England and Wales High Court (Patents Court) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION
PATENTS COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		9 February 2001

Mr Justice Laddie:

Introduction

1. This is a case about a new type of intellectual property called database right. It was introduced into our law by means of the Copyright and Rights in Databases Regulations 1997 (SI 1997/3032). This implements European Parliament and Council Directive 96/9/EC of 11 March 1996 on the Legal Protection of Databases ('the Directive'). The parties agreed that the Regulations have to be construed consistently with the Directive and, for the purpose of these proceedings, attention was only paid to the provisions of the latter.
2. Originally, there was only one claimant in this action, The British Horseracing Board Limited ("BHB"). However at the beginning of the trial, two other claimants, The Jockey Club and Weatherbys Group Limited, were added by agreement, thereby disposing of a minor issue concerning title which had been raised by the defendant, William Hill Organization Limited ("William Hill"). As will be apparent from the names of the parties, this case is concerned with horseracing and, in particular, the extent to which, if at all, the claimants can prevent the defendant from using, in a new part of its business and without their licence, certain data which, according to the claimants, have been derived indirectly from them. For the purpose of this judgment I will refer only to BHB and not the other claimants, save where necessary.

BHB and its Database

3. As explained by Mr Paul Greeves, its Racing Director, BHB, is the governing authority for the British racing industry. The members of BHB comprise the Jockey Club, the Racecourse Association Limited, the Racehorse Owners Association and the Industry Committee (Horseracing) Limited. It is a company limited by guarantee and 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 with responsibility for integrity, licensing, discipline and security. Its concern is the application and operation of the Rules of Racing. BHB took on the remainder of the administrative functions of racing's governing body. Those functions, as set out in BHB's Memorandum and Articles of Association include: (a) improvement of the financial position of horseracing; (b) funding of the administration of horseracing; (c) encouraging the maintenance and improvement of standards in horseracing, and exercising regulatory control; (d) encouraging and improving the breeding of bloodstock; (e) establishing the dates of fixtures and the programme content of fixtures; (f) initiating and promoting improvements in the law and local rules, regulations or practices; (g) making and publishing rules of practice and procedure for horseracing; (h) creating the fixture list for each year's racing; (i) weight adding and handicapping; (j) supervision of race programmes; (k) producing various racing publications and stakesbooks; and (l) compiling data related to horseracing.
4. It will be appreciated that the tasks undertaken by BHB are extensive. The scale of its operation is a reflection of the size of horse racing in this country. For example in the year 2000, there were a total of 1,209 separate race programmes and fixtures in the calendar. There will be 1213 race programmes in 2001. These races are held at 59 racecourses. In 2000, race meetings were scheduled on 327 days of the year. At any one time there are 15,000 horses in training, 9,000 active owners, and 1,000 trainers. Each owner must have registered racing colours in which his horses will run. These colours must be unique.
5. BHB undertakes a four to five month annual process compiling, granting and publishing the Fixture List. Publication of the fixture list in July of the preceding year marks the start of the full process which culminates in weekly advertisement in the Racing Calendar of the races for which entries are taken centrally under the Orders and Rules of Racing. Some 7,800 races are run each year, the exact number dependent upon race divisions and on the level of abandonments due to poor weather. In 2000 there were 175,000 entries for races. This number falls for a variety of reasons. For example sometimes trainers withdraw horses from races and sometimes races are oversubscribed, resulting in some entries being rejected. Shortly before a race, trainers must make "declarations" of the horses which will actually run in a particular race. In 2000, there were 80,000 declarations to run and 80,000 related declarations of riders.
6. In support of its functions, BHB maintains a computerised collection of information (which the parties called the "BHB Database") which is constantly being updated with the latest information. Computerisation began in 1964 when the Jockey Club had the functions now controlled by BHB. The scale and complexity of the data kept by BHB has grown with time. According to the particulars of claim:

"The establishment of the BHB Database has involved, and its maintenance and development continues to involve, extensive work, including the collection of raw data, the design of the database, the selection and verification of data for inclusion in the database and the insertion and arrangement of selected data in the database. The cost of establishing the BHB Database was considerable. The cost of continuing to obtain, verify and present its contents is approximately £4 million per annum and involves approximately 80 employees and extensive computer software and hardware."

There has been no substantial challenge to the accuracy of this statement.

7. There are a huge number of records contained within the BHB Database including many which must be accurately stored and processed each day. It covers not only all of the information set out above but much more. 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, the amount of money the racecourse is to contribute to the prize money for the race, the initial name of the race and the like. Close to the day of a particular race, the pre-race information is expanded to include the time at which the race is provisionally scheduled to start, the final name of the race including, where applicable, sponsor's name, the list of horses entered in the race, their owners and trainers and the relative weights these horses will carry. The final stage of pre-race information includes the list of intended or "declared" runners, their riders, the absolute weight each will carry, its saddle-cloth number, the stall from which it will start and its owner's racing colours. The BHB computerised records contain all of this information. After each race, details of the outcome are recorded. An estimated total of 800,000 new records or changes to existing records are made each year.
8. Maintaining the BHB Database does not consist only of receiving and entering data in BHB's computer. It involves extensive checking of data obtained from a number of sources. Evidence in relation to this issue was given by Dr Khan and Mr Peter Clarke on behalf of BHB. No purpose would be served by setting out in this judgment the detail of this collection and verification process, but some flavour of it can gathered by reference to part of the process undertaken for or on behalf of BHB in relation to declarations made by trainers. Such declarations have to be made by a fixed deadline shortly before the race is due to be run. They are normally made by telephone. As Dr Khan explained:

"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."

9. This typifies the care taken to ensure that the data on the BHB computer are accurate. In relation to pre-race information, for example, the unchallenged evidence of Dr Khan was;

"The pre-race information compiled for each race is the product of a painstaking process of verification which is aimed to ensure that the information is wholly accurate and reliable. Without there being virtually 100% accuracy in the information presented for each race, the confidence of those involved in racing, including bookmakers and their customers, could not be assured."

10. As can be seen from paragraph 2 above, maintenance of the BHB Database is only part of BHB's function. It currently costs BHB £15 million per year to carry out all of its functions on behalf of the British horseracing industry. The cost of running the BHB Database therefore accounts for about 25% of BHB's total expenditure. As Mr Greeves explained, 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 £1M annually, thereby meeting somewhere in the region of 25% of BHB's costs of maintaining the BHB Database. Mr Greeves said in his witness statement:

"The racing information contained in the BHB database is of interest to a wide variety of different users of the information. First, and foremost, essential extracts from the database are made available to the participating elements of 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. The information is made available to these parties each day by way of the joint Weatherbys/BHB Internet website @ http://www.bhb-weatherbys.racingadmin.co.uk and via a database site on the Prestel network, plus each week within BHB's official journal, the Racing Calendar. In addition, the racing information is of interest to radio and television broadcasters who cover different race meetings around the country; similarly, it is used by publications such as The Racing Post, Timeform and local and national newspapers who need to provide to their readers information concerning forthcoming race meetings; information services such as Ceefax and Teletext also need access to the racing information; the information is also, of course, of interest to members of the public who follow horseracing."

11. The information, or selected parts of it, is also supplied to other interested parties, including bookmakers. Among the methods of distribution, two are relevant to the issues in this case. 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, on behalf of BHB, makes available to subscribers in electronic form, normally on the day before a race, what is called a Declarations Feed. This contains an accurate, up-to-the-minute 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 data are supplied to a company, Satellite Information Services Limited ("SIS"), which is allowed to use data from it for certain purposes including for onward transmission to, and use by, its own subscribers. The supply from SIS to its own subscribers takes the form of what is called a raw data feed ("RDF").

William Hill

12. For the purpose of this action William Hill provided witness statements from two of its senior employees; Mr William Haygarth, the Managing Director of its e-commerce division, and Mr Michael Ellis, the Information Services Director of the William Hill Group of which the Defendant in this action is a part. Neither witness was required to attend for cross-examination and the content of the following paragraphs is derived from their evidence.
13. The William Hill business is one of the leading providers of off-track bookmaking services in the United Kingdom and elsewhere, to both UK 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 ("LBOs") and (b) telephone betting operations. In 1999 William Hill earned profits of £78.9 million. William Hill's principal product is the taking of fixed-odds bets on sporting and other events. The most popular event on which William Hill offers odds is horse racing. William Hill is one of the UK's largest LBO operators with currently 1,526 trading units. LBOs provide customers with comprehensive sports information and betting services including horse racing, greyhound racing and other significant sporting events. SIS provides live audio and video coverage of horse and greyhound racing. William Hill provides additional pre-event information for each race in the form of a text service displayed on television screens, which includes the name, place and time of each race and the names of the runners. During the course of the day the screens are regularly updated with the latest betting odds and race results. The same information is also available in each LBO in the form of pages from the "betting office" display edition of the Racing Post, which are posted on the walls. The Information in this form is far more comprehensive than that displayed on the screens as it also includes details of the owner, trainer, weight carried and form of each runner in every race and other information such as colours. A copy of the news-stand edition of Racing Post is also available in each LBO.
14. William Hill's LBOs and its main operations headquarters in Leeds are connected via land-line communications links that are used to transmit statistical and other information (such as risk management information) between LBO managers and the William Hill headquarters. William Hill subscribes to the SIS audio and video service for each of its LBOs, which it augments by overlaying its own promotional audio material. William Hill supplies its own text service to its LBOs via a satellite channel provided by SIS. The text service contains information supplied by SIS, together with William Hill's own material. The other principal channel for William Hill's bookmaking services is via the telephone, by which it offers bookmaking services from businesses established in the UK, in Antigua and in the Isle of Man.
15. It will be appreciated that a very great deal of the information displayed or used by William Hill in the horse-racing related business referred to above comes, directly or indirectly, from the BHB Database. No objection is taken to any of such use. In respect of all of it BHB takes the view that such data are all taken with its express or implied consent and for all of it BHB receives, directly or indirectly, financial compensation. For example, William Hill gets the essential information for its telephone betting service from the Declarations Feed. It is licensed to use the information in this way and for this purpose.

The current dispute

16. This action is concerned with a form of 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 horse racing, with real time changes in the odds being offered. This enhanced service was launched on two internet sites; the "International Site" on 3^rd February 2000 and on the "UK Site" on 13^th 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.
17. William Hill formulates and publishes its own betting odds for horse races, 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 2000 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. A number of examples of what a user of William Hill's internet service will see on his computer screen are to be found in the evidence used at the trial. Three of these are annexed to this judgment. The first (Annex A) was taken off the website on 13 March 2000 at 12.20 pm. It relates to the 2.00 pm 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. The third (Annex C) was taken off the site a week later, ie on 21 March, 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, the William Hill website showed the Lincoln Handicap, a 1 mile race to be held at Doncaster on 25 March, had 58 proposed runners. By 23 March the site shows that the field had shrunk to 46. The prints of the latter two website pages are not annexed to this judgment but are to be found in Trial Bundle B.
18. 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 am and 10.15 am 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 horseracing in the UK. The relevant data for all races (including races where Early Bird odds are offered) taking place that day are derived from the RDF supplied to it by SIS and published between 5.00 am and 7.00 am depending on when the RDF comes through each day, with the runners listed in alphabetical order. By the time William Hill publishes this data on its Internet sites, it has been available from sources other than SIS since the previous morning. For example it will have been published in the press and on various teletext services.
19. It is not disputed that the information displayed on William Hill's internet sites, that is to say the identity of all the horses in the race, the date and time of the race and the identity of the racecourse where the race will be held, is to be found in and comes from the RDF feed received by it from SIS. It is also not in dispute that the same information is supplied to SIS from the data stored on BHB's computer. It is information to be found in the Declarations Feed made available through Racing Pages. It is not in dispute that SIS has no right to sublicense William Hill to use any of BHB's data on its internet site and has not purported to do so. As will be explained hereafter, it is not admitted that the SIS information used by William Hill originated with the information supplied to SIS from the BHB Database, even though it is identical to it.
20. BHB's case is that it owns database right in the BHB Database and that William Hill is making unlicensed use of that data in its internet business. It says that what William Hill takes from the SIS RDF is derived by SIS from the BHB Database. It says that William Hill's activities constitute breaches of BHB's database right in two ways. First it says that each day's use by William Hill of data taken from the SIS RDF is an extraction or re-utilization of a substantial part of the contents of its database contrary to art 7(1) of the Directive. Second, it says that, even if the individual extracts are not substantial, nevertheless the totality of William Hill's actions amount to repeated and systematic extraction or re-utilization of insubstantial parts of the contents of the database contrary to art 7(5).
21. William Hill advances a number of defences to this claim. As I have already mentioned, that based on lack of title is no longer pursued as a result of the addition of the second and third claimants to the action. As I understand its position, William Hill accepts that on, say, 13 March 2000, the BHB Database qualified as a collection of data in respect of which database right exists. I do not understand it to be disputed that the BHB Database as it was at, say, the same time on 21 March was also a collection of data in respect of which database right exists. As we shall see later a potentially important point arises as to whether on these two dates, taken by way of example only, the BHB Database is the same database for the purposes of this area of law. William Hill also indicates that if its current activities infringe any rights owned by BHB, it has alternative methods of presenting the essential race data on its internet sites which it believes would clearly avoid such rights. BHB disputes this. Therefore the parties have addressed not only what William Hill currently is doing but also what it threatens to do.

An overview of database right.

22. The reason for introducing a new database right into the domestic law of the Member States of the European Union is largely explained in the recitals to the Directive. Databases used to be protected by copyright in all or most states. Unfortunately there existed major differences between the relevant national laws. A collection of data which could be protected in one State might not be protected at all in another, or the scope of protection might be different. In an attempt to resolve some of the difficulties created by this lack of uniformity in national laws, the Directive does two things. First it requires Member States to implement certain common features in their national copyright law insofar as they deal with databases. These are set out primarily in arts 3 to 6 inclusive. Second, it creates an entirely new kind of right, which it refers to as "sui generis". This is dealt with primarily in arts 7 to 11 inclusive. This is the database right. It is independent of any copyright or other intellectual property rights which may exist in the database or in any of the individual pieces of data or information collected together within the database. In this case BHB say that it may well be entitled to copyright protection as well as database right, but it is content to sue solely on the latter and the proceedings are limited accordingly.
23. The fact that database right and copyright in databases can exist side by side and that the former is described as sui generis is important. Although it is apparent that there are some features of the database right which are similar to features of copyright, it must not be assumed that the former is based upon or is to be construed as a mere continuation or development of the latter and, in particular, that it is a mere variation of United Kingdom copyright law. There may be a natural tendency, particularly for those familiar with copyright, to look at database through copyright eyes, but there are significant differences between the two rights. They may have concepts in common, but, if so, that is only because those concepts happen to fit both, not because database is a species of copyright. This means that courts have to guard against the assumption that principles which have become familiar in the copyright field automatically apply to the new right. The existence and scope of the new right has to be determined from the Directive itself, and any admissible preparatory texts.

The nature and effect of database right

(a) What is a database?

24. Although William Hill is now prepared to accept that database right exists in the BHB Database (subject to the issues addressed in paragraph 65 et seq below), it is still necessary to say something about the nature of a database which qualifies for protection because this has an impact on the other issues between the parties.
25. Originally, it had been proposed that the Directive should only cover collections of information in electronic form. This restriction was abandoned so as to extend protection to databases in a non-electronic form (see Recital 14). Accordingly the expression 'database' is defined in art 1 as follows:

"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. Protection under this Directive shall not apply to computer programs used in the making or operation of databases accessible by electronic means."

26. The breadth of this definition is reinforced by Recital 17;

"Whereas the term 'database' should be understood to include literary, artistic, musical or other collections of works or collections of other material such as texts, sound, images, numbers, facts, and data; whereas it should cover collections of independent works, data or other materials which are systematically or methodically arranged and can be individually accessed; whereas this means that a recording or an audio visual, cinematographic, literary or musical work as such does not fall within the scope of this Directive;" …

27. This is very broad. Any two or more pieces of data put side by side could be said to be a collection and, therefore, a database. This was not intended, hence the exclusion of recordings, audio visual, cinematographic, literary or musical works. Although, for example, a musical work may involve an arrangement and juxtaposition of musical notes, it is treated as a single work and, for this reason, not a collection. However not all databases are protected by copyright or database right. In relation to copyright, this can be seen in Recitals 15 and 16;

"(15) Whereas the criteria used to determine whether a database should be protected by copyright should be defined to the fact that the selection or the arrangement of the contents of the database is the author's own intellectual creation; whereas such protection should cover the structure of the database;

(16) Whereas no criterion other than originality in the sense of the author's intellectual creation should be applied to determine the eligibility of the database for copyright protection, and in particular no aesthetic or qualitative criteria should be applied;"

28. Therefore, for copyright to subsist, it must be shown not only that there is a relevant collection of information but that it is also original. Although there is no requirement to demonstrate aesthetic or qualitative criteria, there must be a quantitative baseline of originality before protection is acquired. This can be seen in Recital 19 which also demonstrates an equivalent baseline requirement for the existence of database right (a topic I will consider below);

"Whereas, as a rule, the compilation of several recordings of musical performances on a CD does not come within the scope of this Directive, both because, as a compilation, it does not meet the conditions for copyright protection and because it does not represent a substantial enough investment to be eligible under the sui generis right;".

29. The only 'condition for copyright protection' which is not met must be one of sufficient originality. There is another feature of the definition of database which needs to be considered in view of some of the arguments advanced in this case. Art 1 refers to collections in which the materials are "arranged in a systematic or methodical way and individually accessible". This might be thought to mean that for all purposes a database must be arranged systematically or methodically. However, if that were right it might exclude many computerised databases. Recital 21 shows that this cannot be what was intended;

"(21) Whereas the protection provided for in this Directive relates to databases in which works, data or other materials have been arranged systematically or methodically; whereas it is not necessary for those materials to have been physically stored in an organized manner;"

30. It seems, therefore, that the expression 'database' has a very wide meaning covering virtually all collections of data in searchable form. On any view, the BHB Database at a particular point in time, say 12.20 pm on 13 March 2000, is a 'database' within art 1. Whether a database qualifies for protection by copyright, database right or both depends upon how the database was made, that is to say whether sufficient relevant effort was put into its creation. To determine this involves understanding what is the purpose of the intellectual property rights granted in a database.

(b) The objective behind the creation of database right

31. Understanding the objective behind the creation of database right is important because, as both parties agree, correctly in my view, the purpose throws light on what types of activity amount to breaches. The Directive repeatedly refers to protecting and encouraging investment in database creation. There are 11 such references in the Recitals (see Recitals 1, 10 –12, 19, 39 – 42, 54 and 55) and it is also referred to in arts 1 and 10, which are considered below. Further it is apparent what type of investment is considered worthy of protection. Recital 40 and art 7(1) respectively provide:

"(40) Whereas the object of this sui generis right is to ensure protection of any investment in obtaining, verifying or presenting the contents of a database for the limited duration of the right; whereas such investment may consist in the deployment of financial resources and/or the expending of time, effort and energy;"

and

"7. Object of protection

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-utilization of the whole or of a substantial part, evaluated qualitatively and/or quantitatively, of the contents of that database."

32. In the light of this it appears that for database right to exist, there must be investment in its creation and, in particular, that investment must be directed at obtaining, verifying or presenting the contents. As is apparent from art 7(1), the investment must be substantial enough to justify protection. However the terms of Recital 19, set out at paragraph 28 above, indicate that the qualifying level of investment is fairly low. In any event, whatever the level, it is not suggested that the investment in the BHB Database falls below it.
33. Of more significance to this dispute is the type of investment involved. As one would expect, effort put into creating the actual data which is subsequently collected together in the database is irrelevant. This is confirmed by art 7(4) which draws a distinction between rights in the database and rights in the data within the database:

"The right provided for in paragraph 1 shall apply irrespective of the eligibility of that database for protection by copyright or by other rights. Moreover, it shall apply irrespective of eligibility of the contents of that database for protection by copyright or by other rights. Protection of databases under the right provided for in paragraph 1 shall be without prejudice to rights existing in respect of their contents."

34. For this reason, the costs and effort involved in BHB fixing the date of a racing fixture does not count towards the relevant investment to which database right is directed. On the other hand, the efforts which go into gathering all the data together, including the dates of fixtures, is relevant. Such activities constitute "obtaining" the data, one of the types of investment referred to in art 7(1) and Recital 40. In practice where one person both creates the underlying data and gathers it together, as BHB does, it may be difficult to draw a sharp dividing line between the two activities.
35. Investment in 'verification' is also relevant to the subsistence and protection of database right. Verification consists of ensuring the accuracy of a collection of data. Thus even if the content or form of the collection of data are substantially unchanged, effort put into ensuring its accuracy, or continuing accuracy, is relevant and is to be protected. This point is reinforced by Recital 55:

"(55) Whereas a substantial new investment involving a new term of protection may include a substantial verification of the contents of the database"

36. The latter Recital indicates that even if the contents of a database do not change substantially, if sufficient investment is put into ensuring that it is up to date and accurate, it is protected by the new right.
37. Finally, investment in 'presentation' also counts. It appears to me that this must cover at least the effort and resources put into making the data more readily accessible by the user. Effort put into the design of the layout of the information should count. It should be noticed that art 7(3), which is set out in paragraph 25 above, provides that protection under the Directive does not apply to computer programs used in the making or operation of databases accessible by electronic means. It is not entirely clear how one is to determine where the borderline lies between investment put into presentation and investment put into designing computer programs which make the data more readily searchable. Fortunately that is an issue which does not arise in this case.

(c) Breach of database rights

38. As mentioned at paragraph 19 above, BHB puts is case on infringement in two ways. However before considering each of those separately, one general issue has to be addressed. Breaches under both art 7(1) or (5) of the Directive involve the claimant in proving derivation. That is to say the claimant has to prove that the defendant is misusing in a relevant manner parts of its database rather than data obtained from some other, independent, source. This is a live issue in this case because William Hill says that BHB has failed to prove derivation. I have already explained that the information on the former's websites is admitted to be derived from the SIS RDF. This is identical in content to some of the data supplied to SIS by BHB from the BHB Database. It is also identical to the information supplied through Racing Pages in the Declarations Feed which William Hill uses for its telephone betting service. William Hill says that, although this is so, it cannot be assumed and it is not proved that the information it obtains from SIS is derived from the information supplied to SIS by or on behalf of BHB because it is possible that SIS obtained the same relevant information from another source. For the purpose of this action it is only necessary to consider the data relating to imminent races, that is to say information relating to the declared runners.
39. William Hill argues that some at least of the information it takes from the RDF may well have been obtained by SIS directly from the various race courses up and down the country and BHB has not proved otherwise. This argument is based upon the existence and terms of an agreement between the Racecourse Association Limited and SIS (the "RCA Agreement"), a copy of which is exhibited to the witness statement of Mr Robert Ricketts. The Racecourse Association Limited is the representative of the operators of racecourses throughout Great Britain. SIS, for the purpose of supplying real-time broadcasts of race meetings to its subscribers, which includes companies like William Hill, must be able to have broadcasting equipment on site at each racecourse it intends to send race information from. To enable it to do this, it has entered into the RCA Agreement, which, amongst other things, allows it to send its employees and their transmitters to various races. Clause 6.1.1. of that agreement obliges all Racecourse Association Limited members, save in exceptional circumstances, to send to SIS three months in advance details of race days, races and race times at each racecourse. This, says William Hill, may be the source of much of the information on the RDF which it uses on its internet site.
40. It should be noticed that even if this argument were valid, vital information in the RDF, namely the identity of the runners, could not come from this source. The Racecourse Association Limited members are under no obligation to supply that information to SIS. But even in relation to the rest of the information, this argument fails. The purpose of the RCA Agreement is to allow SIS to ensure that its staff are at the right place on the right day. To enable it to do this, SIS needs to know well in advance which racecourses are holding meetings on which days. It is for that reason that Clause 6.1.1. requires Racecourse Association Limited members to furnish SIS with this data three months in advance. Precise details of which races will run at which times and with which runners is irrelevant to this exercise and the RCA Agreement imposes on members no obligation to supply them. It does not require any part of the information contained in the Declarations Feed or the RDF to be supplied by such members. Furthermore, all the information on the racing calendar eventually is derived from BHB. This was explained by Mr Greeves during his cross-examination:

"Q. Do you suggest, wearing your BHB hat, that the racecourses cannot without your permission tell anybody what races they are running from time to time and day to day? Is it a secret? Is it something that you go out of your way to patrol and prevent?

A. That is certainly not my intention to convey as the situation, but, as I said just a moment ago, the information which is made up of race days, races, race times, on each racecourse, can only be taken as official and correct and final and absolute by any users once it has been through all the procedures that BHB carries out with racecourses. We work with them to compile a fixture list but we compile it, we take decisions on whether a fixture shall be placed on a particular day or not, bearing in mind other applications. We take decisions on whether races, which the racecourse may submit, is acceptable or not. It might be against the orders and rules of racing and its content. It might be possible that they wish to race on a day over two miles and it is a hurdle and another course on that day has an identical race that we prefer for one reason or another. There is a process of verification and compilation by BHB as the racing authority, the governing authority, putting it all together. So the racecourse, taking your point, would market its product, advertize its races, but, even when it does so, for instance, they produce their own individual fixture list, Haydock Park might extract from the fixture list and publish its fixtures but you will see on any such fixture list copyright of the British Horseracing Board.

Q. Has that always been on there or is it just the last year or so?

A. It has been on there for as long as I can remember, to my knowledge, and no reason why racecourses would not do so.

JUDGE: Just to help me, it has been there as long as you can remember. How long can you remember?

A. I think I mentioned yesterday I have been lucky enough to be involved in racing, as my career, for nearly 30 years. I became involved in this area about four years after I joined the industry when I was appointed as editor of the Racing Calendar and I well remember back in those days the fact that the fixture list was a copyright item. Anybody who wants to reproduce the fixture list has to enter a copyright agreement with us. They may only pay a small charge for it because we want them to advertize the fact that racing is on so we do not make a prohibitive or go out of our way to make it difficult: quite the opposite."

41. In addition to this, Mr Smith, SIS's company secretary, gave evidence of the source of the data in the RDF. He confirmed that the data contained in it before the start of racing comes from Racing Pages (ie from BHB). This is set out in the second exhibit to his witness statement and was confirmed orally before me. Mr Smith trained as a solicitor and does not profess to have technical knowledge but he explained that his description of where SIS obtained its information from was prepared in consultation with SIS's Software Development Manager. Although Mr Platts-Mills challenged this on the ground that it was hearsay, there appears to me to be no basis upon which to doubt the accuracy of what Mr Smith said. William Hill never asked BHB to produce the Software Development Manager to give direct evidence. Furthermore, on a balance of probabilities, it is most unlikely that SIS, which was supplied with definitive up to the minute, accurate data from BHB, would ignore that, and instead use only such race information as it might receive from each of the 59 racecourses in Great Britain. In my view it is virtually certain that the data in the RDF was derived directly or indirectly from the BHB Database and it follows that the defendant's argument on this issue fails.

Infringement by extraction or re-utilization of a substantial part

42. Art 7(1) provides that one of the purpose of database right is "to prevent extraction and/or re-utilization of the whole or of a substantial part, evaluated qualitatively and/or quantitatively, of the contents of" the database. BHB argues that William Hill's activities breach these rights in that the information they have taken via the RDF represents a substantial part of the contents of the BHB Database. It is, if anything, the most vital information. Although the BHB Database contains much more, the identities of which horses are racing in which races at what time at which racecourses are the core data to which everything else is secondary. Furthermore it is BHB's investment in ensuring that this information is up to date and accurate which William Hill is using for the purpose of its new commercial operation. William Hill wants to ensure that its customers can be confident that they are betting on horses which are racing at identified races. Accuracy is what they want and this is ensured by taking data which originates from the only definitive source, BHB.
43. Although Mr Platts-Mills' arguments are interrelated, they may be grouped under four headings. (1) What William Hill has used is not a part, in the relevant sense, of the BHB Database, (2) even if it is a part, it is not a substantial part, (3) the use does not amount to an 'extraction' from the BHB Database and (4) it is not a 're-utilization' of that Database. It is the first of these which is the most fundamental submission and it goes to the very heart of what is protected by the new right.
44. Mr Platts-Mills says that one must distinguish between the data or information within a database and the characteristics which give rise to the new type of protection. Database right does not protect the information within a database per se. It is crucial that BHB cannot use any database right which it may own to prevent William Hill, or anyone else, from making use of any facts within its Database. Taking the facts, and only the facts, can never infringe database right, whether one fact is taken or all of them. What is protected is what he calls the "database-ness" of the collection of information.
45. "Database-ness" he describes in the following way: The features of a database which make it a database are described in art 1(2), namely

"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."

Since no right is created in the works, data or other materials, the 'database-ness' of a database must lie in the fact that the independent materials are arranged in a systematic or methodical way, and are individually accessible. Following from this he says that the acts amounting to infringement of a database must in some way take unfair advantage of this "database-ness". Any acts which do not make any use of the arrangement of the contents of the database, nor take advantage of the way in which the maker has rendered the contents individually accessible, cannot infringe the database right. It is in this sense that data within a database is not part of the database for infringement purposes. The defining characteristic of that which can be the subject matter of an infringement action is form or structure which makes it accessible to a searcher. Echoing the sentiments expressed by Lord Pearce in the well known copyright case, Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 273, Mr Platts-Mills says that that which would not attract database right except by reason of its arrangement and/or individual accessibility will, when robbed of that arrangement and/or individual accessibility, not be a part of the database right and therefore the courts will not hold its unlicensed copying or use to be an infringement. These propositions dovetail with a more general point, namely that database right is to be construed so as to be narrower than the protection which used to be afforded to compilations under English copyright law.

46. The first of these points, namely that one must distinguish between database right on the one hand and the rights in the individual pieces of data or information on the other, is supported by a number of Recitals and Articles. For example arts 27, 45, 46 and 58 are respectively in the following terms:

"(27) Whereas copyright in such works and related rights in subject matter thus incorporated into a database are in no way affected by the existence of a separate right in the selection or arrangement of these works and subject matter in a database;

(45) Whereas the right to prevent unauthorized extraction and/or re-utilization does not in any way constitute an extension of copyright protection to mere facts or data;

(46) Whereas the existence of a right to prevent the unauthorized extraction and/or re-utilization of the whole or a substantial part of works, data or materials from a database should not give rise to the creation of a new right in the works, data or materials themselves;

(58) Whereas, in addition to the protection given under this Directive to the structure of the database by copyright, and to its contents against unauthorized extraction and/or re-utilization under the sui generis right, other legal provisions in the Member States relevant to the supply of database goods and services continue to apply;"

Similarly art 13 provides:

"Continued application of other legal provisions

This Directive shall be without prejudice to provisions concerning in particular copyright, rights related to copyright or any other rights or obligations subsisting in the data, works or other materials incorporated into a database, patent rights, trade marks, design rights, the protection of national treasures, laws on restrictive practices and unfair competition, trade secrets, security, confidentiality, data protection and privacy, access to public documents, and the law of contract."

47. However, in my view none of this leads to Mr Platts-Mills' conclusion. A database containing, say, a large number of poems does not thereby create an additional right in each poem. Database right is concerned with collections of data. Furthermore nothing in the Directive supports the second part of Mr Platts-Mills' submission relating to "database-ness" which runs together two entirely distinct concepts, namely the feature of form which have to exist before a database will be recognised as existing and the features of content or investment which are protected once a database is held to exist. Thus a database consists of a collection of data brought together in a systematic or methodical way so as to be individually accessible by electronic or other means. In much the same way a literary work in copyright has to "written, spoken or sung" before it is recognised as suitable for protection but infringement is not restricted to acts of writing, speaking or singing. A collection of data in the mind of an author does not qualify for protection until it is put into a form where it is searchable. But the fact that it has to take this form before database right can apply to it does not mean that the database right protects that form. Indeed the Recitals to the Directive are quite explicit that the form of a database is what is protected by copyright not by the sui generis right:

"(38) Whereas the increasing use of digital recording technology exposes the database maker to the risk that the contents of his database may be copied and rearranged electronically, without his authorization, to produce a database of identical content which, however, does not infringe any copyright in the arrangement of his database;

(39) Whereas, in addition to aiming to protect the copyright in the original selection or arrangement of the contents of a database, this Directive seeks to safeguard the position of makers of databases against misappropriation of the results of the financial and professional investment made in obtaining and collection the contents by protecting the whole or substantial parts of a database against certain acts by a user or competitor;

(58) Whereas, in addition to the protection given under this Directive to the structure of the database by copyright, and to its contents against unauthorized extraction and/or re-utilization under the sui generis right, other legal provisions in the Member States relevant to the supply of database goods and services continue to apply;"

48. These recitals make it clear that infringement of the new right is not avoided by taking the contents and rearranging them. On the contrary, what has to be protected is not primarily the form but the investment which went into "obtaining, verifying or presenting the contents" of the database as made clear, not only by art 7(1) but also Recital 40 (see paragraph 31 above). It is for this reason that substantial investment in verification (where the form is substantially unchanged) still qualifies for database right in accordance with Recital 55 (see paragraph 36 above). In such a case, the infringer takes advantage of the relevant investment if he makes use of the accuracy of the data in the database, not because he takes it in a particular form. Furthermore Recital 42 also makes it clear that infringement is not limited to activities which create competing databases:

"(42) Whereas the special right to prevent unauthorized extraction and/or re-utilization relates to acts by the user which go beyond his legitimate rights and thereby harm the investment; whereas the right to prohibit extraction and/or re-utilization of all or a substantial part of the contents relates not only to the manufacture of a parasitical competing product but also to any user who, through his acts, causes significant detriment, evaluated qualitatively or quantitatively, to the investment;"

49. In my view Mr Platts-Mills' submission on this crucial issue is wrong. I would only add the following points. First, Mr Platts-Mills' arguments lead inevitably to his submission that taking all the contents of a database is not the taking of a 'part' of the database if the contents are not taken in substantially the same arrangement as in the claimant's product. There is nothing in the Directive which suggests that this surprising result was what was intended by the legislature. Second, the submission is made all the more unattractive in view of Mr Platts-Mills' concession, rightly made in my view, that with modern software any collection of data stored in computer memory is capable of being accessed and searched. It does not need to be in any particular form or arrangement, yet many such collections are undoubtedly databases. Third, I did not find it helpful or meaningful to describe database right as being narrower than copyright. They are different and have to be assessed individually. Fourth, this topic may be approached in an altogether more simple way. What are prohibited are certain kinds of use or manipulation of "parts of the contents" of the database. If one asks the question, "is a collection of data taken from the database a part of its contents?", the answer must be in the affirmative unless there is something in the Directive to give the words "part of its contents" a special meaning. There is nothing.
50. This leads to the next issue, substantiality. For the purpose of illustrating the arguments, it is convenient to consider what was on the William Hill internet site on 13 March 2000. The trial bundle contains a printout of this material. As I understand it, all horse racing on that and the following day is reported there as is a selection of races to be run at different times in the more distant future. For present purposes it is only necessary to consider the material derived from the RDF. Again, for convenience, I shall only consider the data relating to races to be run on 13 March itself. This covered the complete racing cards at three racecourses; Plumpton, Stratford and Taunton. In total 20 races were run, involving 192 different horses. The size of the individual races differed, the smallest including the 3.00 pm at Plumpton with 6 runners, the largest being the 4.20 pm at Stratford with 16 runners. It is noticeable that the times of the various races are interleaved so that no two races commence at the same time. Thus, for example, the 1.50 pm at Stratford is followed by the 2.00 pm at Plumpton then by the 2.10 pm at Taunton, the 2.20 pm at Stratford and so on.
51. Art 7(1) requires substantiality to be assessed "qualitatively and/or quantitatively". Mr Platts-Mills says that the information used by his client cannot be regarded as substantial on either basis. As far as quantity is concerned, he points to the fact that there is a vast amount of information in the BHB Database. The question of whether the defendant's use is quantitatively substantial must be looked at by comparing what has been taken or used with what is in the claimant's database, not by seeing how important the material is to the defendant. The same approach must be adopted in relation to the qualitative assessment of what has been taken. Once again, if one compares the totality of the important data held on the BHB Database with what is on the William Hill internet site it can be seen that the latter contains very little but the names of a few horses.
52. I accept Mr Platts-Mills' submission that this issue has to be assessed primarily by comparing what has been taken or used with what is in the claimant's database. However the importance of the information to the alleged infringer is not irrelevant. In some cases, of which this is an example, the significance of the information to the alleged infringer may throw light on whether it is an important or significant part of the database. If one of the purposes of the database is to service businesses of the same general type as that run by the alleged infringer with the same type of information taken by him, then the collection, verification and presentation of that type of information within the database is likely to be an important or substantial part of its contents.
53. Art 7(1) provides that substantiality is to be assessed by looking at the quantity and quality of what is taken but it does not require them to be looked at separately. It contemplates looking at the combination of both. This exercise does not admit of precision but, as Mr Prescott submits, in undertaking it the court must bear in mind that one of the objectives of the Directive is to protect the investment in obtaining, verifying or presenting the contents of databases. I have set out at paragraph 3 above, some of the primary functions of the BHB. At its heart, as its corporate name makes clear, is horseracing. The ultimate purpose of the BHB Database is to help the BHB control and facilitate horseracing and raise funds for horseracing derived from gambling. To this end, one, if not the, major purpose of the BHB Database is to ensure that all the relevant data relating to horse racing is accurately stored and available. It is the data relating to the races themselves which represents the ultimate and crucial information within the Database. Here what the defendant is doing is making use of the most recent and core information in the BHB Database relating to racing. William Hill is relying on and taking advantage of the completeness and accuracy of the information taken from the RDF, in other words the product of BHB's investment in obtaining and verifying that data. This is a substantial part of the contents. No useful purpose would be served by trying to assess this issue first on a quantitative basis and then, separately, on a qualitative basis. They should be looked at together.
54. However this finding does not dispose of the dispute. William Hill only breaches BHB's database right if its activities amount to an "extraction" or "re-utilization" of a substantial part. These two words are defined in art 7(2) as follows:

"(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-utilization' 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 right holder 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-utilization."

55. Mr Prescott says that these are not difficult provisions to understand. Extraction means copying. Re-utilization amounts to making the material available in one form or other to the public. Mr Platts-Mills urges a much more restrictive meaning for each expression. I shall take them separately.
56. According to Mr Platts-Mills, a thing which is extracted has to be extracted from somewhere. This involves an act of 'taking away'. Once it has been extracted or taken away from somewhere it cannot be taken away again (unless it is first put back there). Therefore it is only the first removal of the data that amounts to extraction. To be an infringement of database right, subsequent use of the extracted material must amount to re-utilisation thereof, if anything. He also says that what is removed has to have the same quality of 'database-ness' as the original. This latter point is a repetition of the argument considered above and I do not accept it for the reasons already given. I can therefore concentrate on the suggestion that the definition means first removal. Mr Platts-Mills illustrates this submission by reference to two examples. First he refers to what SIS does. It will be recalled that every day SIS, under licence, receives a stream of data from the BHB Database. This is pulled off the computer by Weatherbys. Mr Platts-Mills says that the only extraction is committed by Weatherbys. What SIS has got is a lot of data, but it is not the BHB Database and the only person who has removed anything from the latter is Weatherbys. He says that because this is so, there is nothing to prevent SIS making multiple copies of the data supplied to it, because making such copies does not involve any taking of information from the BHB Database. As shall be explained more fully below, Mr Platts-Mills also says that if, without licence, SIS makes the totality of what it receives from Weatherbys available to third parties, that is not re-utilization either. Therefore SIS can do what it wants with the data supplied to it by Weatherbys. Secondly he refers to the contents of Racing Post. This is a daily newspaper which is devoted to racing. It is supplied, under licence, with a vast amount of information about every race running on a particular day. This does not cover just which horses are racing in which races at which times at which racecourses. It also includes details of owners, colours, previous performances of each horse, handicapping, jockeys, trainers and so on. The vast majority of this information is then published in the newspaper for the information of the public. Mr Platts-Mills says that not only can the Racing Post make this available to anyone it likes, whether licensed or not, but that anyone else can copy the Racing Post in full and without licence. Neither activity would amount to an extraction.
57. I am not even convinced that these arguments are internally consistent. If, as suggested, extraction means removal or taking away data from the database so that they are not left behind, I can not see how what Weatherbys does when it supplies data to SIS can be categorised as extraction either. What Weatherbys does is read data from the BHB Database and pass that information on to SIS, it does not cut the information out of the Database so that it is no longer there. In any event, I have no doubt that these submissions are wrong. The Directive does not require that extraction should be direct rather than indirect, nor does the definition involve the concept of taking away. All that is required is that a substantial part of the contents be transferred to a new medium. Thus the definition refers to "transfer … to another medium". It says nothing about the resultant state of the database from which the transfer has been made. If someone takes a copy of the contents of a database and loads it onto a new medium, it is no less transferred to the new medium because the same data are left on the original database. In substance, therefore, I accept Mr Prescott's argument on this issue. The only qualification to be noted is the requirement that there be transfer "to another medium". A hacker who accesses a database without a licence, looks at the data and memorises it may well not be guilty of extraction if his actions do not involve the making of a copy of the data in material form.
58. As far as re-utilization is concerned, Mr Platts-Mills says that particular attention must be paid to the words "making available to the public" in the definition. He says that this must involve telling the public something it does not already know. In other words it only covers first publication. Once the contents have been made public, for example through the Racing Post, there can be no re-utilization because if the same information is supplied by an unlicensed source it is not making anything available to the public. The public already has it. This is of particular importance in this case because all the information available on the William Hill websites is but a repetition of information which is published in the Racing Post and, for example, in William Hill's own LBOs.
59. By the end of his submissions I was not sure whether Mr Platts-Mills adhered to this argument. In any event, in my view it is incorrect. If it were otherwise the results would be bizarre. For example, as soon as Racing Post published its daily edition, BHB would be unable to protest if unlicensed copies of it were put on the market. Mr Platts-Mills goes even further and says that the supply of all the relevant information by SIS to his client in the RDF was itself a publication, since his client is a member of the public. On this basis, if the whole of the BHB Database were communicated to William Hill, say by BHB, the latter would immediately lose any ability to restrain further use by the former. There is nothing in the Directive which suggests that the right to prevent re-utilization is restricted to secret data on a database, which would be the effect of this submission. What it is concerned with is the unlicensed use of data derived without permission from a database. Private use of such data are not treated as re-utilization but any use which transmits or makes available the extracted information to the public is covered by database right and can be restrained. If one asks the question, "is the information on the William Hill website being made available to the public by … on-line or other forms of transmission?" the answer must be in the affirmative. The fact that some or all of the data are available to the public from another source is irrelevant. There is no reason why information cannot be made available to the public from more than one source. If the draughtsman of the Directive had intended to cover only first publication, he surely would have said so explicitly and, had he done so, it would have undermined the very purpose of the Directive, namely the protection of the investment which goes into making databases.
60. It follows that William Hill's actions of taking information from the RDF and loading it onto its own computers for the purpose of making it available on its website is an unlicensed act of extracting a substantial part of the BHB Database and the subsequent transmission of that data onto its website for access by members of the public is a reutilization. The defendant infringes BHB's rights in both ways.

Infringement by repeated use of insubstantial parts

61. BHB also puts its case in an alternative way. It says that even if the totality of what William Hill takes from the RDF on a particular day is not, by itself, a substantial part of the BHB Database, nevertheless it can rely on art 7(5) which is in the following terms:

"The repeated and systematic extraction and/or re-utilization 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."

62. In this case, William Hill extracts the details of races, horses and times for each race course day by day and week by week throughout the racing season. The claimant argues that such activities clearly and unreasonably prejudice its legitimate interests, not least because a significant part of its income has historically been derived from selling racecard information to bookmakers and newspapers. In this way it derives an income from those who use that information to facilitate betting. For example, William Hill, both prior to the commencement of its new business and now, pays directly or indirectly for the racecard information derived from BHB and used in its telephone betting operations and in its LBOs. I understand that all other bookmakers in the country do likewise. The new internet business is an alternative method by means of which bookmakers make their services available to the public. Instead of a punter being required to walk into a William Hill LBO, the betting shop is made available to him over the internet on his own computer. The mechanics of allowing a punter to place a bet may have changed but the same critical information is used by William Hill for the same purpose. To allow it to use the information without payment on the internet site would undermine this well established part of BHB's operations. In the medium term this may well severely undermine the value of the licences granted for use of data in LBOs and in telephone betting operations.
63. Mr Platts-Mills advances a number of arguments against this. He repeats his 'database-ness' argument, saying that what his client has taken is not an insubstantial part, because of its small size and its lack of necessary database quality it is not a part at all. I have dealt with this above. He also runs the same arguments relating to extraction and re-utilization which I have dealt with under art 7(1). There are two other arguments which deserve separate consideration.
64. First, Mr Platts-Mills draws my attention to the wording of art 7(5). It is only repeated and systematic extractions or re-utilizations of insubstantial parts of the contents of the database which gives rise to this type of liability. He points to art 10(3) which reads as follows:

"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."

65. He says that the evidence of the enormous amount of work done by BHB in updating and continuous verification of the BHB Database means that from week to week or, perhaps, day to day new databases come into existence each of which is protected for its own new term. On this basis, what William Hill is doing is not repeatedly and systematically extracting or re-utilising insubstantial parts of the contents of one database. Rather it amounts to taking one insubstantial part of a sequence of different, though inter-related, databases.
66. This argument can be illustrated by the following example. A database is made up by weekly additions of new data and the removal of some of its old data. On a particular week the contents may be depicted pictorially as set out below:
    
    [Diagram or picture not available]
67. The data represented by 1 + (a) is 5 weeks old, that by 2 + (b) is 4 weeks old and so on. The most recent data are 5 + (e). Assume for the purpose of illustration that (a), (b), (c), (d), and (e) are, by themselves, each so small and individually trivial that they represent insubstantial parts of the content of the database. For example in relation to a database concerned with horseracing, each represents the name of a single horse. The next week the contents of the database can be represented as follows:



[Diagram or picture not available]

68. The oldest data, 1 + (a), has been jettisoned but the most recent data, 6 + (f), has been added. Once again the piece of data, (f), is an insubstantial part of the contents of the database in this form. By the sixth week the database looks like this:

[Diagram or picture not available]

By this time, all of the data which had been in the database 6 weeks previously (as depicted in paragraph 66 above) has been removed.

69. The weekly additions and subtractions are sufficient to justify a separate term of protection for the database each week. Assume now that the unlicensed third party takes data (e) in his first week, (f) in his second week and so on up to (j). When he takes (a) he clearly is not taking anything from the next week's or any subsequent collection of data. When he takes data content (e) in the second week, again he is not taking it from any subsequent week's collection of data, nor has he taken his data from the previous week's collection. The same analysis goes all the way through to and including data content (j). Here there has been no repeated and systematic extraction re-utilization of insubstantial parts of a single database, rather there has been a taking of a single insubstantial part of each of a sequence of databases. This is not covered by art 7(3).
70. Although this is an attractive argument, I do not think it is right. The Directive has to be construed to make sense. This argument starts from the assumption that all databases have to be considered as discrete 'frozen' products, rather like separate editions of a book, each with a discrete period of protection. So, a database which is modified over time is to be broken down into a series of steps, each of which is protected by database right. However that would cause serious difficulties. If a database which is under constant revision is to be considered as a series of discrete protected works, how is one to know where one ends and the next begins? Presumably the size of the steps would be determined by the amount of effort needed to trigger the operation of art 10(3), yet it will be virtually impossible in any case either to determine where the starting point is in the creation of the series of databases or to define the size of the steps. In the case of the BHB Database, for example, are the steps to be measured in seconds, hours, weeks or months? How many of the 800,000 entries made in it each year create a new database? How many separate databases are created each year? Furthermore, were this argument correct, in most if not all cases, art 7(3) would have no application.
71. It seems to me that there is a false assumption underlying Mr Platts-Mills' argument. There is nothing in the Directive which suggests that it was not to apply to dynamic databases in just the same way as it applies to ones which are built and modified in discrete, well defined steps. Many of the most valuable databases are those which are under constant revision. If Mr Platts-Mills' argument were correct, a database which was very large indeed and which involved very large additions of data from moment to moment would have to be split into a myriad of databases, each separated from its immediate predecessor and successor by a very short time interval. Unless an unlicensed copyist took a significant amount of data all of which happened to have been entered or verified virtually simultaneously in the database, he could say that he had taken insubstantial amounts from each member of a series of closely related databases and thereby had not infringed. In fact this argument could be raised in this case in relation to, say, all of the data relating to racing on 13 March which is to be found on the William Hill website. As I have already said that this is a substantial part of the BHB Database. As I have also said, the data reproduced by William Hill includes the names of 192 horses in 20 races held at 3 racecourses. It is inconceivable that all of this material was put on the BHB Database at the same time. It was almost inevitable that each piece was entered at a different time. If Mr Platts-Mills' argument were correct, it would be for BHB to prove that these pieces of data were entered so close to one another that they should be considered to be part of a single database. To do that, it would be necessary to show that between the first and the last there had not been sufficient effort put into the creation of the BHB Database to trigger art 10(3) one or more times.
72. In my view the BHB Database is a single database which is in a constant state of refinement. It seems to have been so regarded by all the witnesses. An attempt to split it into a series of discrete databases, besides being impossible to do, would not reflect reality. Its contents change with time and without any obvious break. So too, the term of protection changes. As new data are added, so the database's term of protection is constantly being renewed. However an unlicensed third party who takes only older data from it only faces a database right which runs from the date when all of that older data was present in the database at the same time. This does not render art 10(3) meaningless. First it emphasises that the term keeps being renewed as the database is renewed. Secondly it makes clear that if someone takes an existing database and adds significantly to it, he obtains protection for the database incorporating his additions. This would be so even if the new author is not the same as the author of the original database.
73. For these reasons it appears to me that the BHB Database is to be regarded as one database in a state of constant revision. William Hill's borrowing from it from day to day come within art 7(5) as repeated and systematic extractions and re-utilizations of parts of its contents.
74. The other argument advanced by Mr Platts-Mills is based on the final part of art 7(5). The extraction or re-utilization of insubstantial parts of a database does not infringe database right unless it conflicts with normal exploitation of that database or unreasonably prejudices the legitimate interests of the maker of the database. He says that William Hill's activities do not so conflict.
75. Each side has given me examples of what it says would or would not count as a conflict with normal exploitation or an unreasonable prejudice. It is not too difficult to think of examples which would fall outside the scope of protection. For example an author of fictional articles about horseracing which are published once a week in a Sunday newspaper might decide to pick the names of the horses in his stories from the list of names of runners published in the Racing Post each week. In so doing, he would make no use of the investment made by BHB in creating or verifying its database and such use would not conflict with normal exploitation of it nor would it prejudice BHB's interests. What amounts to normal exploitation or unreasonable prejudice to legitimate interests may well be impossible to define with precision. In my view this is not a case where it is necessary to try, because the defendant's activities here clearly undermine a significant part of BHB's exploitation of its database and unreasonably prejudices its interests for the reasons advanced by Mr Prescott and set out at paragraph 62 above.
76. For these reasons, I have come to the conclusion that William Hill also breaches BHB's rights under art 7(5).
77. This only leaves one other issue, namely William Hill's proposed modifications to its website if it loses this action. These modifications were explained by Mr Michael Ellis, a senior executive of the defendant:

"If for any reason it was not possible to publish the runners' names or the race names/times, two changes could be made to the Information used on the Internet site. Firstly, instead of the time of the race being stated, the number of the race at the meeting could be identified (e.g. "race 2 at Ascot"). Secondly, instead of the horses entered into the race being identified by name, they could be identified by their number. If an injunction is granted which stops William Hill from using the Information on its Internet site, then (unless the injunction also forbids this) William Hill intends to make one or both of those modifications, but otherwise intends to continue to use the Information [as it currently does]".

78. Database rights protect the unlicensed taking and use of information. What William Hill has in mind involves the manipulation of the same information but its presentation in a different manner. This change would have no impact on the issue of extraction. Substantially the same information, essential for enabling William Hill's customers to place bets, would have to be extracted by William Hill from the RDF or an equivalent source. Infringement of BHB's database right in this respect would be unaffected. Furthermore, I do not see how the modified method of presenting substantially the same data could avoid infringement by re-utilization. If a database happened to be written in English, an unlicensed third party who displayed a substantial part of it would not avoid infringement by doing so in French, German or Chinese ideograms, nor would he avoid infringement if he translated information in denary code to its binary equivalent. As long as substantially the same information is made available on the website, the same acts of extraction and re-utilization will have taken place.

[Annex not available]