation of the work of another. One of the most frequently quoted dicta in the area is that of Petersen J in the case of University of London Press Ltd v. University Tutorial Press Ltd. ([1916] 2 Ch 601) to the effect that "what is worth copying is worth protecting". Copyright infringement will typically be established by showing that a defendant enjoyed access to a protected work and subsequently produced a work which indicated that they had unfairly exploited the work. In Harman Pictures, the allegation was that the screenplay was based excessively upon one work of reference. An even more extreme case is that of Elanco Products v. Mandops ([1980] RPC 213). Here the plaintiff company held a patent in a herbicide. In marketing the product, it supplied detailed instructions for use, these being based on its research and experiences of the product and intended to ensure its maximum effectiveness. When the patent expired it was, of course, open to anyone else to produce the herbicide and this was done by the defendant. In addition to reproducing the plaintiff's product, it also reproduced their instructions. This prompted an objection from the plaintiff and resulted in the defendant making two attempts to modify the terms of the instructions. The changes made failed to satisfy the plaintiff who instituted proceedings alleging that the defendant's instructions infringed their copyright and seeking an interlocutory injunction to prevent the continued use of the instructions pending a trial.
By the time the copyright action was instituted, the defendant's instructions bore few similarities to those originally produced by the plaintiff. Nonetheless, it was held that the copyright had been infringed, Lord Justice Goff ruling that:
It may well be that if the respondents had in fact at the start simply looked at the available information ... and from that decided what they would put in their literature and how they would express it, the appellants would at least have had considerable difficulty in bringing home any charge of infringement, even having regard to the evidence, if the results had been exceedingly similar and the selection of items had been the same. But they chose, on the evidence as it stands at the moment, to proceed by making a simple ... copy, and then they proceeded to revise it. It may well be that the result produced in that way is an infringement.
Given that a patent specification serves as a set of instructions concerning the manner in which the invention is to be performed, the decision in Elanco might be criticised as marking an unwarranted extension to the duration of protection offered to a patent holder. Whilst it must be conceded that the detailed instructions for the application of the herbicide would not be contained in the patent specification, the fact that the plaintiff had enjoyed what might have been a twenty year monopoly over the commercial exploitation of the product must have limited significantly the scope and the incentive for others to conduct research in the field. On a more literary analysis, of course, it is clear that the defendant's conduct even with respect to the third draft of the instructions, could be regarded as a form of plagiarism. A difficulty with the decision in Elanco and, as we shall see, with cases involving copyright in computer programs, may come from the attempt to apply literary criteria to functional and practical works. Significant differences do exist between the two categories of work. In particular, whilst diversity of expression is desirable in a literary context, it may have the opposite effect in the realm of product instructions. Where users have become accustomed to particular instructions for a product as potentially dangerous as a herbicide, there may be undesirable possibilities for confusion and error if the product is supplied with apparently different instructions.
Although a number of English cases considered the extent of copyright protection in software at the interlocutory level, it was only in 1993 with the case of Richardson (John) Computers Ltd. v. Flanders and Chemtec Ltd. ([1994] FSR 144), that the issues were considered in a full hearing. At issue in this case was a computer program designed for use by pharmacists. The program, which was developed to run on the then popular BBC micro-computers performed a number of tasks. Principally, when the computer was attached to a printer to would automate and simplify the task of preparing dosage instructions to be supplied with medicines. The programs other major function was to assist in stock-keeping by keeping a record of the drugs dispensed. The program was marketed by the plaintiff who had also performed a significant amount of work on the original program. Subsequently, the first defendant was employed to work on the project. It was accepted that all relevant copyrights in the work belonged to the plaintiff.
The program achieved considerable commercial success. Relationships between the plaintiff and the defendant were not so fortunate. The defendant resigned from his position in although he continued to perform some work for the plaintiff as an independent contractor for a further period of time.
With the advent of the IBM personal computer one of the plaintiff's major customers expressed interest in a version of the program capable of running on this machine and which could be sold on the Irish market. Following discussions the plaintiff determined not to proceed with the project but at his suggestion the defendant was approached and agreed to perform the work. The program was completed and was sold in Ireland under the name "Pharm-Assist". The defendant subsequently contacted the plaintiff offering him the rights to market the product in the United Kingdom. These discussions proved fruitless and the defendant proceeded to market a modified version of the Pharm-Assist program. At that stage the plaintiff initiated proceedings alleging that "Pharm-Assist" infringed copyright in his original program.
Because of the fact that the programs had been developed to run on different computers, examination of the code used would have revealed few evidences of similarities. The programs did perform the same functions and had very similar appearances when operating on their respective hardware.
In the absence of any relevant UK precedent, the court placed considerable reliance upon the United States case of Computer Associates v. Altai. (982 F 2d 693 (1992)). Here, an employee switched employment between the firms commencing work for the defendants. Unknown to his new employers who put him to work on a project, he had worked on a similar project for his previous employer and had taken a copy of the relevant source code when he left their employ. He completed the task in 4 months only for it to be discovered subsequently that about 30% of the code of the second program was a direct copy from the original. The defendants accepted that this constituted a breach of copyright.
Following discovery of its employee's conduct, the defendant sought to produce a legitimate program which performed the same functions. To achieve this a new team of programmers was put together, selected because none of them had any previous experience on the project. The programmers were given a description of the facilities which were to be developed and were required to produce the programs necessary. The project took some 6 months to complete and the revised product was put on the market. In circumstances not dissimilar to those at issue in the Elanco case discussed above, these subsequent attempts to vitiate an original act of copying did not satisfy the plaintiffs who alleged that the revised version also infringed the copyright in the non-literal aspects of their program.
This argument was rejected by the District Court, a conclusion upheld by the Court of Appeals for the Second Circuit. Although it was recognised that copyright protection had to extend beyond the strictly textual form in order to avoid the situation where a plagiarist could avoid liability by making immaterial variations, it was held that the question whether infringement had occurred could be answered only after making a detailed analysis of the circumstances at issue. In particular, the court recommended the utilisation of a three stage test - referred to as the `abstraction-filtration-comparison test' in order both to identify the elements of a work which were entitled to protection and to determine whether infringement had occurred.
The first stage of the test involves the court in its own form of reverse engineering. Beginning with the complete program and its code the process should work back to the basic idea producing what might be considered a chart of the development process.
After the identification of the elements of the original program that are to form the basis of the decision, the filtration stage seeks to determine why particular elements are found in the original program and to filter out or discard those whose presence is due to circumstances which would serve to deny copyright protection. As identified by the courts in Computer Associates these might include aspects of the program whose form is dictated by reason of efficiency. In many cases, there may be a limited number of ways in which a task may be performed effectively. Also identified as factors which might be filtered out were elements of the program whose form was dictated by external factors. Particular reference was made to:
The final stage of the process entails the comparison of the remaining elements of the original program, what was referred to as the "golden nugget" of protected material against the features of the allegedly infringing program with a view to determining whether any similarities should be regarded as involving a substantial part of the original work and as having been produced as a result of copying rather than by independent creative activities.
The result of the application of this test in Computer Associates was a finding that no infringement had occurred. In Richardson, Ferris J sought to apply a modified and less technical version in which, after determining the eligibility of the program as a whole for copyright protection - a fairly straightforward matter - the court proceeded to answer the question "whether any similarity attributable to copying which is to be found in the defendants program amounts to the copying of a substantial part of the plaintiff's program?"
The first stage in this process involved an examination of the functions performed by the two programs and the manner in which these were carried out. This revealed 17 points of similarity concerning the functions performed by the two programs and the manner in which these took place. When these were subjected to more detailed examination, however, 6 of the points of similarity were considered explicable by reasons other than copying. To give perhaps the simplest illustration, both programs used the same format of date with numbers for day, month and year. Other similarities related to the manner in which the quantity to be prescribed was dealt with before the drug itself was identified. Beyond the fact that only two alternatives were available, evidence was presented indicating substantial commercial demand for the adoption of the quantity first option.
In respect of a further 8 similarities, the view was taken either that no copying had occurred or that the similarities did not constitute a substantial part of the original program. An example of the latter situation concerned the inclusion of a feature in both programs whereby the message "operation successful" appeared on the computer monitor accompanied by a double beep. Such a feature was considered lacking in "originality and cannot have required any significant skill or effort to devise it."
In the case of only 3 of the similarities was the conclusion reached that copying had occurred of a substantial part of the original program. One of these related to the procedures whereby a user could amend entries so as to correct any errors which had been made. Although the need for such a facility would exist in any program, the original routines were unusual in their nature with features that could be regarded as inefficient and misleading to the user. The fact that the later program functioned in the same manner was considered indicative of the fact that it had been copied from the original. It appears to be a feature of cases in this area that similarities of mistakes rather than of valuable features is more damaging to an alleged copyist. Overall, however, such copying as was established in this case was described as constituting "a fairly minor infringement in a few limited respects."
Allegations of copyright were again before the High Court in the case of Ibcos Computers v. Barclays Mercantile Highland Finance and Another ([1994] FSR 275) Again there was a background of the major defendant having worked for the plaintiff on the development of a software product intended for use by agricultural dealers which was marketed under the name ADS. On leaving its employment, he developed a further and competing product which was marketed under the name of Unicorn. The plaintiff alleged that sufficient features of this were copied from the original to constitute an infringement of copyright.
In determining the criteria which would be applied in determining the question whether infringement had occurred, Jacobs J was somewhat critical of the extensive references to the United States decision in Computer Associates and warned against "overcitation of US authority based on a statute different from ours." The approach to be adopted was for the court to determine whether there was a sufficient degree of similarity between the two works which, coupled with evidence of access to the original work, would establish an inference of copying. The onus would then switch to the defendant to establish that the similarities were explicable by causes other than copying. Evidence that "functional necessity" served to narrow the range of options open to the defendant would be relevant. Trivial items may well provide the most eloquent testimony. As was said in Bilhofer v. Dixon ([1990] FSR 105):
It is the resemblances in inessentials, the small, redundant, even mistaken elements of the copyright work which carry the greatest weight. This is because they are the least likely to have been the result of independent design.
In the present case, evidence was presented that the same words were mis-spelled in the same manner, the same headings were used in the two programs and both shared the same bit of code which served no useful purpose for the functioning of the program. Beyond this, there were considerable similarities at the level of the code itself. In respect of one element of the programs it was held that:
... there are 22 identical variables, 8 identical labels, 1 identical remark, 31 identical code lines and one identical redundant variable. This to my mind plainly indicates copying and enough in itself to constitute a significant part.
The court recognised in Ibcos that copyright protection must extend beyond the literal aspects of the program code to aspects of "program structure" and "design features". In the case of the former element, it was held that copyright subsisted in the compilation of individual programs which made up the ADS system. Although some differences existed between ADS and Unicorn it was held that the defendant had taken "as his starting point the ADS set and that set remains substantially in Unicorn." Although the two programs had a different visual appearance and it was recognised that "Unicorn is undoubtedly to the user a much friendlier program than ADS was at the time", the defendant, it was held had taken "shortcuts by starting with ADS and making considerable additions and modifications".
Comparison of the decisions in Richardson and Ibcos causes perhaps more confusion than enlightenment. In its essentials, Ibcos may be considered a relatively straightforward case of copyright infringement although a number of the surrounding circumstances raise more difficult issues. The fact that the two programs used the same operating system and were designed to run on the same types of computers facilitated both direct copying and its detection. In Richardson, the move from the BBC to the IBM system meant that there was far less opportunity for, and evidence of, direct copying. Whilst each case has to be determined on its own merits, the finding of "fairly minor" copyright infringement in Richardson is to be contrasted with much more extensive findings in Ibcos.
It has always been a feature of copyright that protection has extended beyond the purely literal aspects of a work. Any other approach would allow a party to make the most minor changes in order to escape liability. Without exception, the cases which have been dealt with before the United Kingdom courts, whether at interlocutory level or at the stage of a trial, have been grounded in the situation where there has been a prior relationship between the parties. In most cases this has involved a contract of employment. In these situations there has been access to the most detailed elements of the original program. Ordinary users will obtain a much less detailed picture of the program. It may be possible for such a party to decompile elements of the object code so as obtain access to the source code. Where the goal is to produce a competing rather than an inter-operable product, such an approach may be unnecessary. Where the desire is to replicate elements of the visual appearance of the original program, the necessary information can be acquired merely to studying the program in operation. The writing of the enabling code will then be a comparatively simple matter. In Computer Associates, it took only 2 months longer for the second, clean, version of the software to be produced than was occupied in the making of the original infringing version.
The most intractable problem in applying principles of literal copyright to computer programs is that these possess both passive and active aspects. Even where the same computer language is used there will be a very great number of ways in which even the simplest function may be implemented. To give a mathematical analogy, whilst 2+3=5, so does 1+4 or 8-3. The range of permutations is almost infinite. Protecting software as a literary work is akin to protecting a particular formulation of a question when it is the answer which is all-important.
Two further factors may also be identified as complicating matters yet further. First, the operation of the copyright system may be criticised as paying inadequate regard to the interests of computer users. The acquirer of a computer program will require to invest time and effort in learning how to make use of it. If a change is made to another program, the learning process will have to be commenced anew unless the features of the two programs are similar. Whilst the interests of the readers of literary works are best served by the availability of the most diverse forms of expression possible, different considerations apply with functional objects such as software. The consequence of strong copyright protection for computer programs may be to tie a customer into a relationship with the original producer. In the case of British Leyland Motor Corp. Ltd. v. Armstong Patents Co. Ltd.([1986] 1 AllER 850), the House of Lords held that held that the plaintiffs were not to be permitted to use their copyright in engineering drawings in order to maintain a monopoly in respect of spare parts for their vehicles. The right of car owners to maintain their vehicles was to take precedence over the producer's intellectual property rights. Although the situations are not precisely comparable in that the purchaser of a computer program is likely to be seeking its replacement rather than its repair, some regard should be paid to the user's interests.
The second factor concerns the implications of intellectual property rights in an employment relationship. Where individuals are employed as software programmers as was the case in Richardson and Ibcos, it may be extremely difficult to distinguish between their own expertise and the proprietary information belonging to the employer. Most programmers when faced with a particular task will not attempt to create new code. Rather they will seek to reuse existing modules. These may be taken either from their own portfolio of work and experience or from reference manuals. Too extensive an application of intellectual property rights may inhibit significantly the future career prospects of employees.
As was stated at the outset of this chapter, copyright has proved a flexible legal tool. Even the most flexible structures have limits beyond which they cannot be stretched without both causing and suffering damage and it may be that these limits have been reached. In Computer Associates, the Court of Appeals commented that:
Generally, we think that copyright registration - with its indiscriminating availability - is not ideally suited to deal with the highly dynamic technology of computer science. Thus far, many of the decisions in the area reflect the court's attempt to fit the proverbial square peg in a round hole.
and went on to suggest that the patent system might provide a more appropriate regime with protection being dependent upon "exacting up-front novelty and non-obviousness requirements". Considerable concern has however been expressed at the ability of patent examiners to make informed decisions as to novelty and obviousness in the face of such a rapidly changing subject as computer science. It may be that the application of copyright should be restricted to the situation where there is direct copying of a program. This may well encompass most instances of software piracy and many aspects of the decision in Ibcos. For the case where a similar product involves independent effort the need may be for a sui generis system, perhaps involving compulsory licences, to ensure that the original creator receives a fair reward for the skill and labour expended whilst allowing users to find other sources of supply.
A final reference may be made to the suitability of the legal system for resolving disputes concerned with intellectual property issues in software. In many instances the litigation is prolonged. The trial in Computer Associates lasted for 8 days, whilst the hearings in the English cases of Richardson and Ibcos occupied 24 and 13 days respectively. The expenses incurred in such lengthy proceedings must inhibit parties taking disputes to the courts whilst a lack of case law furthers a perception of uncertainty as to the legal position. Even in the cases discussed above, the time necessitated to bring the dispute before the courts entails that the decision relates to technology which may be functionally obsolete. The original program in Richardson was written for the BBC micro-computer in the early 1980's. This computer had a maximum memory of 96 kilobytes of data. This text has been written on a fairly basic computer of the 1990's which has a memory of 12 megabytes. The new machine is about 125 times more powerful than the old one. Similarly, in Ibcos the initial version of the programs at issue was written in 1979. Although court decisions may allocate financial responsibility for past events, it seems unlikely that they can have significant impact upon present and future activities.