Ian Jeffrey thinks that database right is proving to be even more important than predicted. This article looks at the scope of the right and applies it in a number of commonly occurring scenarios.
Database right is now headline news again. At the time of its introduction into UK law in January 1998, there was a certain amount of debate as to its precise effect. As is so often the case, commentators concluded that a good deal would turn on the court’s interpretation of the new legislation. With those sage observations on record, it was left mainly to commercial practitioners to weave the new rights and concepts into their agreements, knowing that only litigation would bring the legislation back into public focus. All that has now happened, and rather more besides. The decisions of various European courts on database right have demonstrated the importance of this statutory backwater for the online publishing world, not least the recent decisions of Laddie J and subsequently the Court of Appeal in the British Horseracing Board case. In addition, the cases have highlighted the need for clarification of various issues. This article considers the current scope of database right as a guardian of online content and highlights issues of particular significance to those involved in publishing content online.
The Legislation
Readers will be familiar with Council Directive 96/9/EC of 11 March 1996 on the Legal Protection of Databases (the Database Directive) and with the Copyright and Rights in Databases Regulations 1997 (the Database Regulations).
The Database Regulations gave effect to the Database Directive in the United Kingdom by making certain amendments to the Copyright Designs and Patents Act 1988 (the CDPA) and by providing a statutory framework for the new database right. The adjustments to the CDPA added databases as a class of literary work but imposed a special test of originality before a database could qualify for copyright protection. To qualify for protection a database had, by reason of the selection or arrangement of its contents, to be the author’s ‘own intellectual creation’.
Whether or not a database was protected by copyright, it stood to enjoy the new database right under Part III of the Database Regulations if there had been ‘substantial investment’ in developing it. The right would be infringed by a person who extracted or re-utilised all or a substantial part of the database, either on one occasion or systematically over a period of time. Database right appeared to be shorter in duration than copyright, but the entire period of protection (15 years) would effectively re-start if the contents of the database changed substantially – as of course they very often will.
British Horseracing Board v William Hill
As the first major UK decision on database right, brief summaries of the facts and decisions are worth setting out.
At first instance,1 Laddie J held that William Hill had infringed BHB’s rights in its comprehensive database of horse racing information. The BHB licensed use of information in its database in various ways, in particular making a ‘declaration feed’ available to various subscribers, through a subsidiary, and secondly by permitting an intermediary to provide a ‘raw data feed’ to certain subscribers. William Hill had access to information held on the BHB database through these intermediaries. When William Hill launched its online betting operations, it made information about certain races available on its Web site. BHB claimed this information could only have come from its databases and that its use was an infringement of its database rights.
In addition to a failed challenge to BHB’s factual argument that its database was the ultimate source of the data used by WH, the bookmaker took a whole series of points in relation to the nature of the right and the question of what constituted an act of infringement. Among the most interesting of the arguments advanced were the following.
-
First, that since database right cannot be used to protect the individual items of underlying data as such, it follows that there must be some special quality attached to the collection of data which leads to it being protected (this quality was rather awkwardly labelled ‘database-ness’ during argument). Unless advantage of that special quality were taken, the right could not, so the argument went, be infringed. Laddie J dismissed this argument on the basis that it confused the features of form which gave rise to the existence of the right and the features of content or investment which were to be protected once the content in the database existed. In other words (borrowing an approach for these purposes from the design community), it is the logical structure of the database which qualifies its population for protection.
-
The next key point of William Hill’s argument was that what had been taken was not a substantial part of the database. Save in the one respect referred to below, this raised the type of argument relatively familiar to copyright lawyers and the judge was able to conclude fairly happily that a substantial part, considered qualitatively, had been taken.
-
More crucially, WH submitted that the taking of information did not amount to the ‘extraction’ or ‘re-utilisation’ of content. It was argued that the indirect capture of data from the database could not be extraction since the concept of extraction incorporated the concept of removal rather than the concept of replication. Put simply, it meant cut rather than copy. This argument was dismissed.
-
WH’s argument on re-utilisation turned on the submission that re-utilisation, requiring an act of making available to the public, could only effectively be established where the information was made available to the public for the first time. This argument too was dismissed.
-
Finally, WH sought to take advantage of a rather ingenious argument along the following lines. Although repeated acts of extraction or re-utilisation of an insubstantial part of a database can, when taken together, amount to extraction or re-utilisation of a substantial part, the fact that the database changed in composition so often meant the minor acts of taking were carried out in relation to different databases and so could not be aggregated to establish infringement. Although he considered it an attractive argument (which at first blush it is), the judge doubted that it could be right.
Laddie J’s argument therefore firmly favoured the content owner. However, the Court of Appeal has now directed2 certain questions arising from his judgment be referred to the European Court of Justice on the basis that the Database Directive requires clarification and that the clarification required is important. Although some 15 questions were proposed for reference by William Hill’s counsel, the Court of Appeal approved the following topics for reference (subject to agreement over the precise framing of the questions):
-
Whether ‘database-ness’ is an essential quality for any part of the contents of the database.
-
Whether information (in this case in the form of lists of runners) created at the same time as they are published can be a relevant part of the contents of the database.
-
Whether ‘extraction’ or ‘re-utilisation’ involves having access to the database or a copy of it.
-
Whether ‘extraction’ or ‘re-utilisation’ extends to a person who is a subscriber to a service provided by a licensee of the database right owner and who thereby receives part of the contents of the database and who makes available that part of the contents to the public in the course of his business.
-
Where there is a constantly updated database, is there a new database separate from the previous database whenever any substantial change has occurred.
As the Claimant’s counsel observed in the Court of Appeal, references to the ECJ can take a great deal of time and we will therefore be left on tenterhooks for some time as to the answers to these significant questions. In the meantime, the online publishing community must deal with these uncertainties as best it can. The balance of this article considers the applicability of the right to some core publishing technologies and to some common industry practices.
Database right and online publishing
In addition to the British Horseracing Board case, there have been a number of decisions in other European jurisdictions on the nature and extent of database right.3 These decisions have taken a number of different perspectives on the Database Directive. Many, although not all, of the decisions have arisen in the context of online publishing, and so provide some guidance to the applicability of the right, or at least some illustrations as to how it might be deployed by publishers in this country who seek to protect their content from use by third parties.
The content delivery system as a database
The Database Directive defines a database as ‘a collection of independent works, data or other materials arranged in a systematic or methodical way and individually accessible by electronic or other means’. At the time the Directive was implemented it was rightly observed that this definition could extend well beyond what most IT lawyers (and certainly most systems developers) would regard as a database, for example to the content of a law firm’s library. However, for the purpose of this article, it is assumed that any mainstream commercial system for the publication of online content is likely to come within the definition of a database.
One could test that thinking in relation to a variety of channels. But, taking Web publishing as one example, many Web sites are built on top of a database built from a database development package such as Microsoft SQL Server. By way of simplification, each piece of content which is to be published on the site will be held as a record in a table in the underlying database. Not only the text string itself will be stored, but also its datatype and size as well as other information (such as a unique record identifier) required by the system design. If the displayed text is to act as a hypertext link to another piece of content, then the record must also contain the data needed to access that other content. So, for example, an online newspaper will publish on its homepage various headlines which act as links to fuller stories elsewhere within the site. The headline and story will be held separately in the database. The record from which the headline is published will hold data showing where it links to and the database structure will therefore allow the story to be retrieved via a reference to the headline.
Indeed, much less sophisticated systems than that described above may come within the Directive’s definition of a ‘database’. A very basic Web site comprised of a series of flat HTML pages has a logical structure comprised of one or more folders containing files constituting the pages and the images which appear on those pages.
The investment threshold for database right to subsist
Assuming a content delivery system constitutes a database for the purpose of the Database Directive, database right will subsist if there has been qualitatively or quantitatively a substantial investment in the obtaining presentation or verification of its contents. On the facts of the British Horseracing Board case, this was plainly established (the database cost around £4 million annually to maintain), but smaller Web publishing operations stand to be challenged on this point in actions relying on database right. In the German Baumrakt.de4 case, a Web site publisher challenged a search engine which framed selected results. Although it was accepted that a collection of Web pages could constitute a database, the claimant could not show on the facts that he had made a sufficiently substantial investment in the development of the database. It is this question that will perhaps be most important in assessing whether the extraction or re-utilisation of content from a simple site could infringe database right, rather than the issue of whether the system fits the Directive’s definition of ‘database’.
Extraction or re-utilisation?
The Database Directive defines ‘extraction’ as being ‘the permanent or temporary transfer of all or a substantial part of the contents of a database to another medium by any means or in any form’. As stated above, Laddie J dismissed the argument that extraction required once and for all removal of the content, although clearly some causal connection must be shown between the storage of the data in the original database and its arrival in the new medium.
Re-utilisation is defined as being ‘any form of making available to the public all or a substantial part of the contents of the database by the distribution of copies, by renting by online or other forms of transmission’. Laddie J dismissed the notion that this meant putting information into the public domain for the first time.
The acts of William Hill in the BHB case were found to infringe BHB’s database rights in both ways (ie extraction and re-utilisation), but it is clear that findings of this type need not be confined to situations in which the content is first copied onto the defendant’s systems and then made available from them. Starting from that point, one can consider the application of the right to various common online scenarios that have vexed the minds of pure copyright lawyers (if such creatures still exist) for some time.
Deep linking
The emerging case law on database rights sets the scene for a fresh examination of the celebrated deep-linking problem. The facts of cases such as Shetland Times5 are familiar to most of us but how should such cases now be decided if database right arguments are run?
In Shetland Times (decided before the Database Regulations came into force), the claimant newspaper wanted to prevent the Defendant from deep linking to a small number of news stories on the claimant’s Web site. Although an application for an interim interdict was granted on the basis of some tenuous copyright arguments, there would, on similar facts, almost certainly now be a database right argument along the following lines. The news stories and the uniform resource locators for them were stored on the claimant’s Web server within a system which had the characteristics of a database (even a collection of flat HTML pages ought to satisfy the definition – a site driven from a ‘proper’ database certainly would). By linking to the stories the defendant was carrying out the following acts, each of which could arguably amount to an extraction or re-utilisation.
-
Storing the URLs within its own Web site (underlying the links from the defendant’ site to the claimants); this could amount to extraction or re-utilisation or both of the URLs.
-
Causing the user to access the story when the link from the defendant’s site was activated. On a wide interpretation, this could amount to re-utilisation of the underlying works in the claimant’s database even though these are never copied across the defendant’s servers (a wide approach of this type seems to have been taken by the German Courts in Step Stone v Ofir 28 February 2001).
Framing
Framing raises similar issues to deep linking. Where site A frames content from site B, it is well known that the servers from which site A is published contain only directions as to the relevant sections of site B to be called for by the user’s browser. It is therefore not straightforward to allege copyright infringement since no part of site B need be reproduced or stored on site A’s servers, and so allegations of copyright infringement would depend on showing, for example, that the publisher of site A was wrongfully purporting to authorise users to download materials from site B. However, the width of the re-utilisation right provided for by the Database Directive assists the publisher of site B, since it prevents others from ‘any form of making available to the public’ and this seems (following the deep linking principles referred to above) to include providing a hypertext link to the resource in question. Having said that, the test of substantiality would still need to be considered in each case; in other words one would have to frame a substantial part of the site, or make an insubstantial part of a site available a sufficient number of times.
In the Baumrakt.de case6 a Web site publisher challenged the practice of a search engine which framed selective results. Although it was accepted that a collection of Web pages could constitute a database, the claimant failed to show on the facts sufficiently substantial investment in the development of the database. However, other German decisions do appear to provide support for the framing arguments set out above.7
Search engines
The use of search engine technology can avoid the need for any manual configuration of links to a protected database. A huge variety of search engines exist, including many specialist engines customised to find information relevant to a particular topic. If these specialist engines run well, it is likely that they will continually identify material from the same key sources and therefore the potential will exist for infringement by repeated and systematic re-utilisation of insubstantial parts of the most popular sites, namely by returning the URLs for resources on those sites time and again.
In the Berlin Online case8 for example, the repeated use of a meta-search engine on an online database was held to amount to repeated and systematic extraction of insubstantial parts of that database and therefore to an infringement of the database rights in it.
Other protection for online content
Database right is not, of course, the end of the story for online publishers. Various other rights are likely to be important to them in maximising the value of their content delivery systems.
Copyright, but not database right, will subsist in computer programs that form part of the content delivery system. Copyright will subsist in the database itself so far as it constitutes the author’s own intellectual creation. Copyright may of course subsist in the underlying works stored in the database. By way of brief illustration, a system for publishing an online celebrity news magazine could be built on a content database containing various photographic images and text records. Database right will subsist in the database provided sufficient investment has been made in establishing it. Copyright will subsist in the database if the selection or arrangement of the content comprises the author’s own intellectual creation. Separate software copyrights will subsist in, for example, scripts or utilities written to facilitate the better display of the content in the user’s browser. Separate copyrights will exist in relation to most items of content. Patent protection may be available for some of the system software. The look and feel of the site will be protected to an extent by a mixture of copyright and trade mark protection. Personal information held within the databases will, of course, be subject to data protection legislation.
Conclusions
The emerging case law on database right has a number of significant implications for online publishers. Although the extent of the right is still far from certain, the emerging case law is firming the right up as perhaps the most valuable weapon available to publishers to prevent a number of threats to exclusivity of their content and so to their income streams. It offers greater scope than copyright to counteract deep linking, automatic searching and framing of sites. It also appears to provide stronger protection generally for underlying information, since the requirement of form is less rigorous (subject to the ECJ’s view on the requirement for ‘database-ness’). This would be a surprising result since it would contradict at least one of the recitals to the Directive itself.
Publishers will, of course, have to establish that sufficient investment has been made in their content delivery systems to qualify them for database right and that sufficient data has been extracted or re-utilised for the right to be infringed, but it looks at present as though those hurdles may not be high in practice and that publishers are therefore currently enjoying something of a golden era in terms of content protection. Now may be the time for publishers to settle outstanding linking scores.
Ian Jeffrey is a Partner at Lewis Silkin, specialising in IT law and copyright.
Endnotes
1. British Horse Racing Board Limited and Others v William Hill Organisation Limited, Laddie J 9 February 2001.
2. British Horseracing Board and Others v William Hill Organisation Limited [2001] EWCA Civ 1268.
3. Many of these are summarised in a very useful paper by P. Bernt Hugenholtz which can be found at www.ivir.nl/publications/hugenholtz/fordham2001.html-noot36.
4. Oberlandesgericht (Court of Appeal) Dusseldorf 29 June 1999.
5. Shetland Times v Dr Jonathan Wills [1997] FSR 604.
6. Oberlandesgericht (Court of Appeal) Dusseldorf 29 June 1999.
7. In particular the Medizinisches case referred to by Hugenholtz.
8. Landgericht Berlin 8 October 1998.
