UK: Electronic Databases: Protecting Your Investment: An Analysis Of The Legal Rights In Electronic Databases Under UK Law

1. Introduction

Electronic databases are an important part of the information economy.¹ The Internet, and the development of on-line services as an effective business tool, has meant that electronic databases are now one of the key platforms for the delivery of information and content. As such, electronic databases are now viewed as a valuable business asset, with database producers and owners keen to ensure that the commercial value inherent in their databases is properly protected.

In 1996, the European Commission adopted directive 96/9/EC on the legal protection of databases ("Database Directive").² This was implemented in the UK by the Copyright and Rights in Databases Regulations 1997 ("UK Regulations"). One of the principal objectives of the Database Directive was to promote investment in the creation of databases throughout the EU. This was to be achieved by the introduction of a new sui generis database right.³

This sui generis right built on existing legal and equitable rights which were capable of applying to electronic databases under UK law – namely, copyright, confidentiality and contract.

The European Court of Justice, in a series of judgments handed down in November 2004, has ruled on the extent and scope of the protection offered by database right.⁴ As discussed briefly below, these rulings have significantly curtailed the scope of that right. As a consequence, the other legal rights which can apply to electronic databases are likely to assume a more important role in allowing database producers to protect and exploit the commercial value inherent in their databases.

The UK courts’ recent decision in Navitaire⁵ has important implications for the copyright protection afforded to database application software (and computer programs generally) and, in particular to the copyright protection afforded to "look and feel" elements of those software programs.

This article examines the legal and equitable rights that are capable of applying to electronic databases, and the role they can play in the protection and commercialisation of such databases.

An electronic database typically consists of an operating platform (the hardware and operating system software on which the database resides), the database application software (the family of programs used to manipulate the database), the data, and the database itself (by which we mean the aggregated collection of data). In addition, between the database software application layer and the data sits the database architecture, comprising the structure and schema of the database. Each of these components can attract a range of different rights under UK law.

We are referring here to the database in its narrowest sense – the aggregated collection of data – as opposed to any of the other components which make up an electronic database. There are two key rights which can subsist in the database: copyright and database right. In general terms, copyright protects the intellectual skill and labour in the selection and arrangement of the database; database right protects the investment needed to obtain, verify and present the contents of the database. The two rights are independent, with different conditions required for subsistence, different durations and different prohibited acts. Whether a database attracts one or both of these rights will depend on the nature of the database and the date of its creation.

3.1.1 Copyright in Data Compilations

Prior to the implementation of the Database Directive, databases were capable of copyright protection under UK law as "tables and compilations".⁶ Provided a data compilation was "original", copyright would subsist. The originality threshold was the one that applied, and which continues to apply, to most other literary works: a sufficient degree of skill, labour and judgement must have gone into the creation of those works. There is no statutory definition of originality in this context, but case law has established that the threshold is low and, in general terms, can be regarded as having been met as long as the compilation has not been copied from another work, and that more than "negligible" or "trivial" effort has gone into its creation. Under this originality test, copyright has been found to subsist in a wide range of data compilations, including telephone directories, football fixture lists, and television programme listings. Most electronic databases would, therefore, have attracted copyright prior to implementation of the Database Directive.

The Database Directive changed this position in two important ways. First, it introduced into UK law a statutory definition of database: "a collection of independent works, data or other materials which (a) are arranged in a systematic or methodical way, and (b) are individually accessible by electronic or other means".⁷ Second, it introduced a new originality threshold which databases falling within this statutory definition must meet in order to attract copyright – these databases must "by reason of the selection or arrangement of [their contents,] constitute the author’s own intellectual creation".⁸ These changes apply to all databases created on or after 27 March 1997. Databases created before that date, even if they fall within the statutory definition, will be subject to the traditional originality threshold.

The new originality test introduced by the Database Directive is significant for two reasons: first, it focuses on the intellectual creation invested in the selection and arrangement of the database contents – any skill and labour invested in gathering together the database contents are irrelevant; second, this originality threshold is almost certainly higher than the traditional originality test which applies to other literary works, in that it appears to require some subjective and qualitative contribution from the author. Although there is no further guidance in the Database Directive as to what is required under this test, the travaux preparatoires to the Database Directive support the interpretation that it is a higher test than the traditional originality test under UK law.⁹ The Commission has also more recently confirmed its view that the intellectual creation threshold is higher than the traditional originality threshold which applied in the UK and other common law countries.¹⁰ This interpretation, if correct, will have significant implications for many databases including, in particular, computer-generated databases. If those databases have been created in circumstances where there has been no human intervention, it seems unlikely that they will attract copyright protection under the new originality threshold.

If a collection of data (or other materials) meets the statutory definition, but does not meet the new originality threshold, it will not attract copyright protection, but it may still attract database right. This is significant: database right lasts for 15 years from the end of the calendar year in which the making of the database was completed; copyright protection lasts for 70 years from the end of the calendar year in which the author of the database died. (In this respect, it is worth noting, however, that Laddie J’s judgement in William Hill¹¹ at first instance on the extension of database right for dynamic databases is still good law in the UK).¹²

If, on the other hand, a collection of data does not meet the statutory definition, it may still attract copyright under UK law as a data compilation or table, provided it satisfies the traditional originality test.

Database right will subsist in a database falling within the statutory definition if there has been "substantial" investment in obtaining, verifying or presenting the contents of the database.¹³ This investment can be qualitative or quantitative. The right is infringed if a person, without the owner’s consent, "extracts or re-utilises all or a substantial part of the contents" or carries out "repeated and systematic extraction/re-utilisation of insubstantial parts".¹⁴ "Extraction" involves transferring the contents to another medium; re-utilisation is making available to the public by any means.

The ECJ has delivered its first judgments on the scope of database right, and the interpretation of the requirement that, for database right to subsist there must have been substantial investment in obtaining, verifying or presenting the contents of the database.¹⁵ These judgments, and their implications, have been considered in detail in previous issues of this journal¹⁶ and a detailed analysis is outside the scope of this article.

For present purposes, it is sufficient to note that the ECJ has interpreted the above requirement to mean that the substantial investment required for database right to subsist must be applied to already existing materials: if there has been substantial investment only in the creation of the material which is contained in the database (and no substantial investment in its verification or presentation) then database right will not subsist.

These rulings have significant implications for producers of single-source databases, which create and present the data and information comprised in their databases, and for fast-moving databases where the act of creating the contents of the database is, in practice, inseparable from the act of creating the database itself: in either case, those databases are unlikely to attract database right.

Although, in practical terms, these programs can be considered part of the electronic database, for legal purposes, neither will attract protection under the Database Directive: the Directive states that "computer programs used in the making operation of databases accessible by electronic means" are not protected.¹⁷

These programs are, of course, capable of attracting copyright protection as literary works under UK law.¹⁸ In addition, a program’s preparatory design materials will be capable of copyright protection, both as literary works in their own right, and as part of the computer programs to which they relate.¹⁹ (As discussed below, in light of recent UK case law in this area, this latter point may well assume an increasingly important role in situations where there has been copying of non-literal elements of a computer program.)

Finally, it is to be noted that computer programs which are made up of a series of sub-programs, routines and sub-routines have been held to be copyright protected as compilations under UK law.²⁰

Where there has been copying of a program’s code, the analysis under UK law is relatively straightforward: provided the program copied is "original", (applying UK law’s traditional originality threshold which, as we have already noted, is low), copyright will be infringed where there has been "substantial" copying (unless one of the statutory permitted exceptions can be relied on). Substantiality, in this context, is a qualitative rather than quantitative test. It is not to be judged against whether the program would work without the code in question, or by the amount of use the program makes of the code in question, but in light of the degree of skill and labour which went into the design and coding of the piece of code which has been copied.²¹ Where there has been "over borrowing" of the skill, labour and judgement which went into the work in question, a substantial part of that work will have been copied.²²

The legal analysis is less straightforward where there has been no copying of program code, and copying has taken place at a more abstract level, for example, where two programs behave or function in a similar way, or where the program outputs are substantially similar. "Non-literal" copying of this nature often includes copying of user commands and interfaces and other aspects of a program’s "look and feel".

Non-literal copying can be particularly relevant in the context of database migrations. Typically, the migration process will involve the creation of an interim program into which data from the existing database is exported, before being migrated to the new database. Frequently, the interim program (and, in some cases, the new database) will reproduce at least some of the structure and architecture of the original. In some cases (particularly where the new database is designed specifically for the end-user), the end-user will want to migrate to a database that is functionally similar to the original.

It is settled law in the UK that non-literal elements of a computer program are capable of copyright protection. In John Richardson v Flanders,²³ it was confirmed that the structure of a computer program was copyright-protectable. In Cantor Fitzgerald, Pumfrey J acknowledged that a program’s "architecture" was capable of protection if a substantial part of the programmer’s skill, labour and judgment has gone into it.

Copying of non-literal elements of a computer program has most recently been considered by the UK courts in Navitaire.²⁴ The case is of particular relevance to this article in that it involved alleged infringement of copyright in an electronic database. It provides a good illustration of the challenges in applying UK copyright law in cases where non-literal copying has been alleged.

Navitaire, Inc was the owner of OpenRes, a ticketless airline reservation system, which it licensed to customers in the airline industry, including easyJet. EasyJet (through a third party software developer) developed a similar system called eRes. Navitaire alleged that easyJet had infringed its copyright in OpenRes, both during the migration process, and in creating a final database which was similar in function and appearance to OpenRes. It was accepted that the defendants had no access to, and therefore did not copy, the source code in OpenRes. Instead, it was alleged that certain non-literal elements of OpenRes had been copied including the user interfaces (comprising the user command sets and the layout of certain of the GUI (graphical user interface) screens), and what Navitaire described as the system’s "business logic" (essentially, its functional structure). Each of these claims was the subject of a detailed analysis by the court.²⁵

The OpenRes database was interrogated by users, using various commands. These included what Navitaire described as "simple" commands (individual command names that had only one outcome), and "complex" commands (commands which had a "syntax" and which could, therefore, have more than one outcome depending on their combination).²⁶ EasyJet’s eRes database used very similar sets of commands.

Navitaire argued that (i) the individual and complex commands were each copyright works in their own right and (ii) the collection of commands as a whole attracted copyright as a compilation. The trial judge, Pumfrey J, rejected both arguments.

Relying on the court’s decision in Exxon,²⁷ he held that individual command words and letters were too simple to possess the necessary qualities to be considered literary works. The complex commands, on the other hand, failed to attract copyright protection on the grounds that they were not written (in the sense that their syntax was not "stated" anywhere in the OpenRes source code) and, therefore, were not literary works.

While Pumfrey J’s conclusions in relation to the simple commands are understandable and uncontroversial, those relating to the complex commands are less so. As noted earlier in this article, it is settled law in the UK that non-literal elements of a computer program (such as structure and architecture) are capable of copyright protection. This is the case even where those non-literal elements are not "stated" in the program’s code. In fact, Pumfrey J, in his judgment, recognised that this was not a particularly satisfactory way to resolve the issue, not least because it failed to address the situation where the commands’ syntax had been stated in the source code. In his view, even if the syntax was stated in the source code, the commands would not be copyright protected. This was because they were, in his view, a form of computer programming language and, as such, were not capable of copyright protection under the Software Directive.²⁸

Recital 14 of the Software Directive states that "…to the extent that logic, algorithms and programming languages comprise ideas and principles those ideas and principles are not protected under this Directive". Art 1(2) states that "ideas and principles which underlie any element of a computer program, including those which underlie its interfaces, are not protected by copyright under this Directive". It is to be noted, however, that neither Recital 14 nor Art 1(2) categorically excludes computer languages from copyright protection. This lack of certainty was acknowledged by Pumfrey J, who said that, ultimately, the position would need to be clarified by the ECJ.

Finally, in this respect, Navitaire’s claim that the collection of commands as a whole attracted copyright as a compilation also failed. This was on the grounds that they there had been no overall design in their creation and that each command had been added in an ad-hoc manner over time.

The judgment in Navitaire would, therefore, on its face, appear to rule out copyright protection for a computer program’s command sets. However, two aspects of the judgment leave some scope for future argument that command codes should attract copyright protection (at least in some circumstances). First, while the Software Directive appears to imply that computer programming languages are not copyright-protectable, as noted above, the position is not without doubt. Likewise, Pumfrey J’s finding that command sets are a form of computer language is also open to question. Indeed, Pumfrey J gave Navitaire leave to appeal on his legal findings on these points (among others). Unless and until the ECJ rules on this issue, there is still scope for argument that computer languages are capable of copyright protection where they are "expression" rather than "idea".

Second, and perhaps more importantly, Pumfrey J’s judgment suggests that, where command codes are compiled in a structured manner and are part of an overall methodical design, they may attract copyright protection as a compilation. Detailing the command sets and their overall design in the program’s documentation is likely to assist any argument in this respect.

Navitaire’s claim for copyright infringement of the user interface screens was based on similarities in the look and feel of the programs’ interfaces. The court distinguished between those screens which were purely character-based, and those which made up the graphical user interface. As noted above, the Software Directive precludes ideas underlying computer program interfaces from copyright protection.²⁹ However, this applies only to computer programs as literary works. Therefore, the character-based screens, as literary works which amounted to ideas underlying the program’s interfaces, were not copyright protected. The GUI screens, on the other hand, were artistic works: as such, the Software Directive did not apply and they were copyright protected. In copying elements of those GUI screens, the defendants had infringed copyright.

Navitaire also claimed that, because OpenRes and eRes behaved in a similar manner (in that they gave similar outputs, based upon similar sequences of inputs), the "business logic" or "functional structure" of OpenRes had been copied (in much the same way as the plot of a book can be copied without any copying of the words).

This argument also failed. Pumfrey J considered that the "business function" of a computer program was an idea too abstract and general to attract copyright. Although the business function of the program was envisaged by the authors as the end result of the program, that did not constitute sufficient skill and labour to pass the originality threshold required under UK law. This conclusion was, in the court’s view, consistent with the policy of the Software Directive to exclude computer languages and the ideas underlying a program’s interfaces from protection: "it should not be possible to circumvent those exclusions by seeking to identify some overall function or functions that it is the sole purpose of the interface to invoke". To allow the business function of a program to attract copyright protection as a literary work was "an unjustifiable extension of copyright protection".

Pumfrey J’s decision in Navitaire has subsequently been confirmed in Nova Productions,³⁰ a case which also involved allegations of infringement of non-literal elements of a computer program. Again, the trail judge, rejected claims that the "look and feel" of the program had been infringed, on the grounds that the elements alleged to infringe were at such a level of abstraction as to be general ideas, and thus were not capable of copyright protection by virtue of Art 1(2) of the Software Directive.

The message from the decisions in Navitaire and Nova Productions is clear. Where there has been copying of nothing more than a program’s outputs and/or elements of the functional structure of a computer program, it will be very difficult for program owners to successfully claim copyright infringement. It should be noted, however, that in both these cases, the claimants were unable to show that the functional structure of their programs had been documented in the preparatory design materials. Although, the claimant in Nova Productions raised this as an argument, on the facts it failed. The trial judge in that case expressly declined to comment on whether Nova Productions’ copyright infringement claim would have succeeded had the relevant elements of the program’s functional structure been documented in the preparatory design materials. These judgments would, therefore, appear to leave open the possibility of a successful copyright infringement claim where the claimant can show that the functional elements which have been copied have been embodied in sufficient detail in the program’s preparatory design materials.

The architecture of a database comprises its structure and schema (the creation scripts which defines the tables and fields within the database and the relationship between those fields and tables).

The structure of a database is capable of copyright protection by virtue of the Database Directive: Recital 14 states that the copyright protection afforded to a database should extend to the structure of the database. Although Section 3A of the CDPA does not contain language to that effect, the UK courts will interpret Section 3A in a manner consistent with the Directive.

Again, this is an issue of particular relevance in the context of database migrations, where it is often necessary (or at least desirable) to reproduce the structure of the original database, either in the interim programs or the new database itself.

If the database structure is reproduced in this way, then copyright is likely to be infringed unless a permitted exception applies. Under UK law, the most likely possible permitted exception is that set out in Section 50D of the CDPA. Under this section, it is not an infringement of database copyright for a lawful user to do "anything necessary" to use access and use the contents of the database. Contract terms which seek to prevent this are void.

The scope and effect of Section 50D was considered by the court in Navitaire. Navitaire alleged that the defendants had infringed its copyright in the OpenRes database structure through its reproduction in eRes (and the interim databases). Pumfrey J considered that "necessary" in the context of Section 50D should be construed "so as to exclude the merely desirable or convenient, but not so as to require the merely absurd".

To facilitate the migration process, easyJet had provided Bulletproof with screen shots of OpenRes, a tape of the complete OpenRes database, and hyperlink access to the OpenRes database.

Pumfrey J held that the use of the screen shots was permitted by Section 50D, since this was necessary to find out what data was being held. However, in providing Bulletproof with a copy of the OpenRes database, copyright in that database had been infringed. Section 50D could not be relied upon in these circumstances, since the data migration could have been achieved through the use of alternative techniques (a neutral file format for example)³¹ which did not involve copying the OpenRes database. Providing Bulletproof with hyperlink access to OpenRes was also an infringement since this allowed Bulletproof to access the database fields even when there was no access to data.

Thus, following the judgment in Navitaire, it would appear that, where migration can be achieved through the use of a neutral file format technique, then even though it is likely to be more time-consuming and less convenient than other techniques, directly accessing and copying the database will not be exempted by Section 50D.

Under UK law, there are no property rights in "raw" data or information per se. Information and data can, however, be protected by rights of confidence and by contract.

A successful claim for breach of confidence, under UK law, typically requires three components: (i) the information in question must be "confidential"; (ii) it must have been imparted in circumstances in which an obligation of confidence arises; and (iii) there must have been a breach of that confidence by the person receiving the information to the detriment of the disclosing party.

Two important points are to be noted in respect of these requirements. First, "confidential", in this sense means that the information in question must not be "publicly accessible". The fact that the underlying information contained in an electronic database is in the public domain, will not, in itself, preclude the law of confidence from applying: it is well established, under UK law, in a line of cases starting with Albert (Prince) v. Strange,³² that, where information in its aggregated form is not publicly available, that information (in its aggregated form) is capable of protection through the law of confidence, even though the underlying information may itself be in the public domain.

Second, in the context of many databases (particularly those that are publicly accessible) it will not be possible to imply a special relationship which would give rise to an obligation of confidence. For database producers to rely on the law of confidence in respect of any unauthorised use or disclosure of data comprised within a database, it will be necessary to ensure that anyone accessing that database is required to accept licence terms which contain appropriate confidentiality notices and restrictions on re-use and disclosure of the information and data contained within the database.

As an extension of the last point, it is, of course, possible under UK law, to restrict the use and disclosure of information accessed via a database by contract.

It is settled law in the UK that a person is entitled, by agreement, to make a charge for access to, and use of, its information, whether or not that person has any intellectual property rights in that information. This has most recently been confirmed in Victor Chandler.³³ The case involved the British Horse Racing Board and Victor Chandler, a bookmaker based in Gibraltar. Victor Chandler claimed that, as a result of the ECJ’s ruling in BHB v. William Hill,³⁴ the data licence it had entered into with the British Horseracing Board for access to and use of pre-race data was void, and that it should not have to pay for access to and use of that data. The trial judge, Laddie J, dismissed Victor Chandler’s claim, holding that licence terms which restricted Victor Chandler’s use of the pre-race data and which entitled BHB to make a charge for Victor Chandler’s use of that data, were valid.

The ECJ’s and UK court’s recent judgments make it more difficult for database owners and producers to assert statutory intellectual property rights in databases and the application software used to manipulate those databases.

The UK courts’ recent judgments in Navitaire and Nova Productions underscores the difficulties that claimants will face in attempting to assert copyright in the functional structure and other "look and feel" elements of database application software. To enhance the prospects of asserting copyright in the functional structure of a computer program, program owners should ensure that those elements which they wish to protect are properly documented in the program’s preparatory design materials.

The ECJ’s recent judgments have significantly curtailed the scope of the sui generis database right by distinguishing between resources used in the creation of the contents of a database and those used to obtain, verify and present that content in the database. Only the latter investment will be taken into account in determining whether a database attracts database right. This will be of particular importance for producers of single-source databases, who both create the data and establish the database and for fast-moving databases where the act of creating the data is inseparable from creation of the database.

As a result, rights of confidence and contract rights assume a more significant role in allowing database owners to control access to their databases. Database producers should ensure that all users of their databases are required to agree to legally binding contract or licence terms, governing the circumstances in which their databases can be accessed and used. Those terms should include express confidentiality notices and restrictions on re-use and disclosure of the information and data contained within the database.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.