OPINION OF ADVOCATE GENERAL
STIX-HACKL
delivered on 8 June 2004 (1)

Case C-203/02

The British Horseracing Board Ltd and Others

v

William Hill Organization Ltd

(Reference for a preliminary ruling from the Court of Appeal (England & Wales) (Civil Division))

(Directive 96/9/EC – Databases – Protection – Sui generis right – Beneficiaries – Obtaining and verification of the contents of a database – (In)substantial part of the contents of a database – Extraction and re-utilisation – Normal exploitation – Unreasonable prejudice to the legitimate interests of the maker – Significant change to the contents of a database – Sport – Betting)

I –  Preliminary observations

1.       This reference for a preliminary ruling is one of four parallel sets of proceedings (2) concerning the interpretation of Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases (3) (‘the Directive’). Like the other cases, this case concerns the so-called sui generis right and its scope in the area of sporting bets.

II –  Legal background

A – Community law

2.       Article 1 of the Directive contains provisions on the scope of the Directive. It provides inter alia:

‘1.

This Directive concerns the legal protection of databases in any form.

2.

For the purposes of this Directive, “database” shall mean a collection of independent works, data or other materials arranged in a systematic or methodical way and individually accessible by electronic or other means.’

3.       Chapter III regulates the sui generis right in Articles 7 to 11. Article 7, which concerns the object of protection, provides inter alia:

‘1.

Member States shall provide for a right for the maker of a database which shows that there has been qualitatively and/or quantitatively a substantial investment in either the obtaining, verification or presentation of the contents to prevent extraction and/or re-utilisation of the whole or of a substantial part, evaluated qualitatively and/or quantitatively, of the contents of that database.

2.

For the purposes of this Chapter:

(a)

“extraction” shall mean 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;

(b)

“re-utilisation” shall mean any form of making available to the public all or a substantial part of the contents of a database by the distribution of copies, by renting, by on-line or other forms of transmission. The first sale of a copy of a database within the Community by the rightholder or with his consent shall exhaust the right to control resale of that copy within the Community;

Public lending is not an act of extraction or re-utilisation.

3.

The right referred to in paragraph 1 may be transferred, assigned or granted under contractual licence.

…

5.

The repeated and systematic extraction and/or re-utilisation of insubstantial parts of the contents of the database implying acts which conflict with a normal exploitation of that database or which unreasonably prejudice the legitimate interests of the maker of the database shall not be permitted.’

4.       Article 8, which governs the rights and obligations of lawful users, provides in paragraph 1:

‘1.

The maker of a database which is made available to the public in whatever manner may not prevent a lawful user of the database from extracting and/or re-utilising insubstantial parts of its contents, evaluated qualitatively and/or quantitatively, for any purposes whatsoever. Where the lawful user is authorised to extract and/or re-utilise only part of the database, this paragraph shall apply only to that part.’

5.       Article 9 provides that Member States may provide for exceptions to the sui generis right.

6.       Article 10, which concerns the term of protection, provides in paragraph 3:

‘Any substantial change, evaluated qualitatively or quantitatively, to the contents of a database, including any substantial change resulting from the accumulation of successive additions, deletions or alterations, which would result in the database being considered to be a substantial new investment, evaluated qualitatively or quantitatively, shall qualify the database resulting from that investment for its own term of protection’.

B – National law

7.       The Directive was transposed into United Kingdom law by the Copyright and Rights in Databases Regulations 1997 (SI 1997 No 3032). The parties to these proceedings and the national court agree that these Regulations have to be construed consistently with the Directive.

III –  Facts and main proceedings

8.       In the main proceedings, the claimants, the British Horseracing Board (‘BHB’), the governing authority for the British horseracing industry, its Members, the Jockey Club, the Racehorse Association Limited, the Racehorse Owners Association and the Industry Committee (Horseracing) Limited, and Weatherbys are in dispute with the defendant, William Hill. The case concerns the taking of bets on the internet by William Hill and some of its competitors.

9.       The BHB is a company which was formed in June 1993 to take over part of the function formerly carried on by the Jockey Club. After that date the Jockey Club retained the principal regulatory function within British horseracing. Its concern is now the application of the Rules of Racing. The BHB took on the remainder of the administrative functions of racing’s governing body, in particular, the compilation of data related to horseracing.

10.     Weatherbys maintains and publishes the General Stud Book, which is the official document of registration of thoroughbred horses in Great Britain and Northern Ireland. Weatherbys is also a registered bank and has a publishing arm. In 1985, Weatherbys, on behalf of the Jockey Club, started to compile an electronic database of racing information comprising (amongst other things) details of registered horses, their owners and trainers, their handicap ratings, details of jockeys, information concerning fixture lists comprising venues, dates, times, race conditions, entries and runners. The Jockey Club still uses the database for some of its functions.

11.     In 1999, the database of racing information and the Stud Book were integrated into a single electronic database. This is ‘the BHB Database’ with which the proceedings are concerned. It is constantly being updated with the latest information. It is agreed between the parties in the main proceedings that the BHB Database is protected by the sui generis right and that the sui generis right is owned by one or more of the claimants in the main proceedings.

12.     The cost of continuing to maintain the BHB Database and keep it up to date is approximately GBP 4 million per annum and involves approximately 80 employees and extensive computer software and hardware.

13.     The BHB Database contains a huge number of records including many which must be accurately stored and processed each day. It consists of some 214 tables, containing over 20 million records. Each record contains a number of items of data. It includes a collection of data accumulated over many years by way of the registration of information supplied by owners, trainers and others concerned in the racing industry. It contains the names and other details of over one million horses, tracing back through many generations. It contains details of registered owners, racing colours, registered trainers and registered jockeys. It also contains pre-race information, that is to say information relating to races to be run in Great Britain and made available in advance of the race. This covers the place and date on which a race-meeting is to be held, the distance over which the race is to be run, the criteria for eligibility to enter the race, the date by which entries must be made, the entry fee payable and the amount of money the racecourse is to contribute to the prize money for the race.

14.     There are three principal functions performed by Weatherbys, which lead up to the issue of final pre-race information. The first is the registration of information concerning owners, trainers, jockeys, horses, etc. By way of example, annually, Weatherbys registers the names of some ten thousand newly named horses. In addition, the performances of horses competing in each race are recorded. Weatherbys employs a team of some 15 staff whose principal task it is to create and maintain data concerning horses and people.

15.     In addition, it must be ensured that the identities of the horses which take part in races are indeed the same as those whose names are included in the pre-race lists issued.

16.     The second principal function leading up to the issue of final pre-race information is weight adding and handicapping. All entries for handicap and non-handicap races, which total 180 000 entries per annum, must be assigned a weight.

17.     The third activity leading up to the issue of final pre-race information is the compilation of the lists of runners. This is carried out by Weatherbys’ call centre, manned by up to 32 operators at any one time, who receive telephone calls (and faxes) to enter horses in races. Weatherbys check that each horse is qualified for the race in two stages.

18.     As regards the events described in paragraphs 24 to 31 and 32 to 35 of the order for reference, see the annex to this Opinion.

19.     The racing information contained in the BHB Database is of interest to a wide variety of different users. Essential extracts from the database are made available to the racing industry itself, including representatives of the different racecourses around the country, racehorse owners, trainers, riders and their agents, the Jockey Club, pedigree compilers and overseas racing authorities. As stated above the information is made available to these parties each day by way of the joint Weatherbys/BHB internet website and via a database site, and each week in the BHB’s official journal, the Racing Calendar.

20.     In addition, the racing information is of interest to radio and television broadcasters, magazines and newspapers and to members of the public who follow horseracing.

21.     The information is made available on the morning of the day before the race. The names of all the participants in all the races in the United Kingdom are made available to the public on the afternoon before the race through newspapers and Ceefax/Teletext.

22.     The information is also supplied to bookmakers. First, data are made available to a company called Racing Pages Ltd which is controlled and owned by Weatherbys and the Press Association jointly. Racing Pages Ltd forwards data to its various subscribers which include some bookmakers. In particular, Racing Pages makes available to subscribers in electronic form, normally on the day before a race, what is called a Declarations Feed. This contains a list of races, declared runners and jockeys, distance and name of races, race times and number of runners in each race together with other information. Secondly, one of Racing Pages’ subscribers is Satellite Information Services Limited (‘SIS’), which is allowed to use such data for certain purposes. SIS supplies these data to its own subscribers in the form of what is called a raw data feed (‘RDF’). These items of data represent the core of the pre-race information without which punters could not place bets.

23.     William Hill is one of the leading providers of off-course bookmaking services in the United Kingdom and elsewhere, to both United Kingdom and international customers. It and its subsidiaries offer odds on a large number of events at any given time, providing betting services to their customers through two principal channels: (a) a nationwide network of Licensed Betting Offices and (b) telephone betting operations. William Hill’s principal product is the taking of fixed-odds bets on sporting and other events. William Hill also provides betting services over the internet. The most popular event on which William Hill offers odds is horseracing.

24.     William Hill is a subscriber to both the Declarations Feed and the RDF. However, it does not use the Declarations Feed for the activities at issue in these proceedings.

25.     William Hill’s internet service is described in paragraphs 40 to 47 of the order for reference (see annex).

26.     In proceedings before the High Court of Justice the BHB alleged infringement of the sui generis right by William Hill. The Jockey Club and Weatherbys were added as claimants. Mr Justice Laddie made an order ruling that William Hill had infringed the claimants’ database right under both Article 7(1) and Article 7(5) of the Directive. William Hill filed an appeal against the order of Mr Justice Laddie on 14 March 2001. The appeal is pending before the Court of Appeal.

IV –  Questions referred

27.     The Court of Appeal seeks a ruling from the Court of Justice on the following questions:

(1)

May either of the expressions:

(a)

‘substantial part of the contents of the database’; or

(b)

‘insubstantial parts of the contents of the database’

in Article 7 of the Directive include works, data or other materials derived from the database but which do not have the same systematic or methodical arrangement of and individual accessibility to be found in the database?

(2)

What is meant by ‘obtaining’ in Article 7(1) of the Directive? In particular, are the facts and matters in paragraphs 24-31 above [in the annex to this Opinion] capable of amounting to such obtaining?

(3)

Is ‘verification’ in Article 7(1) of the Directive limited to ensuring from time to time that information contained in a database is or remains correct?

(4)

What is meant in Article 7(1) of the Directive, by the expressions:

(a)

‘a substantial part, evaluated qualitatively ... of the contents of that database’? and

(b)

‘a substantial part, evaluated quantitatively ... of the contents of that database’?

(5)

What is meant in Article 7(5) of the Directive, by the expression ‘insubstantial parts of the database’?

(6)

In particular, in each case:

(a)

does ‘substantial’ mean something more than ‘insignificant’ and, if so, what?

(b)

does ‘insubstantial’ part simply mean that it is not ‘substantial’?

(7)

Is ‘extraction’ in Article 7 of the Directive limited to the transfer of the contents of the database directly from the database to another medium, or does it also include the transfer of works, data or other materials, which are derived indirectly from the database, without having direct access to the database?

(8)

Is ‘re-utilisation’ in Article 7 of the Directive limited to the making available to the public of the contents of the database directly from the database, or does it also include the making available to the public of works, data or other materials which are derived indirectly from the database, without having direct access to the database?

(9)

Is ‘re-utilisation’ in Article 7 of the Directive limited to the first making available to the public of the contents of the database?

(10)

In Article 7(5) of the Directive what is meant by ‘acts which conflict with a normal exploitation of that database or unreasonably prejudice the legitimate interests of the maker of the database’? In particular, are the facts and matters in paragraphs 40-47 above [in the annex to this Opinion] in the context of the facts and matters in paragraphs 32-35 above [in the annex to this Opinion] capable of amounting to such acts?

(11)

Does Article 10(3) of the Directive mean that, whenever there is a ‘substantial change’ to the contents of a database, qualifying the resulting database for its own term of protection, the resulting database must be considered to be a new, separate database, including for the purposes of Article 7(5)?

V –  Admissibility

28.     In many respects the questions referred do not so much concern the interpretation of Community law, in other words the Directive, as the application of the directive to a specific set of facts. That being so, I must endorse the Commission’s view that, in proceedings on a reference for a preliminary ruling under Article 234 EC, that is not the role of the Court of Justice but that of the national court and that the Court of Justice must confine itself to interpreting Community law in the case before it.

29.     According to the settled case-law of the Court of Justice, in proceedings under Article 234 EC, which is based on a clear separation of functions between the national courts and the Court of Justice, any assessment of the facts in the case is a matter for the national court. (4)

30.     The Court therefore has no jurisdiction to give a ruling on the facts in the main proceedings or to apply the rules of Community law which it has interpreted to national measures or situations, since those questions are matters for the exclusive jurisdiction of the national court. The analysis of individual events in connection with the database at issue in these proceedings thus requires a factual assessment, which it is for the national court to make. (5) That apart, the Court has jurisdiction to answer the questions referred.

VI –  Assessment of the merits

31.     The questions referred for a preliminary ruling by the national court relate to the interpretation of a series of provisions of the Directive and in the main to the construction of certain terms. The matters addressed fall within different fields and must be dealt with accordingly. While some of the questions concern the scope ratione materiae of the Directive, others relate to the requirements for granting the sui generis right and its content.

A – Scope ratione materiae: the term ‘database’

32.     As regards the requirement of independence of the materials in a database, William Hill has taken the view that the ‘materials’ must be independent of the maker. That view is mistaken. As is clear from the reference William Hill itself makes to the necessity of obtaining the data, this argument raises a point which should, rather, be clarified in the context of the interpretation of the defining element ‘obtaining’ used in Article 7(1) of the Directive.

B – Object of protection: Conditions

33.     In order to be covered by the sui generis right under Article 7 of the Directive a database must fall within the defining elements laid down by that provision. These proceedings concern the interpretation of some of those criteria.

34.     In that connection, reference should be made to the legal debate on the question whether the sui generis right covers the creation, in the sense, essentially, of the activity of creating a database, or the outcome of that process. On that point, it must be observed that the Directive protects databases or their contents but not the information they contain as such. Ultimately it is thus a matter of protecting the product, while at the same time indirectly protecting the expenditure incurred in the process, in other words, the investment. (6)

35.     The requirements laid down by Article 7 of the Directive must be read in conjunction with those laid down by Article 1(2). The resulting definition of the object of protection is narrower than that of ‘database’ in Article 1.

36.     The sui generis right introduced by the Directive derives from the Scandinavian catalogue protection rights and the Dutch ‘geschriftenbescherming’. However, that background must not mislead us into importing the thinking on those earlier provisions developed in academic writings and case-law into the Directive. Rather, the Directive should serve as a yardstick for the interpretation of national law, even in those Member States which had similar provisions before the Directive was adopted. In those Member States, too, the national legislation had to be brought into line with the precepts of the Directive.

1.     ‘Obtaining’ within the meaning of Article 7(1) of the Directive (second question)

37.     One issue in the present case is whether there was any ‘obtaining’ within the meaning of Article 7(1) of the Directive. That provision only protects investment in the ‘obtaining’, ‘verification’ or ‘presentation’ of the contents of a database.

38.     We must base our discussion on the thrust of the protection conferred by the sui generis right, in other words the protection of the creation of a database. Creation can then be seen as an umbrella term for obtaining, verification and presentation. (7)

39.     The main proceedings deal with an often discussed legal problem, that is to say whether, if so, under what conditions, and to what extent the Directive protects not only existing data but also data created by the maker of a database. If obtaining is only to relate to existing data, the protection of the investment would only cover such data. Thus, if we take that interpretation of obtaining as a basis, the protection of the database in the main proceedings depends on whether existing data were obtained.

40.     However, if we take the umbrella term creation, in other words the supplying of the database with content, (8) as a basis, both existing and newly created data could be covered. (9)

41.     A comparison of the term ‘obtaining’ used in Article 7(1) with the activities listed in the 39th recital in the preamble to the Directive might shed some light. However, it must be pointed out at the start that there are divergences between the various language versions.

42.     If we start with the term ‘Beschaffung’, used in the German version of Article 7(1), it can only concern existing data, as it can only apply to something which already exists. In that light Beschaffung is the exact opposite of Erschaffung (creation). Analysis of the wording of the Portuguese, French, Spanish and English versions, which are all based on the Latin ‘obtenere’, to receive, yields the same result. The Finnish and Danish versions also suggest a narrow interpretation. The wide interpretation of the English and German versions advocated by many parties to the proceedings is therefore based on an error.

43.     Further assistance with the correct interpretation of ‘obtaining’ in the terms of Article 7(1) of the Directive might be provided by the 39th recital in the preamble, which is the introductory recital for the subject of the sui generis right. That recital lists only two activities in connection with the protected investments, that is to say ‘obtaining’ and ‘collection’ of the contents. However, here too, problems arise over the differences between the various language versions. In most versions, the same term is used for the first activity as that used in Article 7(1). Moreover, although the terms used do not always describe the same activity, they essentially concern the seeking and collecting of the contents of a database.

44.     The language versions which use, in the 39th recital, two different terms from those used in Article 7(1) of the Directive are to be construed so that the two activities listed are viewed as subspecies of obtaining within the meaning of Article 7(1) of the Directive. Admittedly, that raises the question why the 39th recital only defines obtaining but not verification or presentation more precisely. The latter two terms appear first in the 40th recital.

45.     On the other hand, the language versions which use the same term in the 39th recital as in Article 7(1) of the Directive will have to be construed so that the term obtaining in the 39th recital is understood in a narrower sense, whereas the term used in Article 7(1) of the Directive is to be understood in a wide sense, in other words as also encompassing the other activity listed in the 39th recital.

46.     All the language versions thus allow of an interpretation according to which, although ‘obtaining’ within the meaning of Article 7(1) of the Directive does not cover the mere production of data, that is to say, the generation of data, (10) and thus not the preparatory phase, (11) where the creation of data coincides with its collection and screening, the protection of the Directive kicks in.

47.     In that connection, it should be pointed out that the so-called ‘spin-off theory’ cannot apply. Nor can the objective pursued in obtaining the contents of the database be of any relevance. (12) That means that protection is also possible where the obtaining was initially for the purpose of an activity other than the creation of a database. For the Directive also protects the obtaining of data where the data was not obtained for the purposes of a database. (13) That implies that an external database which is derived from an internal database should also be covered by protection.

48.     It is the task of the national court, using the interpretation of the term ‘obtaining’ set out above, to assess the activities relating to the BHB Database. It is primarily a matter of classifying the data and its handling from its receipt to its inclusion in the database. That entails inter alia the assessment of the three main functions of Weatherbys before the issue of the pre-race information, that is to say the registration of a series of items of information, weight adding and handicapping and compilation of the lists of runners. There is, in addition, the registration of the results of the races.

49.     However, even if those activities were classified as the creation of new data, there might be ‘obtaining’ within the meaning of Article 7(1) of the Directive. That would be the case if the creation of the data took place at the same time as its processing and was inseparable from it. That could be so in the case of the receipt of information and its entry in a database immediately thereafter.

2.     ‘Verification’ within the meaning of Article 7(1) of the Directive (third question)

50.     This question essentially asks whether some of the activities undertaken in connection with the BHB Database should be considered to be ‘verification’ within the meaning of Article 7(1) of the Directive.

51.     Unlike ‘obtaining’, ‘verification’ applies to data which already form the contents of the database. That suggests, prima facie, that the time at which the verification under Article 7(1) takes place is after the registration which is to be verified. The provision thus appears not to cover those verifications which concern materials which have yet to be registered as they do not constitute the existing contents of the database.

52.     It is essentially a matter of monitoring the ‘materials’ of a database in respect of completeness and accuracy, which includes checking whether a database is up to date. However, the outcome of such verification could also require the obtaining of data and their entry in the database.

53.     It is not disputed that those working on the BHB Database undertake a series of checks. These include the various identity checks on applicants and the horses and the checks on the licences to participate.

54.     However, it is disputed whether and, if so, which verifications concern the existing contents of the database such as certain facts about the trainers, or whether the verification of the information takes place before registration, that is to say before the material to be verified becomes part of the database.

55.     However, even if some of the verifications are where necessary undertaken before inclusion in the database, that still does not mean that the other monitoring activities are not to be classified as verification within the meaning of Article 7(1) of the Directive. As regards the updating and correction of the contents of the database which actually took place, we can assume that the requirement of the Directive for a verification has been met. Thus it is sufficient if many of the activities undertaken are to be classified as verification within the meaning of Article 7(1) of the Directive and the substantial investment also concerns at least the part of those activities covered by Article 7(1).

56.     It is the task of the national court to establish whether the monitoring activities at issue in the main proceedings are to be held to be ‘verification’ within the meaning of Article 7(1) of the Directive.

C – Content of the protected right

57.     It must first be observed that, strictly speaking, the introduction of the sui generis right was intended not to harmonise existing law but to create a new right.  (14) That right goes beyond previous distribution and reproduction rights. That should also be taken into account in the interpretation of prohibited activities. Accordingly, the legal definition in Article 7(2) of the Directive assumes particular importance.

58.     At first sight Article 7 of the Directive contains two groups of prohibitions or, from the point of view of the person entitled, that is to say the maker of a database, two different categories of right. Whereas paragraph 1 lays down a right to prevent use of a substantial part of a database, paragraph 5 prohibits certain acts relating to insubstantial parts of a database. On the basis of the relationship between substantial and insubstantial, paragraph 5 can also be understood as an exception to the exception implied by paragraph 1. (15) Paragraph 5 is intended to prevent circumvention of the prohibition laid down by paragraph 1, (16) and can thus also be classified as a protection clause. (17)

59.     Article 7(1) provides for a right of the maker to prevent certain acts. That entails a prohibition on such preventable acts. The preventable and thus prohibited acts are, first, extraction and, second, re-utilisation. Legal definitions of the terms ‘extraction’ and ‘re-utilisation’ are given in Article 7(2) of the Directive.

60.     However, the prohibition laid down by Article 7(1) is not absolute, but requires the whole or a substantial part of a database to have been affected by a prohibited act.

61.     The two defining elements must therefore be examined on the basis of the criterion determining application of Article 7(1) and (5): ‘substantial’ or ‘insubstantial’ part as the case may be. Thereafter the prohibited acts under Article 7(1) and (5) are to be considered.

1.     Substantial or insubstantial parts of a database

a)     General observations (first question)

62.     It was contended in the proceedings that Article 7(1) of the Directive only prohibits acts which entail that the data are arranged in as systematic or methodical a way and are as individually accessible as in the original database.

63.     That argument must be understood as laying down a condition for the application of the sui generis right. Whether there is in fact any such condition must be determined on the basis of the provisions on the object of protection and in particular on the basis of the legal definition laid down in Article 7(2) of the acts prohibited under Article 7(1).

64.     Neither Article 7(1) nor Article 7(5) of the Directive lays down the above condition expressly or makes any reference to it. Rather, the fact that express reference is made in Article 1(2) to arrangement ‘in a systematic or methodical way’ whereas no such reference is made in Article 7 suggests the opposite conclusion, that is to say, that the Community legislature did not intend to make that criterion a condition for the application of Article 7.

65.     Moreover, the very purpose of the Directive precludes such an additional criterion.

66.     The protection provided for in Article 7 would be undermined by such an additional criterion because the prohibition laid down by that article could be circumvented by simple alteration of parts of the database.

67.     The 38th recital in the preamble to the Directive demonstrates that the Directive was also intended to prohibit possible breaches consisting in the rearrangement of the contents of a database. That recital refers to that risk and to the inadequacy of copyright protection.

68.     The purpose of the Directive is precisely the creation of a new right, and even the 46th recital cannot refute that as it concerns another aspect.

69.     Even the 45th recital, according to which copyright protection is not to be extended to mere facts or data, does not support the argument for an additional criterion. That, of course, does not mean that the protection covers the data themselves or individual data. The object of protection is and remains the database.

70.     Accordingly it must be considered that the fact of having the same systematic or methodical arrangement as the original database does not constitute a criterion for the determination of the legality of the actions taken in connection with the database. Therefore, the view that the Directive does not protect data which are compiled in an altered or differently structured way is fundamentally mistaken.

71.     The answer to the first question referred should therefore be that the expressions ‘a substantial part … of the contents of that database’ or ‘insubstantial parts of the contents of the database’ in Article 7 of the Directive can also cover works, data or other materials derived from the database but which do not have the same systematic or methodical arrangement of and individual accessibility to be found in the original database.

b)     The expression ‘substantial part of the contents of a database’ in Article 7(1) of the Directive (first, fourth and sixth questions)

72.     This question seeks an interpretation of the term ‘substantial part of the contents of a database’ in Article 7(1) of the Directive. In contrast with other key terms in the Directive there is no legal definition of this term. It was removed in the course of the legislative procedure, at the stage of the Common Position of the Council, to be precise.

73.     Article 7(1) of the Directive provides for two alternatives. As is clear from the wording a part may be substantial in quantitative or qualitative terms. The wording chosen by the Community legislature must be interpreted as meaning that a part may be substantial even when it is not substantial in terms of quantity but is in terms of quality. Thus the argument that there must always be a minimum in terms of quantity must be dismissed.

74.     The quantitative alternative must be understood as requiring the amount of the part of the database affected by the prohibited act to be determined. That raises the question whether this must be assessed in relative or absolute terms. In other words whether a comparison must be made of the amount in question with the whole of the contents of the database (18) or whether the affected part is to be assessed in itself.

75.     In that connection, it must be observed that a relative assessment would tend to disadvantage the makers of large databases (19) because the larger the total amount the less substantial the affected part. However, in such a case, a qualitative assessment undertaken at the same time could balance out the equation where a relatively small affected part could none the less be considered substantial in terms of quality. Equally, it would be possible to combine both quantitative approaches. On that basis even a part which was small in relative terms could be considered substantial because of its absolute size.

76.     The question also arises whether the quantitative assessment can be combined with the qualitative. Of course, it only arises in cases where an assessment in terms of quality is possible in the first place. If it is, there is nothing to prevent the affected parts from being assessed according to both methods.

77.     In a qualitative assessment, technical or economic value is relevant in any event. (20) Thus, a part which is not large in volume but is substantial in terms of value may also be covered. Examples of valuable characteristics of lists in the field of sport would be completeness and accuracy.

78.     The economic value of an affected part is generally measured in terms of the drop in demand (21) caused by the fact that the affected part is not extracted or re-utilised under market conditions but in some other way. The affected part and its economic value can also be assessed from the point of view of the wrongdoer, that is to say in terms of what the person extracting it or re-utilising it has saved.

79.     In the light of the objective of protecting investment pursued by Article 7 of the Directive, the investment made by the maker will always have to be taken into consideration in the assessment of whether a substantial part is involved. (22) According to the 42nd recital, the prohibition on extraction and re-utilisation is intended to prevent detriment to investments. (23)

80.     Thus, investments, and in particular the cost of obtaining data, can also be a factor in the assessment of the value of the affected part of a database. (24)

81.     There is no legal definition in the Directive of the point at which a part becomes substantial. The unanimous view expressed in legal writings is that the Community legislature intentionally left such demarcation to the courts. (25)

82.     However, the question whether a substantial part is affected may not be allowed to depend on whether there is significant detriment. (26) Mere reference to such detriment in a recital, that is to say at the end of the 42nd recital, cannot be sufficient to cause the threshold for protection to be set so high. It is, moreover, debatable whether ‘significant detriment’ can be relied on as a criterion for defining substantialness at all since the 42nd recital could also be construed as meaning that ‘significant detriment’ is to be seen as an additional requirement in cases in which a substantial part is affected, that is to say in cases where substantialness has already been established. Even the ‘serious economic and technical consequences’ of prohibited acts referred to in the eighth recital cannot justify too strict an assessment in relation to detriment. Both recitals serve, rather, to emphasise the economic necessity for protection of databases.

c)     The expression ‘insubstantial parts of the contents of the database’ in Article 7(5) of the Directive (fifth and sixth questions referred)

83.     The expression ‘insubstantial parts of the contents of the database’ in Article 7(5) of the Directive is likewise no longer legally defined as it was in Article 11(8)(a) of the amended Commission proposal (93) 464 fin.

84.     Interpretation of the criterion ‘insubstantial’ must begin with the objective of the provision for which it is legally determining. Article 7(5) of the Directive is intended to cover those areas which are not covered by Article 7(1), which only concerns substantial parts. Accordingly, ‘insubstantial part’ must be interpreted as meaning a part which does not reach the threshold for a substantial part in terms of quality or quantity within the meaning of Article 7(1). That threshold forms the upper limit. There is, however, a lower limit which derives from the general principle of the Directive that the sui generis right does not cover individual data.

85.     It is for the national court to assess the parts at issue in the main proceedings by applying the above criteria to the specific facts.

2.     Prohibitions relating to the substantial part of the contents of a database

86.     The right of the maker enshrined in Article 7(1) of the Directive to prevent certain acts implies a prohibition on such acts, namely extraction and/or re-utilisation. Such acts are therefore described as ‘unauthorised’ in a series of recitals. (27)

87.     I now turn to the interpretation of the terms ‘extraction’ and ‘re-utilisation’. In that connection the corresponding legal definitions in Article 7(2) of the Directive must be analysed. Here too, the objective of the Directive of introducing a new form of right must be borne in mind. Reference will have to be made to that yardstick for guidance in the analysis of the two terms.

88.     The principle applies, with regard to both prohibited acts, that the objective or intention of the user of the contents of the database is not relevant. Thus, it is not of decisive importance whether the use is purely commercial. Only the defining elements of the two legal definitions are of relevance.

89.     Again, with regard to both prohibited acts, and in contrast to the position under Article 7(5), it is not only repeated and systematic acts which are covered. As the acts prohibited under Article 7(1) have to concern substantial parts of the contents of a database, the Community legislature has less stringent requirements of such acts than those applicable in respect of insubstantial parts under Article 7(5).

90.     In that connection, an error in the structure of the Directive must be pointed out. (28) As the legal definition of Article 7(2) also focuses on the whole or a substantial part, it duplicates the requirement laid down by Article 7(1) unnecessarily. In combination with Article 7(5), the legal definition laid down in Article 7(2) even entails a contradiction since Article 7(5) prohibits the extraction and re-utilisation of insubstantial parts. Analysis of extraction and re-utilisation according to the legal definition in Article 7(2) yields the odd result that Article 7(5) prohibits certain acts in relation to insubstantial parts only when such acts concern the whole or substantial parts.

91.     Several parties also raised the question of competition. This aspect should be considered in the light of the fact that the final version of the Directive does not contain the rules on the distribution of compulsory licences originally planned by the Commission.

92.     Opponents of extensive protection for the maker of a database fear that extensive protection gives rise to a danger of the creation of monopolies, particularly in the case of hitherto freely accessible data. For instance, a maker who has a dominant position on the market could abuse that position. In that connection it must be borne in mind that the Directive does not preclude the application of the competition rules in primary law and in secondary legislation. Anti-competitive conduct by makers of databases is still subject to those rules. That is clear both from the 47th recital and from Article 16(3) of the Directive, under which the Commission is to verify whether the application of the sui generis right has led to abuse of a dominant position or other interference with free competition.

93.     In these proceedings the issue of the legal treatment of freely accessible data was also addressed. In that connection, it was those governments submitting observations in the proceedings which expressed the view that public data were not protected by the Directive.

94.     On that point, it must first be emphasised that the protection covers the contents of databases and not of data. First, the risk that the protection might extend to the information contained in the database can be countered by interpreting the Directive narrowly in that respect, as proposed here. Second, recourse to the national and Community instruments of competition law where necessary is mandatory.

95.     As regards the protection of data which make up the content of a database of which the user of the data is unaware, it must be pointed out that the Directive prohibits only certain acts, that is to say, extraction and re-utilisation.

96.     Although the prohibition of extraction laid down in the Directive presupposes knowledge of the database, that is not necessarily the case as regards re-utilisation. I will come back to that issue in connection with re-utilisation.

a)     The term ‘extraction’ in Article 7 of the Directive (seventh question referred)

97.     The term ‘extraction’ in Article 7(1) of the Directive is to be interpreted on the basis of the legal definition in Article 7(2)(a).

98.     The first element is the transfer of the contents of a database to another medium, such transfer being either permanent or temporary. The wording ‘by any means or in any form’ implies that the Community legislature gave the term ‘extraction’ a wide meaning.

99.     It thus covers not only the transfer to a data medium of the same type (29) but also to one of another type. (30) That means that merely printing out data falls within the definition of ‘extraction’.

100.   Furthermore, ‘extraction’ clearly cannot be construed as meaning that the extracted parts must then no longer be in the database if the prohibition is to take effect. Nor, however, must ‘extraction’ be so widely construed as also to cover indirect transfer. Rather, direct transfer to another data medium is required. In contrast to ‘re-utilisation’ it does not require any public element. Private transfer is also sufficient.

101.   As regards the second element, that is to say the affected part of the database (‘whole or substantial part’), reference can be made to the arguments on substantialness.

102.   It is the task of the national court to apply the above criteria to the specific facts of the main proceedings.

b)     The term ‘re-utilisation’ in Article 7 of the Directive (eighth and ninth questions referred)

103.   According to the legal definition in Article 7(2)(b) of the Directive, re-utilisation involves making data publicly available.

104.   By deliberately using the term ‘re-utilisation’ rather than ‘re-exploitation’ the Community legislature wanted to make clear that the protection was to cover acts by non-commercial users too.

105.   The means of ‘re-utilisation’ listed in the legal definition such as ‘the distribution of copies, by renting, by on-line … transmission’ are to be understood simply as a list of examples, as is clear from the additional words ‘or other forms of transmission’.

106.   In cases of doubt, the term ‘making available’ is to be construed widely (31) as the use of the additional words ‘any form’ in Article 7(2)(b) suggests. On the other hand, mere ideas (32) or a search for information as such using a database (33) are not covered.

107.   Many of the parties expressed the view that the data were in the public domain. Whether that is so can be determined by examination of the specific facts, which is a matter for the national court.

108.   However, even if the national court reaches the conclusion that the data are in the public domain that does not preclude parts of the database containing data in the public domain from also enjoying protection.

109.   In Article 7(2)(b) of the Directive there are also rules on the exhaustion of the right. The right is exhausted only under certain conditions. One of those conditions is described as ‘the first sale of a copy of a database’. That suggests that there can be exhaustion of the right only in respect of such physical objects. If re-utilisation happens in some other way than though a copy, there is no exhaustion. As regards on-line transmission that principle is expressly laid down in the 43rd recital. The sui generis right thus does not only apply on the first ‘making available to the public’.

110.   As the Directive does not mention the number of transactions following the first ‘making available to the public’ that number cannot be relevant. Thus, if a substantial part of the contents of a database is involved that is protected even if it was obtained from an independent source such as a print medium or the internet and not from the database itself. Unlike extraction, ‘re-utilisation’ also covers indirect means of obtaining the contents of a database. The defining element ‘transfer’ must therefore be interpreted widely. (34)

111.   It is for the national court to apply the above criteria to the specific facts of the main proceedings.

3.     Prohibitions concerning insubstantial parts of the contents of a database (10th question referred)

112.   As already pointed out, Article 7(5) of the Directive lays down a prohibition on the extraction and/or re-utilisation of insubstantial parts of the contents of a database. This provision differs from Article 7(1) firstly in that not every extraction or re-utilisation is prohibited but only defined instances. ‘Repeated and systematic’ acts are required. Secondly, the prohibition in Article 7(5) differs from that in Article 7(1) as regards its subject‑matter. This prohibition applies even to insubstantial parts. Thirdly, to offset this lesser requirement of the affected part in comparison with Article 7(1), Article 7(5) requires unauthorised acts to have a specific effect. In that regard, Article 7(5) provides for two alternatives: the unauthorised acts must either conflict with a normal exploitation of that database or unreasonably prejudice the legitimate interests of the maker of the database.

113.   The provision on the connection between the act and its effect must be understood to mean that it is not necessary for every individual act to have one of the two effects but that the overall result of the acts must have one of the two prohibited effects. (35) The objective of Article 7(5) of the Directive and of Article 7(1) is the protection of the return on investment.

114.   However, the interpretation of Article 7 generally raises a problem in that the German language final version of the Directive was formulated rather more weakly than the Common Position. It is now sufficient for the act to ‘imply’ (‘hinausläuft’) rather than ‘have the result of’ (‘gleichkommt’) one of the effects described. The other language versions are formulated more directly and essentially concern extraction and/or re-utilisation which conflicts with a normal exploitation of that database or which unreasonably prejudices the legitimate interests of the maker of the database.

115.   In this connection related international law should be discussed. Both the effects mentioned in Article 7(5) of the Directive are modelled on Article 9(2) of the Berne Convention as revised and in fact on the first two stages of the three‑step test laid down therein. However, that does not mean that both provisions must be interpreted in the same way.

116.   First, Article 9 of the Berne Convention as revised serves a different purpose. That provision gives the parties to the Convention the authority to derogate from the strict rule of protection under the conditions in the three‑step test. Provision is made for that sort of construction (that is to say, the option of exceptions for Member States) in Article 9 of the Directive.

117.   Secondly, Article 9 of the Berne Convention as revised differs in that it does not formulate ‘conflict with a normal exploitation’ and ‘unreasonable prejudice’ as alternatives but as two of three cumulative defining characteristics. (36)

118.   Other international rules similar to Article 7(5) of the Directive are to be found in Article 13 of the TRIPs Agreement and certain WIPO agreements. However, as the latter were adopted after the Directive they should be left out of account.

119.   As regards the interpretation of Article 13 of the TRIPs Agreement, similar reservations can be raised as in connection with the Berne Convention as revised. For Article 13, like Article 9 of the Berne Convention as revised, regulates the limits on and exceptions to the exclusive rights imposed by the Member States. However, unlike Article 9 of the Berne Convention as revised, both effects, that is to say ‘conflict with a normal exploitation’ and ‘unreasonable prejudice’, are given as alternatives as in the Directive.

120.   These considerations demonstrate that the interpretation of the above rules of international law cannot be transferred to Article 7(5) of the Directive.

121.   The acts of extraction and re-utilisation prohibited under the Directive and the effects of such acts regulated by it have in common that the purpose of the acts is not of decisive importance. Article 7(5) of the Directive cannot be interpreted in that way in the absence of any rule concerning purpose. If the Community legislature had wanted purpose to be taken into account it could have used in Article 7 of the Directive a formulation like that in Article 9(b) of the Directive.

a)     ‘Repeated and systematic extraction and/or re-utilisation’

122.   The requirement for ‘repeated and systematic’ acts is intended to prevent the undermining of protection by successive acts each concerning only an insubstantial part. (37)

123.   On the other hand, it is unclear whether Article 7(5) lays down two alternative or two cumulative requirements. Any interpretation should begin with the wording of the provision. However, that does not yield any unequivocal result. Some language versions link the two requirements with ‘and’, (38) others with ‘or’.  (39) Most of the language versions and the objective of the Directive, however, indicate that the two characteristics are to be understood as cumulative requirements. (40) A repeated but not systematic extraction of an insubstantial part of the contents of a database is therefore not covered.

124.   There is a repeated and systematic act when it is carried out at regular intervals, for example, weekly or monthly. If the interval is less and the affected part small, the act will have to be carried out more frequently for the part affected overall to fulfil one of the two requirements laid down by Article 7(5) of the Directive.

b)     The expression ‘normal exploitation’ in Article 7(5) of the Directive

125.   The term ‘normal exploitation’ in Article 7(5) of the Directive must be interpreted in the light of the objective of that protective clause. That is clear in particular from the preamble to the Directive. In the 42nd recital the prevention of detriment to investment is cited as a reason for the prohibition of certain acts. In the 48th recital the objective of the protection enshrined in the Directive is expressly described ‘as a means to secure the remuneration of the maker of the database’.

126.   That indicates a wide interpretation of the term ‘normal exploitation’. Thus ‘conflict with … exploitation’ must be understood not only in the technical sense that only effects on the technical usability of the affected database are covered. Rather, Article 7(5) also relates to purely economic effects on the maker of a database. It is a matter of protecting the economic use made of it under normal circumstances. (41)

127.   Thus, Article 7(5) is applicable not only in relation to acts which result in the creation of a competing product which then conflicts with the exploitation of the database by its maker. (42)

128.   In individual cases Article 7(5) may cover the exploitation of potential markets not exploited by the maker of the database. Accordingly, it is, for example, sufficient, if the person extracting or re-utilising the data fails to pay licence fees to the maker of the database. If such acts were allowed, that would provide an incentive for other persons to extract or re-utilise the contents of the database without paying licence fees. (43) If there were thus the possibility of exploiting the database without charge, that would have serious implications for the value of licences, resulting in reduced income.

129.   The rule is not limited to cases where the maker of the database wishes to use its contents in the same way as the person extracting or re-utilising the data. Nor is it relevant either that the maker of a database cannot exploit its contents in the same way as the person extracting or re-utilising it because of a statutory prohibition.

130.   Finally, ‘conflict with … exploitation’ must not be interpreted so narrowly that only a total ban on exploitation would be prohibited. According to all the language versions other than the German the prohibition is applicable as soon as there is any conflict with exploitation, that is to say, even in the case of negative effects on a limited scale. That is also where the threshold lies above which detriment to the maker of the database can be assumed, thus triggering the prohibition.

131.   As many of the parties have pointed out, it is for the national court to assess the specific acts and their effect on the database at issue in the proceedings on the basis of the above criteria.

c)     The expression ‘unreasonably prejudice’ in Article 7(5) of the Directive

132.   As regards the interpretation of the term ‘unreasonably prejudice’ in Article 7(5) of the Directive it must first be recalled that there had already been discussion, in connection with the Berne Convention as revised, as to whether such an unspecific legal term was usable at all. Reference to the ways in which it differs from ‘normal exploitation’ is also crucial to the interpretation of the expression ‘unreasonably prejudice’.

133.   As regards the scope of protection the provision at issue makes lesser demands of the expression ‘unreasonably prejudice’ than of ‘normal exploitation’ in so far as in the former case ‘legitimate interests’ are protected. The protection therefore goes beyond legal position and also covers interests, that is to say ‘legitimate’ and not only legal interests.

134.   To offset this, Article 7(5) lays down stricter requirements as regards the effect of the unauthorised acts. They must ‘unreasonably prejudice’ and not merely prejudice. However, the term ‘unreasonably’ must not be interpreted too strictly. Otherwise the Community legislature would have required damage or even significant damage to the maker here too.

135.   In the light of the language versions other than the German, it will have to be interpreted as requiring the acts to have damaged interests to a certain extent. In that connection the Directive focuses, here as elsewhere, on detriment to the maker. The main proceedings show very clearly that the protection of the maker’s rights affects the economic interests of others. However, that does not mean that great importance can be attached, in the interpretation of Article 7(5) of the Directive, to the effects of the sui generis right on the interests of other persons or to any possible ‘damage’ to the relevant Member State as a result of possible effects on income from taxation. The Directive is intended to prevent detriment to makers of databases. Unlike the other effects, this objective is expressly enshrined in the Directive.

136.   The maker’s investments and the return on them constitute the core interest referred to in Article 7(5). Thus, here too, the economic value of the contents of the database is the starting point for assessment. The focus of the assessment is the effects on the actual or anticipated income of the maker of the database. (44)

137.   We can use the expression ‘normal exploitation’ as a basis for assessing the extent of protection. If we interpret that term narrowly, as not covering the protection of potential markets, such as new ways of exploiting the contents of a database, (45) we will at least have to describe the impact on potential markets as prejudice to legitimate interests. Whether such prejudice is unreasonable, will depend on the facts of the individual case. However, the fact that the person extracting or re-utilising the data is a competitor of the maker of the database cannot be decisive.

138.   In this connection, too, it must be recalled that it is for the national court to investigate the specific facts and to ascertain whether they must be considered to ‘unreasonably prejudice’ the legitimate interests of the maker of the database at issue in the proceedings.

D – Change to the contents of a database and term of protection (11th question referred)

139.   In the present case the issue is whether every ‘substantial change’ to the contents of a database qualifying the resulting database for its own term of protection entails that the resulting database must be considered to be a new, separate database for the purposes of Article 7(5).

140.   Under Article 10(3) of the Directive any substantial change – under certain conditions – qualifies the database for its own term of protection. One of those conditions, namely the criterion ‘substantial change to the contents of a database’, and the consequences it entails will be examined below. In the present case the issue is to be investigated from the perspective of ‘repeated and systematic extraction and/or re-utilisation’ within the meaning of Article 7(5) of the Directive.

141.   Essentially this question concerns the object of the extended term of protection. In that connection, it must be clarified whether substantial changes result in the creation of a new database. If it is concluded that a new database exists alongside the existing database what is decisive is which database the prohibited acts relate to.

142.   In response to various submissions, the question whether Article 10(3) of the Directive is to be interpreted as meaning that it governs only the term of protection and not the object of protection is to be considered at the same time.

143.   The wording of Article 10(3), according to which a substantial change, under certain circumstances, ‘shall qualify the database resulting from that investment for its own term of protection’, suggests that the Community legislature assumed that such a change resulted in a separate database. That conclusion is confirmed by the other language versions.

144.   Nor can the systematic interpretation be adduced in refutation. For instance, although the title of Article 10 is ‘term of protection’ that does not mean that that article merely regulates the period of time and not also the object of that protection.

145.   The view taken by the Community in the context of the WIPO also supports the argument for a new database in the event of a substantial change under certain circumstances. (46)

146.   It is obvious that the new term of protection laid down by Article 10(3) can only relate to a specific object. It is clear from the background to the drafting of this provision that the result of the further investment was meant to enjoy protection. (47) Limiting the object of protection to the resulting database is consistent with the objective of providing for a new term of protection. (48)

147.   It should be recalled, here, that the database at issue in these proceedings is what is known as a dynamic database, that is to say, a database which is constantly updated. It must be borne in mind that not only deletions and additions but also, as is clear from the 55th recital, verifications are to be considered changes within the meaning of Article 10(3) of the Directive.

148.   It is characteristic of dynamic databases that there is only ever one database, namely the most recent. Previous versions ‘disappear’. That raises the question of what the new term of protection covers, in other words, what the object of protection, that is to say, the new one, is.

149.   The point of departure must be the objective of the changes, which is to bring the database up to date. That means that the whole database is the object of the new investment. Thus, the most recent version, that is to say, the whole database, is always the object of protection. (49)

150.   The background to the drafting of the Directive also supports that interpretation. Although Article 9 of the original proposal (50) still made provision for extension of the term of protection of a database, in its explanatory statement to the proposal the Commission expressly referred to a new ‘edition’ of the database. (51) A clarification as regards constantly updated databases was then included in an amended proposal. (52) In the legal definition in Article 12(2)(b) the successive accumulation of small changes typical of dynamic databases is expressly mentioned.

151.   Viewed in that light Article 10(3) of the Directive provides for a ‘rolling’ sui generis right.

152.   Ultimately, the solution proposed here for dynamic databases reflects the principle that it is always the result, that is to say, the new and not the old database, which is protected. Dynamic databases differ from static databases simply in that, in the case of dynamic databases, the old database ceases to exist because it is constantly transformed into a new one.

153.   Further, the fact that in the case of dynamic databases the whole database and not only the changes as such enjoy a new term of protection can, regardless of the objective and subject‑matter of the new investment, be justified by the fact that only an assessment of the whole of the database as such is practicable.

154.   The objective of protecting investments and of providing an incentive for investment lends further support to the argument for assessment as a whole. In the case of dynamic databases these objectives can only be attained if updates are also covered. (53) Otherwise investment in dynamic databases would be disadvantaged.

155.   It is for the national court to assess the specific changes to the database in the main proceedings. In the course of that assessment the national court must take account of the fact that even insubstantial changes in sufficient number are to be classified as substantial changes. As is clear from the 54th recital in the preamble to the Directive, the burden of proof that the criteria in Article 10(3) exist lies with the maker of the new database.

156.   It is for the national court to judge at what point the threshold above which a part becomes substantial has been crossed. In that connection it must be ascertained whether the new investment is substantial. The assessment of the substantial nature of a part must be based on the requirements in Article 7 of the Directive. The relevant conditions with regard to investment must also be observed. That is so, regardless of the fact that Article 10(3) refers expressly to ‘new investment’, whereas Article 7 concerns initial investment. (54)

VII –  Conclusion

157.   I therefore propose that the Court should answer the questions referred as follows:

1.

Whether the works, data or other materials derived from the database have the same systematic or methodical arrangement and individual accessibility as in the original database is not relevant to the interpretation of the expressions ‘a substantial part … of the contents of that database’ or ‘insubstantial parts of the contents of the database’ in Article 7 of Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases.

2.

The expression ‘obtaining’ in Article 7(1) of the Directive must be interpreted as meaning that it also covers data created by the maker if the creation of the data took place at the same time as its processing and was inseparable from it.

3.

The term ‘verification’ in Article 7(1) of the Directive is to be interpreted as meaning that it is not limited to ensuring from time to time that information contained in a database is or remains correct.

4.

The expression ‘a substantial part, evaluated qualitatively ... of the contents of that database’ in Article 7(1) of the Directive is to be interpreted as meaning that the technical or commercial value of the affected part must be taken into account. The expression ‘a substantial part, evaluated quantitatively ... of the contents of that database’ in Article 7(1) of the Directive is to be interpreted as meaning that the amount of the affected part is relevant. However, in both cases it is not solely the relative amount of the affected part as a proportion of the contents as a whole that is relevant.

5.

The expression ‘insubstantial parts of the contents of the database’ in Article 7(5) of the Directive is to be interpreted as meaning that such parts are more than individual data and less than ‘substantial parts’ within the meaning of Article 7(1).

6.

The term ‘extraction’ in Article 7 of the Directive covers only the transfer of the contents of the database directly from the database to another data medium.

7.

The term ‘re-utilisation’ in Article 7 of the Directive covers not only the making available to the public of the contents of the database directly from the database, but also the making available to the public of works, data or other materials which are derived indirectly from the database, without having direct access to the database.

8.

Acts which prevent the economic exploitation of the sui generis right by its proprietor even on potential markets are to be considered to be ‘acts which conflict with a normal exploitation of that database’. Acts which damage the legitimate economic interests of the maker to a degree which exceeds a certain threshold are to be considered to be ‘acts which unreasonably prejudice the legitimate interests of the maker of the database’.

9.

Article 10(3) of the Directive is to be interpreted as meaning that any ‘substantial change’ to the contents of a database which qualifies the database for its own term of protection, entails that the resulting database must be considered to be a new, separate database, including for the purposes of Article 7(5).

ANNEX

(Order for reference)

24. The third activity leading up to the issue of final pre-race information is the compilation of the lists of runners. This is carried out by Weatherbys’ call centre, manned by up to 32 operators at any one time, who receive telephone calls (and faxes) to enter horses in races. For most races, a horse must be entered by noon, five days before the race. The caller identifies himself by reference to a personal identification number assigned to him. The caller is then asked to quote the code number, as published in the Racing Calendar, of the race for which he wishes to make an entry, the name of the horse and the name of its owner.

25. Weatherbys check that each horse is qualified for the race in two stages. The first stage is in ‘real-time’ when the horse is first entered for the race. The age and sex of the horse is compared with the conditions of the race in question and, if the horse is not qualified on either ground, a warning appears on-screen, and the entry is not accepted. If the person entering the horse for the race has not been formally authorised to act by the lodgement of a written authority to act with Weatherbys, or if the owner is not registered, or if the trainer is not licensed, or has not advised Weatherbys that the horse is in his care, or if the horse’s name is not registered, this fact is made apparent by the computer system to the operator, who then will not accept the entry. Each entry is allotted its own unique reference number for ease of identification in the ‘declaration’ process that then ensues and which is described below.

26. The fact that a horse is ‘entered’ in a race does not necessarily mean it will run in the race. First, it must be ascertained whether the horse is eligible to run in the race. Secondly, the trainer must confirm that he wants the horse to run in the race (this is called ‘declaring’ the horse and is done the day before the race). Thirdly, even a declared horse may not be permitted to run if, for example, there are too many declarations for the race. Because a horse must be ‘declared’ before it runs in the race, trainers can and do ‘enter’ horses for more than one race on the same day, knowing that they may later ‘declare’ the horse for only one or none of these races.

27. Once the original deadline for entries has passed, the entries received are processed by the computer, listing them by race, within race-meeting. Once this is done, the list of ‘provisional entries’ (i.e. entries which have not undergone the double-check described below and to which no weights have yet been allotted) is made available through the BHB/Weatherbys Joint Internet Site and the BHB’s Information Service on Videotex.

28. All the telephone conversations are tape-recorded. During the afternoon, they are replayed and checked against an audit report produced by the computer. The operator ‘playing-back’ will never be the same operator that took the call. In this way, a double-check is carried out to ensure, as far as possible, that the caller’s wishes were correctly heard and actioned and that the issued list of entries will be accurate.

29. Weatherbys then undertakes the second stage of the checks as to the eligibility of the horses to run in the race in question. Because these make reference to the detailed records of each horse’s past performances and compare them with the relevant conditions of the race in question, they are not carried out in real-time, for fear of slowing down the throughput of entries. Rather, the computer carries out this task after the deadline has passed by reference to the past performance and handicapping data. At the same time it calculates and assigns its weight as described above. Thereafter, a ‘confirmed’ list of entries, now double-checked and with weights assigned, is issued through the same channels.

30. A further transaction needs to take place before a horse takes its place in the final list of runners issued. A horse which is ‘entered’ must be ‘declared’ by its trainer if it is to run in the race. The deadline for declarations is normally the day before the race – currently at 10.00 a.m. in the summer months, 10.15 a.m. in the winter. This process entails the trainer telephoning the Weatherbys call centre before the allotted deadline and ‘declaring’ (i.e. confirming the intention that the horse should run in the race). No declaration calls are entertained after the deadline. When the declaration is made by telephone, the call centre will identify the entry by reference to the reference number allotted to that entry when it was first made, which the caller will quote.

31. After the overnight declarations deadline, the computer assigns a saddle cloth number to each horse. This is done by reference to the final weight allotted (inclusive of any penalties which might have been added as recently as the morning of declaration). For horses allotted the same weight, the order is either random (in the case of handicaps) or in alphabetical order of horse name (for non-handicaps). In addition, for flat races, the computer carries out a random generation of starting stall numbers, to determine from which stall each horse will start. The starting position is an element of information of which punters are known to take account – the significance of the draw varying according to the racecourse, distance of race, etc. A further check is carried out of any very recent performances of declared horses. Where, under the conditions of the race, these performances result in the allocation of a penalty, this penalty is added to the basic weight allotted. On occasions, the weights of the declared runners may need to be adjusted by the computer, depending on the race conditions. Moreover, if the number of declared runners exceeds the maximum number of participants, regulated by the Jockey Club for reasons of safety (being information which is also stored on the database), the race may need to be divided, according to set procedures, or some horses will be eliminated (i.e. not put on the final list of runners) again according to set procedures.

32. The maintenance of the BHB Database (including the steps mentioned above that lead to the lists of actual runners being generated) is only part of the BHB’s function. It currently costs the BHB GBP 15 million per year to carry out all of its functions on behalf of the British horse racing industry. The cost to the BHB of running the BHB Database therefore accounts for about 25% of BHB’s total expenditure. The BHB is self-financing, obtaining its income primarily from fees for registrations and licences, fixture fees from racecourses and entry handling charges payable by owners and racecourses. Part of its income is derived from fees charged to third parties for use of information contained on the BHB Database. These fees currently yield an income of just over GBP 1 million annually, thereby meeting somewhere in the region of 25% of the BHB’s costs of maintaining the BHB Database.

33. For example, Weatherbys supplies information from the BHB Database to William Hill and other bookmakers. In particular, there is an agreement between Weatherbys and William Hill, pursuant to which Weatherbys supplies information from the BHB Database to William Hill. For this supply, William Hill and the other bookmakers pay a fee to Weatherbys, who in turn pays a fee to the BHB.

34. Until 1999, off-course bookmakers did not pay the BHB directly to use information from the BHB Database. Since 1999, a number of off-course bookmakers have paid the BHB directly to use pre-race information on the internet. However, at the date these proceedings commenced in 2000, other off-course bookmakers, including the three major off-course bookmakers and the government-owned Tote, had refused to pay licence fees to the BHB for the use of pre-race information on the internet, on the grounds that no such licence was necessary.

35. Some other users of pre-race information (such as the association of on-course bookmakers, electronic publishers and the Racecourse Association) pay the BHB directly for this information.

...

William Hill’s Internet service

40. These proceedings are concerned with a business which has recently been commenced by William Hill and a number of its competitors. It takes the form of providing betting services over the Internet. At the moment it is a minor part of the defendant’s business in turnover terms. William Hill established its first Internet site in June 1996 to promote its telephone betting business. In May 1999 it started betting on horseracing, limited initially to a small number of selected races each day on which William Hill had produced its own odds. It developed this into a comprehensive service covering the majority of horseracing, with real time changes in the odds being offered. This enhanced service was launched on two internet sites; the ‘International Site’ on 3 February 2000 and the ‘UK Site’ on 13 March 2000. Members of the public can access these sites over the internet, see what horses are running in which races at which racecourses and what are the odds offered by William Hill. If they want, they can place bets electronically. Other information (e.g. the jockey or trainer of the horse) is necessary for customers to arrive at an informed view of the horse’s chances of success. If the customer requires any of this information he must find it elsewhere, for example, in the newspapers. Annex F is an example of the sort of information available in the Racing Post newspaper in respect of a single race.

41. William Hill formulates and publishes its own betting odds for horseraces, referred to as Early Bird and ante-post odds. Early Bird odds are set by William Hill odds compilers using their own skill and judgment and are generally offered by it at the start of the day on selected races taking place the same day. William Hill currently gives Early Bird prices for approximately 2 000 horse races in the UK each year. Ante-post odds are those offered by William Hill on a specific race one or more days in advance of the race taking place. Five examples of what a user of William Hill’s internet service will see on his computer screen are annexed to this Schedule. The first (Annex A) was taken off the website on 13 March 2000 at 12.20 p.m. It relates to the 2.00 p.m. race at Plumpton to be held on that day. The horses’ names are the declared runners. The second (Annex B) was taken off the site on the same day and relates to the Grand National, which was to be run on 8 April 2000. The third (Annex C) was taken off the site a week later, i.e. on 21 March 2000, and also relates to the Grand National. Comparing the last two shows how the list of runners and the total number of runners can change as the date of the race approaches. In fact, not only the identity and number of the horses can change, but so can the timing of the race. Annex A is an example of a particularly small race with few runners. Some races are much larger. For example, as at 13 March 2000, the William Hill website showed the Lincoln Handicap, a 1 mile race to be held at Doncaster on 25 March 2000, had 58 proposed runners. By 21 March 2000 the site shows that the field had shrunk to 46. Printouts of the latter two website pages are to be found at Annexes D and E respectively.

42. Between May 1999 and February 2000, William Hill offered Internet betting on only selected races (those races for which it offered Early Bird and ante-post odds). Between 9.00 a.m. and 10.15 a.m. each day the runners for Early Bird races were entered manually together with their odds, with the relevant race data derived from race cards published in the national press. Potential runners in ante‑post races were entered manually from published lists. In both cases runners were displayed in the order of their odds with the shortest (lowest) appearing first. It is since February 2000 that William Hill has offered Internet betting on all mainstream horse racing in the UK. The relevant data for all races (including races where Early Bird odds are offered) taking place that day are currently derived from the RDF supplied to it by SIS and published between 5.00 a.m. and 7.00 a.m. on the day of the race, depending on when the RDF comes through each day. Where William Hill offers Early Bird or ante-post odds, the runners are listed in order of the odds offered. In other cases (or where such odds have not yet been formulated), William Hill offers the Starting Price, in which case the runners are listed in alphabetical order.

43. By the time William Hill publishes this data on its Internet sites (i.e. on the day of the race in question), it has been available from sources other than SIS since the morning of the day before. For example, it will have been published in the press and on various teletext services.

44. As can be seen from the annexes, the information displayed on William Hill’s internet sites comprises the names of all the horses in the race, the date, time and/or name of the race and the name of the racecourse where the race will be held. In terms of the number of records, this is a very small proportion of the total size of the BHB Database. No other information derived from the BHB Database appears on William Hill’s website. For example, William Hill does not display the jockey’s name, saddle cloth number or weight allocated to or to be carried by the horse. William Hill does not display any information relating to the racing form of any of the horses. Nor does it display any of the large amount of other information which is in the BHB Database and is used for the purpose of the Stud Book, the Jockey Club’s functions and/or the other functions of the BHB.

45. The horseraces are not arranged on William Hill’s websites in the same way as they are arranged in the BHB Database. Further, William Hill arranges the lists of runners by odds, with the favourite first, or alphabetically; they are not arranged in the same way as in the BHB Database, except perhaps by coincidence. However, each list of runners published on William Hill’s websites is the complete list of all runners for that race.

46. William Hill does not have direct access to the BHB Database. The information displayed on William Hill’s internet sites has in the past, and may in the future, be obtained by William Hill from two sources: (1) the evening newspapers, published before the day of the race; and/or (2) the RDF supplied by SIS on the morning of the race. The RDF is derived from the BHB Database. The information in the newspapers is also derived from the BHB Database: it is supplied to the newspapers by Weatherbys.

47. It is not in dispute that SIS and the newspapers have no right to sublicense William Hill to use any information derived from the BHB Database on its internet site and they have not purported to do so.

1 –

Original language: German.

2 –

Proceedings are also before the Court in Cases C-46/02, C-338/02 and C-444/02, in which I am also delivering my Opinion today.

3 –

OJ 1996 L 77, p. 20.

4 –

Judgments in Case 36/79 Denkavit [1979] ECR 3439, paragraph 12, Joined Cases C-175/98 and C-177/98 Lirussi and Bizzaro [1999] ECR I‑6881, paragraph 37, Case C-318/98 Fornasar and Others [2000] ECR I‑4785, paragraph 31, and Case C-421/01 Traunfellner [2003] ECR I-0000, paragraph 21 et seq.

5 –

See Case C-448/01 EVN [2003] ECR I-0000, paragraph 59.

6 –

Malte Grützmacher, Urheber-, Leistungs- und Sui-generis-Schutz von Datenbanken, 1999, 329; Georgios Koumantos, ‘Les bases de données dans la directive communautaire’, Revue internationale du droit d’auteur 1997, 79 (117). On the other hand, many writers see the investment as the object of protection (see, for example, Silke von Lewinski, in: Michel M. Walter (Ed.), Europäisches Urheberrecht 2001, paragraph 3 on Article 7, and the writings cited by Grützmacher on page 329 in footnote 14.

7 –

Giovanni Guglielmetti, ‘La tutela delle banche dati con diritto sui generis nella direttiva 96/9/CE’, Contratto e impresa. Europa, 1997, 177 (184).

8 –

Andrea Etienne Calame, Der rechtliche Schutz von Datenbanken unter besonderer Berücksichtigung des Rechts der Europäischen Gemeinschaften, 2002, 115 FN 554.

9 –

Grützmacher (cited in footnote 6), 330 et seq.; Matthias Leistner, Der Rechtsschutz von Datenbanken im deutschen und europäischen Recht, 2000, 152.

10 –

Leistner (cited in footnote 9), 152.

11 –

Guglielmetti (cited in footnote 7), 184; Gunnar W. G. Karnell, ‘The European Sui Generis Protection of Databases’, Journal of the Copyright Society of the U.S.A., 2002, 993.

12 –

As regards the views put forward, see P. Bernt Hugenholtz, ‘De spin-off theorie uitgesponnen’, AMI − Tidschrift voor auteurs-, media & informatierecht 2002, 161 (164 FN 19).

13 –

Von Lewinski (cited in footnote 6), paragraph 5 on Article 7.

14 –

Common Position (EC) No 20/95, adopted by the Council on 10 July 1995, No 14.

15 –

Jens-Lienhard Gaster, Der Rechtsschutz von Datenbanken, 1999, paragraph 492.

16 –

Oliver Hornung, Die EU-Datenbank-Richtlinie und ihre Umsetzung in das deutsche Recht, 1998, 116 et seq.; Leistner (cited in footnote 9), 180; von Lewinski (cited in footnote 6), paragraph 16 on Article 7.

17 –

Common Position (EC) No 20/95 (cited in footnote 14), No 14.

18 –

See inter alia von Lewinski (cited in footnote 6), paragraph 15 on Article 7.

19 –

Grützmacher (cited in footnote 6), 340.

20 –

Gaster (cited in footnote 15), paragraph 495; Grützmacher (cited in footnote 6), 340; von Lewinski (cited in footnote 6), paragraph 15 on Article 7.

21 –

Josef Krähn, Der Rechtsschutz von elektronischen Datenbanken, unter besonderer Berücksichtigung des sui-generis-Rechts, 2001, 162.

22 –

See Guglielmetti (cited in footnote 7), 186; Krähn (cited in footnote 21), 161, and Leistner (cited in footnote 9), 172.

23 –

In that regard, on some views, a theoretical likelihood of detriment is sufficient, see Leistner (cited in footnote 9), 173; see Herman M. H. Speyart, ‘De databank-richtlijn en haar gevolgen voor Nederland’, Informatierecht – AMI 1996, 171 (174).

24 –

Carine Doutrelepont, ‘Le nouveau droit exclusif du producteur de bases de données consacré par la directive européenne 96/6/CE du 11 Mars 1996: un droit sur l’information?’, in: Mélanges en hommage à Michel Waelbroeck, 1999, 903 (913).

25 –

Doutrelepont (cited in footnote 24), 913; Gaster (cited in footnote 15), paragraph 496; Leistner (cited in footnote 9), 171; von Lewinski (cited in footnote 6), paragraph 15 on Article 7.

26 –

See, however, Karnell (cited in footnote 11), 1000; Krähn (cited in footnote 21), 163.

27 –

See, for example, the eighth, 41st, 42nd, 45th and 46th recitals.

28 –

See Koumantos (cited in footnote 6), 121.

29 –

Von Lewinski (cited in footnote 6), paragraph 19 on Article 7.

30 –

Gaster (cited in footnote 15), paragraph 512.

31 –

Von Lewinski (cited in footnote 6), paragraph 27 on Article 7.

32 –

Von Lewinski (cited in footnote 6), paragraph 31 on Article 7.

33 –

Grützmacher (cited in footnote 6), 336.

34 –

Von Lewinski (cited in footnote 6), paragraph 38 on Article 7.

35 –

Leistner (cited in footnote 9), 181; von Lewinski (cited in footnote 6), paragraph 18 on Article 7, FN 225.

36 –

Sam Ricketson, The Berne Convention for the Protection of Literary and Artistic Works: 1886-1986, 1987, 482.

37 –

Gaster (cited in footnote 15), paragraph 558.

38 –

Most of the Romance languages, and the German, English and Greek versions.

39 –

The Spanish, Swedish and Finnish versions.

40 –

Leistner (cited in footnote 9), 181; von Lewinski (cited in footnote 6), paragraph 17 on Article 7.

41 –

That is also consistent with the interpretation of Article 13 of the TRIPs Agreement by a WTO Panel (WT/DS160/R of 27 July 2000, 6.183).

42 –

Leistner (cited in footnote 9), 181.

43 –

See WT/DS160/R of 27 July 2000, 6.186 (cited in footnote 41).

44 –

See WT/DS160/R of 27 July 2000, 6.229 (cited in footnote 41).

45 –

Leistner (cited in footnote 9), 182.

46 –

Standing Committee on Copyright and Related Rights (19 May1998), SCCR/1/INF/2.

47 –

Common Position (EC) No 20/95 (cited in footnote 14), No 14.

48 –

Von Lewinski (cited in footnote 6), paragraph 5 on Article 10.

49 –

Simon Chalton, ‘The Effect of the E.C. Database Directive on United Kingdom Copyright Law in Relation to Databases: A Comparison of Features’, E.I.P.R. 1997, 278 (284); Hornung (cited in footnote 16), 173 et seq.; Leistner (cited in footnote 9) 209; see St. Beutler, ‘The Protection of multimedia products under international law’, UFITA 1997, 5 (24); Guglielmetti (cited in footnote 7), 192; Herman M. H. Speyart (cited in footnote 23), 171 (173).

50 –

COM(92) 24 final.

51 –

Explanatory Statement to Proposal COM (92) 24, No 9.2.

52 –

COM(93) 464 final.

53 –

Grützmacher (cited in footnote 6), 390 et seq.

54 –

See, on that point, at greater length, Leistner (cited in footnote 9), 207 et seq.