England and Wales High Court (Chancery Division) Decisions

IN THE HIGH COURT OF JUSTICE
CHANCERY DIVISION

		Royal Courts of Justice Strand, London, WC2A 2LL
		31st July 2007

INTRODUCTION

1. This is an action for infringement of copyright and breach of contract relating to two suites of the defendant's software known as GA_Money (GAMoney) and Golf Game ("GA Golf). The claimant ("Cantor") claims that the defendant's ("GA") use of these suites constitutes infringement of its copyright and breach of a written agreement which confers a limited licence to use software in which it owns the copyright. The copyright claim and the contract claim are closely related, since the central allegation is that there has been use of the software beyond the scope of the licence, thereby giving rise to infringement. The software in question is compendiously known as the "Cantor Software" and it is alleged that specific parts of it have been wrongfully used by GA.
2. It is convenient, before turning to the issues, to set out aspects of the history.

HISTORY AND DEVELOPMENT OF THE CANTOR SOFTWARE

3. The history of the development of the Cantor Software is set out in the statement of Ms Stratford-Martin, in an account which is not challenged.
4. She explains that BGG International ("BGC") is the parent company of Cantor and it is a substantial undertaking involved in financial services and related activities. In or around November 2000, Mr Kevin O'Neal and Mr David McDowell, approached BGC with an idea for developing and commercialising on-line games. They wanted to create a series of skill based games which would be available on the Internet. The precise detail of their proposal does not matter for present purposes but BGC believed that the proposal had commercial promise. It therefore arranged to employ Mr O'Neal and Mr McDowell and for them to run the project as a business, to be majority owned by Cantor.
5. For purpose of the venture, BGC and the individuals concerned incorporated a company on 12 December 2000, known as "GameAccount Limited" as the corporate vehicle for the project. This company has now changed its name to "Cantor Gaming Limited", the Claimant.
6. Mr O'Neal and Mr McDowell entered into employment contracts with Cantor on 13 March 2001 under which they were respectively appointed as joint Chief Executive Officer and Chief Financial Officer. Three consultants were also engaged.

The development of and title to the copyright in the Cantor Software

7. Mr O'Neal and Mr McDowell were behind the initial ideas for the software and were involved in the plans for its development and exploitation. However, they are not themselves software engineers. The software development was carried out by software engineers who were either employed by or contracted specifically by Cantor or were employed by another BGC group company and assigned to assist with the project. The four individuals particularly responsible for the development of the Cantor Software were Clive Hetherington, Sam Lawrence, Fergus Leane and David McNally, with Kevin O'Neal with David McDowell acting in a more supervisory capacity.
8. There is no dispute that, either as a result of express terms in the contracts of the individuals or by operation of law coupled with appropriate assignments, Cantor came to own the copyright in the Cantor Software so developed. At a certain point during the development of the software a number of the programmers were transferred to work for another company in the BGC group, known as "eSpeed" and appropriate contractual terms ensured that copyright in works developed at eSpeed also vested in Cantor. Cantor also owns the copyright in this software and some of the graphics relating to the Golf Game, which were separately developed by Vaguerant Limited (trading as PixelBrothers) from whom an assignment of copyright has been obtained.

The software in issue in outline

9. The Cantor Software, taken as a whole, comprises several parts. This case is concerned with two suites which form part of the Cantor Software known as, GA_Money ("GAMoney") and Cantor Golf ("Cantor Golf).

GAMoney

10. GAMoney is a complete database. It was been described by Mr Lawrence of GA as a "core database". The core function of GAMoney was to enable applications to connect to it to update and retrieve information about customers' accounts and the games which they had played. GAMoney closely interacted with other aspects of the Cantor Software. There is no dispute and Mr Lawrence accepted in cross-examination that, taken as a whole, it is a very substantial piece of software containing thousands of lines of code. It will be necessary to consider aspects of its structure and function in greater detail below. For present purposes, it is not in dispute that, taken as a whole, GAMoney is a work in which copyright subsists and that skill and labour was exercised in creating parts of it such as the stored procedures and tables considered below.

Cantor Golf

11. Cantor Golf is a suite of software which enables a simulated golf game to be played on-line. It has a number of different elements, including a part which calculates the trajectory of a ball, relative to the golf course, that has been notionally "hit" by a player and another part that generates particular graphics showing the golf course. In addition, it enables players to select a wager and it keeps track of play, declaring a winner at the end of the game. Software of this kind operates using a combination of "client side" code (i.e. software stored and running on a user's computer) and "server side" code (i.e. software stored and running on the server of the undertaking offering the game in question).
12. The issues on liability that arise as regards Cantor Golf are much more straightforward than those in relation to GAMoney and the evidence deals with Cantor Golf more briefly.
13. First, it has been accepted in correspondence that there has been use of the Cantor Software for GA's GAGolf game in a manner not authorized by Cantor and that such attracts liability both under the contract and as a matter of copyright infringement. Details of how that was accepted are outlined below.
14. Second, Mr Lawrence of GA said in his initial witness statement that, on 3^rd August 2004, a new version of the golf game (GAGolf) was released on the GA2 website. He says (para. 51):

"Some of the graphics and some of the game code from the Cantor Golf game were incorporated into GA Golf. "

15. GA now accepts that the parts used constitute a sufficiently substantial part to attract liability. It is therefore unnecessary to deal further in this judgment with Cantor Golf or the similarities that it bears to GA Golf further. It will be necessary to return to it in considering the question of what relief is appropriate.

The developers and Cantor part company

16. Returning to the chronology, Ms Stratford-Martin's evidence explains that, in mid 2002, the relationship between Cantor and Kevin O'Neal and David McDowell broke down as a result of differences over how the project should develop and (it is said) because the development was over budget. At that point, the Cantor Software had not been launched on-line and there was only one contract in place, which was for Cantor to provide and run an on-line game as a promotion in The Sun for News International Limited.
17. The parties therefore agreed that Mr O'Neal and Mr McDowell would leave Cantor but would be given a licence in relation to the Cantor Software so that they could (i) fulfil the contract with News International Limited and (ii) further exploit the Cantor Software through the vehicle of a new company which would be run under the name "GameAccount". That vehicle is the defendant, GA.
18. At the end of September 2002, a Deed of Settlement was entered into between Cantor Gaming and a company then called "Fro-Zen" (which, renamed, is the defendant) under which Cantor Fitzgerald GameAccount Limited Holdings LLC bought all of the shares of Cantor from the shareholders, including those shares belonging to Kevin O'Neal and David McDowell, and by which Kevin O'Neal and David McDowell would enter into Deeds of Termination of their employment.
19. At the same time, a Collaboration and Transfer Agreement ("the Agreement") was entered into between Cantor and Fro-Zen Limited whereby Cantor granted a limited licence of the Cantor Software to Fro-Zen. A trade mark assignment was also entered into between Cantor and Fro-Zen whereby Cantor assigned the trade mark "GAMEACCOUNT" and all other IP in the name "GAMEACCOUNT" to Fro-Zen resulting in the change of name. A copy of the Cantor Software was provided to GA on a CD ROM at completion of the Agreement.
20. One of the issues in this case centres on the Agreement. Under Clause 2 of the Agreement, GA was granted a perpetual, irrevocable, royalty-free, non-exclusive licence which (it is accepted by the parties) was limited in its scope in a number of ways, to use the Cantor Software. The licence was, in particular, limited by clause 10.3.6, which I consider further below.

ISSUES

21. There are two main issues which require determination.

(a) Whether GA has infringed Cantor's copyright and/or acted in breach of contract in its use of GAMoney. This boils down to a question of whether there has been use of the GAMoney in a manner contrary to clause 10.3.6 of the Agreement.

(b) Whether an injunction and delivery up should be granted, in the light of the undertakings offered and what has happened since.

22. Neither issue is wholly straightforward. As to the issue regarding GAMoney, that turns largely on construction of the Agreement and an assessment of the nature of the use that GA has made of GAMoney. That, in turn, requires consideration of how and by reference to what standard, the significance of the admitted use is to be determined. The parties' approaches to this issue diverge considerably although there is consensus that the practical importance as regards GAMoney is limited. Use of GAMoney ceased in 2006 and, while maintaining that there was breach of contract and infringement, there is only faint suggestion by Cantor that the use was of any commercial significance even before that date.
23. As to the second issue, there is a dispute as to whether, in all the circumstances, an injunction is an appropriate remedy, having regard, inter alia, to the undertakings offered before the proceedings commenced and the changes that GA has made as a result of the pre-action correspondence. This issue requires the court to consider the principles upon which the court will grant injunctive relief in cases where it is alleged that there is no threat of future infringement. Although the parties have adopted rather different approaches, the practical significance of the point is also open to question in this case. Public use of the GA Golf game has stopped and, save for one matter to which I will come, had stopped shortly after the initial letter before action. As noted, GAMoney was removed in early 2006.

HEARINGS

24. There were two hearings in this case. The main trial took place on 9th -10^th May 2007. Cantor then applied to re-open the trial and there was a further hearing on 23rd May 2007, at which further evidence was adduced from both sides, relating, almost exclusively, to GA Golf.

WITNESSES

25. At the main hearing, four witnesses gave evidence and were cross-examined.

Claimant's witnesses

Ms Stratford-Martin

26. Ms Stratford-Martin is a solicitor and a senior legal counsel of BGC International, Cantor's parent company. Her evidence dealt primarily with the history of the dispute and established certain formal matters concerning title to the software. Her cross-examination mainly concerned the development of Cantor's case on the correspondence and the pleadings. In my judgment, she gave fair answers to the points put to her. In particular, it was suggested to her on behalf of GA that she had provided a selective picture of GA's approach to the litigation. While it is true that her evidence does not refer to every letter, I do not find that her account was inaccurate. She was a good witness.

Mr Carlisle

27. Mr Carlisle is a senior lecturer and the course leader for the BSc honours programme in Computer Games Design at the University of Bolton. He holds a degree in Computing (networks and distributed systems) from Leeds Metropolitan University and has extensive experience in games programming spanning many years, working inter alia on aspects of the Worms games. He provided two formal expert reports. He was cross-examined on these and on an earlier interim report provided to GA in November 2006. He too gave his evidence fairly. It was put to him that, had he checked, other versions of the software in question he would have realised that certain conclusions which he had reached in his interim report were ones to which he could not have come, points which he accepted. Overall, his evidence was of considerable assistance and I make no criticism of it.

Defendant's witnesses

Mr Lawrence

28. GA's primary factual evidence was from Mr Samuel Lawrence, GA's chief technical officer. Mr Lawrence has managed the growth of the company's technical department and is responsible for the management of all aspects of GA's technology including every aspect of the software development process from conception through to release.
29. Mr Turner for Cantor criticised Mr Lawrence's evidence, in particular in that Mr Lawrence had not been as forthcoming as he might have been on the GA's purported justification in 2004 for such use as they had made of aspects of Cantor Golf for GA Golf. It might fairly be said that Mr Lawrence was not particularly comfortable in the witness box when asked about the golf game and GA's justification for the use of parts of Cantor Golf, reflected in his answers to questions on this topic during cross-examination. However, the focus of the technical dispute at trial was not GAGolf (where liability was admitted) but GAMoney. Admission of infringement of another's copyright may lead to some discomfort in answering questions about how that occurred. On GAMoney, Mr Lawrence's evidence was comprehensive in his witness statement and in cross-examination, save for a slight hesitation before accepting in cross-examination that removal of GAMoney had been motivated by a concern about the litigation.
30. None of this is sufficient, in my judgment, to lead me to reject Mr Lawrence's evidence and I am unable to accept Mr Turner's submission that Mr Lawrence did not do his best to assist the court in this matter. I should add that nothing at the further hearing at which he also gave evidence caused me to change my view. Mr Lawrence's evidence in paragraph 51 of his witness statement was challenged on the basis that events following the main hearing showed that the statement that GA Golf was removed was untrue. As I explain below, it turned out that GA Golf was, to Mr Lawrence's surprise, still accessible to some degree, even at the date of the trial. However, he did not know that it was and he gave an explanation, which in my judgment was cogent, for why a software developer would not have thought that it was accessible in that way. Accordingly, nothing at the further hearing leads me to conclude that Mr Lawrence was not trying to assist the court.

Mr Lee Murray

31. Mr Murray was, until March 2007, employed by GA as a Senior Website Designer. He was primarily responsible for organising the removal of GA Golf in February 2006. He explains what he did to render the software inaccessible. He was cross-examined (at the later hearing) and I accept his evidence.

Mr Du Vergier

32. Mr David Du Vergier was GA's expert. He is employed by LogicaCMG as a technical architect specialising in enterprise systems. His c.v. indicates that he has extensive experience in the development of a wide range of systems in different areas although perhaps less experience in games as Mr Carlisle. Nonetheless, he was, in my judgment, well qualified to give evidence on the matters in issue in this case. No criticism was made or could be made of his evidence. It was cogent, clear and helpful, particularly as regards the general concepts of the operation of relational databases and the operation of GAMoney.

Other witnesses

33. A witness statement from Mr David McDowell, a non-executive director of GA, was placed before the court. He was not available for cross-examination. His statement explained the steps taken to remove the Golf game from the GA web-site. His understanding was that, after removing or redirecting the Golf game page, it would no longer be possible for anyone to play GA Golf. It supported the evidence of Mr Murray and Mr Lawrence. Further, at the hearing on 23rd May 2007, Cantor also adduced evidence from Mr Carl Rohsler, a solicitor at Hammonds, concerning his discovery on 17^th May 2007 that it was still at that date possible to access GA Golf by means of an internet search. The factual accuracy of his statement is not in doubt.

DATABASE STRUCTURE AND THE MAKING OF COPIES OF THE SOFTWARE

Database structure

34. In order to understand the context in which the issues arise, it is necessary to understand something about relational databases and the software for them. Mr Du Vergier explained this in section II of his report, in terms which are not disputed and which I therefore set out in full.

"6. GAII_Build and GAMoney are both what are known as relational databases. A relational database is a database that maintains a set of separate, related entities (known as "tables") and combines data elements from these entities for queries and reports when required. The distinction between relational and other types of database is not relevant for present purposes. I therefore use the term "database " in this report to mean relational database.

7. In essence, a database consists of a series of spreadsheet-type tables consisting of rows and columns, each table holding a different kind of information and related in some defined way to some or all of the other tables. Thus a simple credit score database might consist of a table containing the names of a number of people, a table detailing the credit cards held by each of those people and a table detailing the payment histories for each of those cards. By using the relationships between these tables, this database can be queried to return any combination of data, e.g. the payment history of a particular person on a particular card; the number of cards held by that person; the number of times that person has made late payments on any card and so on.

8. To allow data to be located efficiently, a database will typically also contain a number of 'indexes'. An index relates to a particular table, and allows rapid navigation to the required rows in that table. To continue the previous example, a table containing names of people might also have an index on the surname column; in this way, all persons named 'Smith' can be located directly, without having to search the entire table. An index typically contains a small subset of the data contained in the corresponding table and has a correspondingly smaller disk and memory footprint in comparison to the full-blown table.

9. In principle, then, a database comprises:

The database 'schema'; this is the definitions of each table in the database, the interrelationships between all of those tables, and their indexes (I shall refer to tables of this kind hereafter as "application tables"). GAMoney contains 105 such application tables.

The database 'stored procedures'. Stored procedures are sections of code dedicated to performing an individual function of the database. Thus a stored procedure might be used for the kind of queries outlined above, or for the entering and updating of information within application tables or for practically any other operation the creator/user of the database can conceive of. Stored procedures are stored as part of the database, producing efficiency benefits and providing a clearly defined interface to the data.

The data itself; this is the actual data stored within the application tables and indexes (I shall refer to data of this kind hereafter as "application data ").

10. The combination of schema, stored procedures and data is what is commonly meant by the expression "database". Thus GAII_Build and GAMoney each have their own unique schema, stored procedures and application data. Each is therefore a unique database.

Physical Implementation of a Database

11. Databases are generally created and operated by means of a sophisticated commercial software product called a database "engine ". Examples are Oracle and Microsoft's SQL Server. GAll_Build and GAMoney are both SQL Server databases. A typical SQL Server setup will involve a dedicated computer (a "database server ") running the SQL Server engine, itself running any number of different SQL Server databases.

12. To understand the extent to which copies of GAMoney were made in the memory of GameAccount's computers post August 2004, it is necessary to understand how the SQL Server database engine drives a database. To do this, it is helpful to go briefly through the process by which a new database such as GAII_Biuild is created. Firstly, the SQL Server engine creates a skeleton database, including a number of so-called "system " tables. It is important not to confuse system tables with application tables. System tables are fixed, generic tables that the SQL Server engine uses to build and maintain its own internal picture of each database it has to run. Subsequently, as the creator of the database adds new application tables, indexes and stored procedures according to the purpose of the database under construction, definitions of these items are stored in that database's system tables in a form that is meaningful to the SQL Server engine. So for example the system table "sysobjects" stores, inter alia, the names of all of the application tables that comprise that database.

13. Once a SQL Server database has been constructed, the SQL Server engine uses the information contained in its system tables to recover that database's stored procedures and to navigate its schema.

System Tables and Metadata

14. The information in the system tables is called metadata, because it is data describing data. This term is to be contrasted with application data, which is the user-accessible data stored in the application tables (in the case of GAMoney and GAII_Build, details of players, wagers etc). It is the concept and use of metadata that allows a single generic database engine such as SQL Server to drive any conceivable database. In this sense we can think of the database engine as a general-purpose data storage and retrieval device, and the metadata as a program that configures it to understand each kind of application data, e.g. (as in the case of GAII_Build and GAMoney) details of players and wagers.

15. Database engines are extremely sophisticated pieces of technology, analogous in many ways to a computer operating system. In both cases much goes on beneath the surface hidden from the user's view. In both instances a user, even a sophisticated user such as an application developer or database administrator, has a limited amount of control over the way the database engine or operating system actually performs its functions. Thus, to some extent a database engine such as SQL Server is a piece of "black box" technology, i.e. it performs its function but the user does not always know and cannot always control the detailed manner in which it does so... "

Copying of software

35. It is also necessary, in order to understand Cantor's case, to understand something about how copies of software (or parts of it) may be made during the course of development, maintenance and running of the program. These are explained most fully by Mr Carlisle and are as follows.

Copying during development of the software

36. The first way in which copies may be made is during development of the software. Mr Carlisle explained that, when developing software, it can be important to ensure that changes to the system are managed without causing unexpected errors or crashes in real time. As he says, this requires a group of developers to co-operate, and to ensure that when a change is made to one part of a system, it is tested first in an environment where it can do no harm, before being generally released into the system. Controlling the state of the system precisely can be a challenge. Often, a team of programmers will use a form of version control software which places the code in a "library" and then monitors and controls the "checking in" and "checking out" of versions by programmers. This means that two programmers cannot work on the same file at the same time, and also means that if an unexpected error occurs, it is possible to recreate the software at the last date on which it was working correctly.
37. Another measure which is normal practice is for there to be two servers (or two sets of servers in a multi-server deployment). The first containing the "live" system, and the second acts as a "development" server, in which code can be modified and tested outside the live environment. Once a change to the system has been made and successfully tested, then it will be "uploaded" onto the live server and dropped in as a replacement for the previous version. That is the practice employed by GA, who throughout the relevant period maintained a development server. It is inevitable that each "uploading" of the software leads to the creation of a new copy of the software on the server.

Copying during running of the software (program execution)

38. The second is during the running of the software (during program execution). Mr Carlisle also explained that a further way in which computers create copies of the software operating on them (or at least parts of it) is when a program is being executed. Software in its "executable" form usually exists in the form of one or more files stored on the hard drive of the computer.
39. As he says, when a program is executed by the computer it is copied from the hard drive into "dynamic" memory, creating a new copy of the files in the live memory of the computer. During execution of a program, transient copies of the instructions contained in the program (but expressed in "machine code" rather than a high level programming language) are constantly being copied and recopied as the computer performs logical operations.
40. In a constantly live development such as the Cantor or GA system, one would expect the relevant software to be loaded into the Random Access Memory of the computer immediately at the point when the computer is turned on. It will remain in that memory during the course of the operation of the program, and each interaction within the program logic will cause different instructions to be copied and operated upon. At various points in the operation of the program, a subroutine will be "called" by one part of the program to perform a specific function. Such a call occurs as a result of an instruction in the code.
41. Mr Carlisle explains that this system of logical "calls" and "returns" lies at the heart of all programming. It is what allows the program to distribute its logical functions in an ordered fashion. It also means that programs are not huge monolithic pieces of code like a long narrative novel with constant repetitions of the same functions at every occasion, but can be shorter and logically create a "division of labour", with each subroutine doing a specific job. The way that calls operate in the internal operation of a computer is that the system looks up the address of the named function in memory. It then "jumps" to that address and executes the sequence of instructions it encounters. These instructions are generally copied into memory at the initial running of the program and stay in memory for as long as the program is running. No matter whether the instructions actually perform a useful task, the fact remains that all the instructions capable of being addressed must be copied into the computer's memory in order for the system to operate.
42. Copying of this kind is under the control of the person operating the software. It is he or she who will initiate running of the software and such a person will generally know that software of this kind is running on the system.

Copying during maintenance of the software (back up copies)

43. The third way is where copying of software also occurs when backup copies are made. This is straightforward and takes place whenever a copy of the whole or part of the software takes place for the purpose of backing up software, usually while it is in development.

Copying under control of the software itself or its environment

44. There is a still further way in which copies of software or parts of it may be made in a manner which may be unknown to (and not easily predictable by) the person running the program. Much modern software uses sophisticated techniques of storing in memory either parts of a program or data in a local memory or "cache". The purpose of this is generally to speed up operation and to avoid the need to call that software or data up from hard storage. Software may copy parts of a program that may not be immediately needed but which the program considers are likely to be needed in the near future. One form of this is that certain software will load into Random Access Memory more subroutines than it needs for a particular operation, copying (for example) related programs or data which it considers it may need.
45. In the most general sense, this type of copying is under the control of the person running the software since he or she has initiated the running of the software. But in a more realistic sense, it is not under the control of the user at all. He or she is unlikely to know that it was happening and, as in this case, even the most qualified experts may be unable to determine, without disproportionate effort, precisely what parts of any suite of software have actually been copied in this way. To add to the difficulty it is possible (at least in theory) that what and how much is copied is not fixed but depends on the software's analysis of the operations that have been performed, which may differ on each occasion that the software is run.

Types of use of software

46. In addition to types of copying of database software, it is possible to divide use of database software into a number of different categories. First, there is full scale running of the whole of a database Second, there may be use of specific procedures from a suite of database software. Third, there may be no use of the stored procedures but merely the accessing of parts of a database (such as the application tables) for some other purpose.

THE AGREEMENT - INTERPRETATION

The licence and its limits

47. Clause 2 of the Agreement confers the licence, the scope of which is in issue in this case. It is in the following terms:

"2 Licence

2.1 In consideration for [GA] entering into the Deed of Settlement and undertaking to perform its obligations under this Agreement (including without limitation the payment of the Licence Fee stipulated in clause 2.2 below), [Cantor] hereby grants subject to the terms of this Agreement to [GA], a perpetual, irrevocable, royalty-free nonexclusive licence to:

2.1.1 to use the Licensed Materials in order (and to the extent necessary) to fulfil its obligations under any contract entered into pursuant to clause 4.1 below;

2.1.2 to develop or modify the Licensed Materials in order to devise and develop New Games which may be jointly exploited by the parties pursuant to clause 4.2 below;

2.1.3 subject to clause 2.3 below, to promote the Golf Game, Backgammon Game and any other New Games and to enter into contracts with Customers for the provision of any of the foregoing to End-Users of such Customers;

Subject always to the terms and conditions of this Agreement. "

48. It is noteworthy that this clause does not in terms confer a licence as such under the GameAccount IPR, as defined.

Clause 10.3.6.

49. As noted above, clause 10.3.6 of the Agreement is central to this dispute. It provides as follows:

"10.3 Without prejudice to the generality of clause 10.2.1, the parties agree that the occurrence of any of the following events shall constitute a "material breach" of this Agreement....

10.3.6 A third party Bookmaker (or any member of its Group) directly or indirectly finances, invests or acquires any shareholding in [GA]; or [GA] enters into any form of collaboration or agreement (whether oral or in writing) with third party Bookmaker (or any member of its Group) or any company carrying on an exchange business (or any member of its Group) in which any Intellectual Property Rights of [Cantor] are used or proposed to be used; "

Definitions

50. Clause 1.1 includes the following definitions.
51. "Intellectual Property Rights", defined as "...copyrights (including rights in computer software), patents, trade marks, service marks, business names, design rights, database rights, semi-conductor topography rights, rights in undisclosed or confidential information (such as know how, trade secrets and inventions) whether patentable or not)) and all other intellectual property or similar proprietary rights of whatever nature (whether registered or not and including applications to register or rights for registration) which may now or in the future subsist anywhere in the universe. "
52. "Licensed Materials", defined as "the items listed in Schedule 1". These include the suite of software applications for the "Golf Game" and the applications which support and monitor the running of the backgammon, golf and other on-line game[s].
53. "GameAccount IPR", defined as "all the Intellectual Property Rights owned by [Cantor] and subsisting in the Licensed Materials".

Interpretation

54. Before turning to the rival contentions, it is necessary to record that the parties are agreed that clause 10.3.6 operates so as to limit the scope of the licence to use the software granted pursuant to clause 2. In my judgment this is correct, indeed inevitable, despite its presence in clause 10 which, overall, is concerned with term and termination.
55. Cantor contends that the reference to "use" of the Intellectual Property Rights in that clause refers to use in the sense of "infringe". That is to say, if an act of the defendant would, but for the licence, constitute (for example) an infringement of copyright in the software, that constitutes "use". Cantor also contends that the clause is implicitly limited to the works licensed under the agreement and referred to in Schedule 1 ("the Licensed Materials").
56. GA, on the other hand, contends that "use" cannot be equated with "infringe" (in a copyright or other intellectual property right sense) in that clause. Something further is required. GA points to the fact that clause 10.3.6 does not use the term "infringe" (which it could easily have done had that been meant) and that to construe the clause in that way would involve re-writing it. GA contends that the term "use" must be understood by reference to the intended purpose of the software in question. GA submits that the relevant assessment is in terms of intended functionality. So, if the software is not deployed or invoked (to use neutral terms) for the purpose for which the software was designed or intended, that does not constitute "use" within the meaning of the clause. GA also submits that this accords with the approach taken by Mr Carlisle in an interim report prepared for the case dated 6^th November 2006.
57. GA further submits that any other construction would result in draconian consequences for the following reason. It is well known that verbatim reproduction of even a comparatively small part (quantitatively) of a copyright work may infringe. Liability for copying is strict and it is possible for there to be copying of software without a user of the computer system invariably being aware of it, because certain programs will (for example) engage in automatic cacheing. The parties cannot have contemplated that slight and inadvertent copying of this kind would give Cantor the right to terminate the licence altogether. So GA prays in aid the fact that clause 10.3.6 is within the termination provisions as a reason to limit its breadth. GA submits that it is only by reading "use" as involving more than "infringe" that proper commercial sense can be made of the clause its context. These are serious arguments and they were advanced very attractively by Mr Acland for GA.

Analysis

58. In analysing these submissions, it is important to bear in mind that the word "use" is capable of meaning different things, according to context. It is necessary to have regard to its particular context in determining what "use" means specifically in clause 10.3.6. The following matters are, in my judgment, relevant to what "use" means in that context.
59. First, it is not disputed that the Agreement overall (and in particular clause 2) confers a licence under specified intellectual property rights. Clause 10.3.6, however, refers to the "use" of any "Intellectual Property Rights" (as opposed to either use of the "Licensed Materials" or use of the "GameAccount IPR"). These are both of some significance. Although not terminology with which a purist might be comfortable, it is not improper (or uncommon) to refer to infringement of an intellectual property right as its "use". Certain intellectual property rights, such as confidential information, are generally referred to as being used, even in more specialized legal contexts. So there is a fundamental reason for considering that "use" in this context, governing as it does "Intellectual Property Rights" is intended to denote the concept of doing an act which would otherwise constitute infringement of one or more of those rights. To that extent I accept Cantor's argument.
60. Second, clause 10.3.6 is a clause concerned to limit the activities of GA and provide a basis for bringing the Agreement to an end altogether, if GA acts contrary to its provisions. However, the licence does not automatically terminate if there has been use of the kind prohibited by that clause. Clause 10 as a whole provides that, in those circumstances, there is a right to terminate in the event that a material breach has gone unremedied or is not capable of remedy. Accordingly, the consequences of use within the terms of clause 10.3.6 are not inevitably as draconian as GA submits.
61. Third, it is not, however, in my judgment correct to treat clause 10.3.6 as limited to situations where there is use of "Intellectual Property Rights" in the "Licensed Materials". The clause does not so provide and such is not necessarily to be implied. To do so would arguably be to redefine implicitly the term "Intellectual Property Rights" of [Cantor] in a manner inconsistent with the specific definition of "GameAccount IPR". Clause 10.3.6 is broader: it refers to the Intellectual Property Rights of Game Account (i.e. Cantor). Although the definition could rationally have been chosen to be narrow, a broader definition also makes commercial sense. The parties could have agreed (and in my judgment did agree) that the Agreement would be terminable if any of Cantor's intellectual property rights were used in the prohibited manner, regardless of whether they were the specific rights licensed under the Agreement. I am unable to accept that aspect of Cantor's argument.
62. Fourth, although it is (just) grammatically possible to read clause 10.3.6 as confined to a situation where breach is determined once and for all, upon GA "entering into" the specified collaborations or agreements with a third party Bookmaker, it would not make commercial sense so to interpret it. The clause contemplates that there will be material breach, inter alia, when the rights in question "are used". That in turn contemplates that the analysis may take place at the time of use of the rights, not merely at the time of entering into the agreement. In order to fall within the clause there must not only be use of the Intellectual Property Rights but use (or proposed use) in a collaboration or agreement with a third party Bookmaker.
63. Finally, there is a real difficulty with GA's attempt to read in a further requirement, above a concept akin to "infringe" into the meaning of use in clause 10.3.6. GA contend that the test is one of functional utility or redundancy. But that begs the questions: "what functions and what utility?" The approach leaves open the question as to whether it is possible to regard as functionally redundant a use which formed part of the working system, albeit a use that could have been removed and did not involve the full functionality of that part.
64. I should add that I do not believe that the parties in drafting the Agreement had the specific workings of a database in mind or even the fact that sometimes there was residual software on a system specifically in mind in drafting the agreement or that the terminology used was intended to reflect a particular understanding of what use was or was not to be limited by the agreement.

Conclusion on interpretation

65. For all these reasons, of the two alternative interpretations, Cantor's proposed interpretation is to be preferred. I therefore propose to apply it in what follows. I also consider below the position, if I am wrong on construction and at least some additional functional element is required. Before doing so, it is necessary to make some preliminary observations about Cantor's case on use of GAMoney.

GA'S DEVELOPMENT

The position in 2003 and GA's understanding of what would be required to ensure compliance with the limited licence

66. It is convenient to pick up the chronology and explain how GAMoney came to be used by GA.
67. By 2003, GA knew that it would have to make changes to the software in order to avoid infringing Cantor's rights when it was proposing to enter into arrangements with UKBetting. Mr Lawrence accepted in cross-examination that he always knew that the "database part needed to be migrated away from" in order to comply with the contractual requirements. Mr Lawrence put it as follows in his witness statement:

"15. From early 2003 onwards, GameAccount had been developing a new database called GAII_Build to replace GAMoney. There were two primary reasons for undertaking this project. First, GAMoney fell under the Collaboration and Transfer Agreement with the Claimant. Second, we needed a different design of database to support a new line of business the company was pursuing at the time, namely single player tournaments.

16. The main development of GAII_Build was complete by August 2003. However the new database could not replace GAMoney by simple substitution. This is because the WGA website had been built to communicate with GAMoney. GAII_Build was a different design of database and an entirely new website would ultimately need to be built (called "GA2") before its advantages could be realised. The process of replacing GAMoney with GAII_Build would include five steps as follows:

(1) Data in GAMoney (usernames, account balances etc.) would be transferred over to GAII_Build.

(2) The two databases would then operate together with the WGA website as shown below:



During this phase, GAII_Build would become the definitive database for player account balances. Transaction data still had to be sent first to GAMoney because the WGA website was bound to this database and could not communicate directly with GAII_Build. In other words, the only way for the WGA web site to communicate with GAII_Build database was through GAMoney.

(3) Any new games would be introduced using the new GA2 website, bound to the GAII_Build database. The WGA website would continue to be used for the existing games and customer account pages (e.g. registering a new account or making a cash deposit).

(4) All existing games, with the exception of Cantor Golf and customer account pages would be migrated to the GA2 website. The Cantor Golf game would continue to be served from the WGA website and hence GAMoney would continue to be linked with GAII_Build.

(5) The Cantor Golf game would be replaced with a new version able to work using the stored procedures in the GAII_Build database. The WGA website and GAMoney database would then be entirely redundant. "

68. Development continued as further explained in his evidence and GA made arrangements with bookmakers in 2004, the first being on 23^rd February of that year. The GA system then offered to UKBetting was described by Mr Lawrence as follows:

"41. The following diagram represents the GameAccount system as offered to UKBetting:



The GA2 website would display the UKBetting branded site discussed in paragraph 35. The links to Cantor Golf have been removed to show that players did not have access to this game or the Clubhouse. The link from GAll_Build to GAMoney has been retained in the diagram, to show that player information relating to UKBetting-sourced players was transferred from GAII_Build to GAMoney by operation of the MG_CB_Player_New and MG_CB_Player_Login stored procedures (see SL-1). As discussed in paragraph 29, these procedures copied UKBetting player data to GAMoney because they were called upon for every registration and login made through the GA2 website. However individual player data copied to GAMoney would only be used if the player then played the Cantor Golf game. Since UKBetting players could not play Cantor Golf, their player data in GAMoney was never referenced after it had been copied.

42. Although not available to UKBetting customers, the Cantor Golf game and Clubhouse remained accessible to the general public entering through www.gameaccount.com. As such, the Cantor Golf game, Clubhouse and GAMoney database remained part of the GameAccount live suite of software until I August 2004 though none was used to provide services to UKBetting-sourced players.

Removal of Cantor Golf and redundancy of GAMoney

43. On 1 August 2004, the Cantor Golf game was removed from the WGA website. The Cantor Clubhouse was also removed on 1 August 2004 together with the WGA web pages that were embedded into the Clubhouse. With the removal of the Clubhouse and WGA web pages, there was no longer any webpage being used that was bound to the GAMoney database.

44. At this point, the GAMoney database could have been removed entirely and the task itself would have taken no more than a few hours. Apart from the GAII_Build database, no other processes connected to the GAMoney database since the Cantor Golf game and WGA website had both been removed. None of the procedures in GAII_Build referencing GAMoney and called by GAMoney would therefore have been used and so could simply have been deleted. Only three types of stored procedure in GAII_Build referencing GAMoney remained active (albeit without any purpose):

(1) MG_CB_Player_Login and MG_CB_Player_New for duplicating customer data (as explained in paragraph 29). These procedures were now without purpose since no other process would access the data after it had been copied to GAMoney.

(2) The eight procedures in SL-1 with the prefix Player_ for displaying game history via the GA2 website (one of these procedures is discussed in paragraph 68 below). No game data was stored in GAMoney so no GAMoney game data would be returned by any of these procedures.

(3) MT_Account_SimpleEvent used by the management tool for manually crediting accounts. The reason this procedure updated the GAMoney player balance was to enable wagered games of Cantor Golf to be played. As explained above, Cantor Golf was removed on I August 2004.

Each of these procedures could easily have been removed or altered to leave GAII_Build without any reference to GAMoney.

45. The following diagram represents the GameAccount system after removing Cantor Golf:



69. The latter diagram shows that there is still an arrow from GAII_Build to GAMoney. Although GAII_Build was not actively invoking the stored procedures in GAMoney, it was actively using tables in GAMoney. Mr Carlisle's evidence was

that in the period after August 2004, transient copies of GAMoney continued to be made in the memory of the computer. Mr Du Vergier's evidence was not materially different, although he was unable to say with precision which parts of GAMoney would have been loaded.

70. It is instructive to contrast the diagram in Mr Lawrence's paragraph 41 with the position following the complete deletion of all references to GAMoney, after the letter before action in this case. Mr Lawrence shows this in paragraph 48 of his statement as follows:

"48. The following diagram represents the set up after deletion of all references to GAMoney, while the GAMoney database sits idle on the live database servers:



It can be seen that all operational connections between the system offered to bookmakers and GAMoney have there been removed.

Circumstances leading to Cantor's suspicion of copying of the golf game

71. In July 2005, Cantor discovered that GA was providing an on-line backgammon service to Paddypower, a third party bookmaker, through the medium of its website (www.paddypower.com). Ms Stratford-Martin wrote to GA to remind them of the restrictions contained in the Agreement, and to ask them for their confirmation that no use was being made of the Cantor Software in collaboration with Bookmakers.
72. GA replied that the on-line backgammon game was new software which had been written from scratch and did not make use of the Cantor Software. Nothing was said about any other software. GA also confirmed that the software for their Backgammon game had been written in a new language. In the light of that correspondence, to which I shall return in connection with the injunction issue, Cantor decided not to take the matter further.

Cantor's discovery of the GA's use of GA Golf

73. In February 2006, Cantor became aware that an on-line golf game was also being offered by GA on the Paddypower website. Similarities were noticed with the Cantor Software and a letter before action was written on 16^th February 2006 seeking, inter alia, that GA immediately withdraw the offer and advertisement of the on-line golf game from the web-site and procure the removal of the same from the website of Paddypower.com and any other third party bookmaker websites and provide undertakings in the form attached to the letter.
74. GA's response, by its solicitors, on 21^st February 2006 was, inter alia, that the Golf Game was GA's own original work and that if and to the extent that any elements of the software originally provided by Cantor survived in the version in use (which was not admitted) such elements were not a substantial part of the original software. GA however arranged for all copies of GA Golf on the web to be taken down pending the outcome of further investigations.

Undertakings offered

75. During further inter-solicitor correspondence, it was accepted that the Agreement was terminated. A full response was provided on 24th March 2006 and the letter from GA's solicitors, of that date, accepted that a substantial proportion of the Cantor Golf game graphics and original internal code were retained for GA Golf. Accordingly, GA made a number of admissions relating to the GA Golf game and offered extensive undertakings including an undertaking immediately to cease and desist from "using or advertising or offering to customers howsoever the on-line golf game software" and within 10 days to "deliver up to Cantor...all copies of the Software [as defined] in its possession or control" and verify such delivery up under oath. GA also offered to submit to an inquiry as to damages and/or an account of profits for breach of contract and infringement of copyright in relation to the admitted acts of infringement and pay a contribution to legal costs of £10,000. The undertaking offered concluded with the following:

"It is further acknowledged and agreed that in the event of a breach of any of the above undertakings (1) legal action may be taken in relation to such breach without further notice and (2) Cantor will be entitled to (amongst other things) injunctive relief to restrain further breach. "

76. Those undertakings did not satisfy Cantor and these proceedings were commenced on 4^th April 2006. It is not necessary for present purposes to relate their procedural history.

USE OF GA MONEY AFTER 2004

The significance of the use of GAMoney

77. As noted above, the central question for this part of the case is whether the use of GAMoney, after the date in 2004 when arrangements were made with UKBetting, violated clause 10.3.6 of the Agreement.
78. The starting point is that it is common ground that such use as there may have been by GA in the course of providing services for UK Betting was not of commercial importance. In opening the case, Cantor's counsel candidly and helpfully said that it was not alleged that the use of GAMoney was important to the defendant's system, although it was pointed out that it was playing a role in the defendant's system and that, if it was being copied or used, that was unlicensed use. Cantor's approach at trial was in this respect consistent with what had been said on its behalf in a letter dated 16^th February 2007, in which an open offer to settle the claim was made. Without conceding that there was only a limited amount of commercial damage by the use of GAMoney, Cantor's solicitors accepted that it was possible that such a contention may have merit.
79. That position is also consistent with GA's evidence on the utility of GAMoney. Mr Lawrence said in his statement that as of 1^st August 2004, when the Cantor Golf game and the Clubhouse WGA web pages were removed from the WGA web-site, there was no longer any web-page that was bound to the GAMoney database and that the GAMoney database could have been easily removed. It is also consistent with the expert evidence on the matter, none of which (including Mr Carlisle's initial report of November 2006) suggests that the use of GAMoney was of real significance.
80. Despite the unimportance of the use, is the use of GAMoney nonetheless within the meaning of clause 10.3.6 of the Agreement? On the assumption that it is correct to apply a test for "use" akin to that of infringement, it is worth considering briefly what that test involves, before applying it to the facts.

Infringement by copying

81. Viewed from the perspective of pure UK copyright law, infringement of copyright by copying is a broad concept. It includes reproducing the work in any material form and includes storing the work in any medium by electronic means. Copying also includes the making of copies which are transient or are incidental to some other use of the work (see section 17, Copyright Designs and Patents Act 1988). This is consistent with EU law which, by the Computer Software Directive (Council Directive 991/250/EEEC of May 14, 1991 on the legal protection of computer programs), provides, in Article 4, that the permanent or temporary reproduction of a computer program by any means and in any form, in part or in whole, shall be the exclusive right of the rightholder.

Substantiality

82. In Designers Guild Ltd. V. Russell Williams (Textiles) Ltd [2001] FSR 11, at para. 61, Lord Scott of Foscote referred to the two types of case in which the question of substantiality of a part of a copyright work may be in issue. He said:

"One type of case is, obviously, where an identifiable part of the whole, but not the whole has been copied. For example, only a section of a picture may have been copied, or only a section or two, or even only a phrase, from a poem or a book, or only a bar or two of apiece of music, may have been copied....In cases of that sort, the question whether the copying of the part constitutes an infringement depends on the qualitative importance of the part that has been copied, assessed in relation to the copyright work as a whole. "

83. Designers Guild was not such a case, as Lord Scott pointed out. This case is. Here, if there has been copying at all, there has been literal copying of all or part of GAMoney, not reproduction of more general features. I therefore propose to apply the test of that aspect of Designers Guild.

Facts

84. Certain aspects of the copying relied upon do not (at least according to Cantor) require an assessment of whether the part copied was or was not substantial. That is because, if there was copying at all, it would have been of the whole of GAMoney. I therefore deal with these aspects first.

GAMoney as a whole - storage and backup

85. Cantor relies on two aspects of copying of the whole of GAMoney. First, the fact that it was on the server at all times (as a whole) and was live at all times and had to be so that smaller parts could be copied as and when needed. This is not disputed as a matter of fact although GA contends that such is not sufficient to constitute "use".
86. Second, the fact that back up copies were regularly made (including a back up on 23rd April 2004). Again, it was not disputed by GA's evidence that back up copies were regularly made (see Mr Lawrence's evidence, at pp153-154).
87. It should be noted that neither backups nor the mere fact of GAMoney residing on the database featured particularly strongly in Cantor's original expert evidence. The "back up" point, in particular, appears to have been something of an afterthought. That, however, goes more to the overall significance of this use than whether it was "use" within the Agreement at all. For copyright purposes, both types of copying would constitute infringements, absent licence.

Was GAMoney functionally redundant?

88. Mr Carlisle's evidence was that, without GAMoney running, an error would have been generated each time GAII_Build called the routine within GAMoney and failed to find it. That evidence was not disputed and is consistent with the presence of the arrow between GAII_Build and GAMoney in the Figure in paragraph 41 of Mr Lawrence's statement.
89. In my judgment, GAMoney was not "functionally redundant" between August 2004 and early 2006 (to use GA's terminology). Even if clause 10.3.6 included a requirement that the use be for some particular function (as opposed to being completely redundant and unused), the particular use made of GAMoney would, in my judgment have satisfied that requirement. It is in this connection, noteworthy that Mr Lawrence described the situation in paragraph 48 of his statement as one where GAMoney "sits idle" on the live database servers. That may be contrasted with the position described in paragraph 41, to which that description was not applied. I am therefore unable to accept GA's submission that it was common ground that any transient copies of GAMoney (or parts thereof) were functionally redundant in that they did not contribute to the operation of the GA suite of software. In particular, Mr Carlisle's evidence was to the contrary and Mr Du Vergier's evidence in paragraph 27 of his statement also indicates that parts of GAMoney were loaded into memory "for the purpose of access by GAII_Build".
90. In my judgment, the situation depicted in paragraph 48 of Mr Lawrence's statement is not within clause 10.3.6 of the Agreement and does not constitute breach. That is not because, absent a licence, it would not be an infringement of copyright for GAMoney to reside, even idly, on the GA server. It is because, if it merely so resided, it would not be used "in any collaboration or agreement with a third party Bookmaker." The contract does not prohibit the mere residence of GAMoney on the server. It prohibits such residence in the context of a collaboration or arrangement with a third part Bookmaker. In contrast, the situation depicted in paragraph 41 is within clause 10.3.6.

Parts of GAMoney - stored procedures

91. Although Mr Carlisle considered that GAMoney database might operate by compiling the stored procedures (and therefore making copies or substantial copies of them in RAM) when the system started up, as it turns out, the particular database operated in the SQL environment and does not do so. This can be seen from Microsoft's SQL Manual. Having seen this manual, Mr Carlisle accepted in cross-examination that the stored procedures were not compiled unless they were actually called. It is not suggested that any of the stored procedures in GAMoney were actually called by GAII_Build during the relevant period. It is not therefore possible to be satisfied that any copies were made of these stored procedures after the relevant date.
92. Accordingly, I do not believe that if there was use, it was as a result of copying of stored procedures. Moreover, although I do not believe that the parties did in fact turn their minds to the matter when the contract was made, it is highly unlikely that had they done so, they would have regarded the mere caching of otherwise unused procedures was to be caught by the term "use". The same applies to copies of parts of GAMoney made automatically by the system from time to time.

Parts of GAMoney - tables

93. It is common ground that GAII_Build references certain application tables in GAMoney. It is also not in dispute that, if a stored procedure in GAII_Build which references a particular application table in GAMoney is invoked, some or all of the metadata for that table will be copied into RAM.
94. After August 2004, GAII_Build continued to call 8 tables of GAMoney (GAMoney had 105 tables). Those tables would be copied when GAMoney was running. It is not, on the evidence as a whole (and especially that of Mr Du Vergier which I accept), possible to say with any confidence precisely what transient reproduction would occur as a result of calling those tables. The best one can say with any confidence is that it is highly likely that any table that is referenced is likely to have been partially or completely copied. Accordingly, on the balance of probabilities, at least some tables of GAMoney would have been copied.

Significance of copying of the GAMoney tables

95. GA contended that the use of the tables was not, in any event, shown to be sufficiently significant to constitute infringement in the Designers Guild sense. This is a point of some difficulty and hard to assess on the evidence. However, in my judgment once it is shown that a part of a work literally reproduced is not functionally useless, it is a short step to showing that a sufficiently substantial part has been reproduced which, in my judgment, is the case here. Although that principle cannot be taken to its most extreme conclusion, it is of utility in assessing the significance in the present case.

The overall picture

96. Finally, it is necessary to step back from the detail of individual uses and ask an overall question: would a person skilled in the art of software design consider that GAMoney was being "used", perhaps unimportantly, but nonetheless, materially, in the operation of the software overall for Bookmakers between 2004 and 2006? In my judgment, the answer to that question is "yes". Although not performing the function for which it was designed, namely a full-function database, overall, GAMoney was not useless for the operation of GA's software. To the contrary, it was used and GAMoney was stored in memory, backed up, parts of it were further reproduced and it was regularly accessed in order that GAMoney could perform the limited function for which it was still deployed by GA between 2004 (following the arrangements with the Bookmakers) and 2006 (when it was removed). It could have been removed, as Mr Lawrence says, or rendered "idle" in 2004, but it was not.

Conclusion on "use" of GAMoney

97. For all those reasons, in my judgment, GAMoney was used after August 2004 in the sense required by clause 10.3.6 of the Agreement. It follows that GA was in breach of contract and, that use being unauthorised, constituted an infringement of copyright.
98. In making this finding, however, it should not be thought that I attribute any real significance either to the breach or to the infringement. It is perfectly possible (as here) for a breach of contract to be real but nonetheless of limited, if any, commercial importance. In my judgment this was a minor breach, committed more through laziness than through any deliberate policy. It was capable of easy remedy and was quickly remedied in 2006 under threat of litigation.
99. I then turn to the second issue, namely whether injunctive relief and delivery up should be granted.

INJUNCTIVE RELIEF

General principles

100. The Court has a wide power to grant or withhold an injunction founded on s.37(l) of the Supreme Court Act 1981. Although injunctions are granted almost automatically in certain kinds of case, the court is nonetheless required to have regard to all the circumstances. As Lord Bingham said in South Bucks DC v. Porter [2003] 2 WLR 1547 at 1562-1563:

"Underpinning the court's jurisdiction to grant an injunction is section 37(1) of the Supreme Court Act 1981 conferring the power to do so "in all cases in which it appears to the court to be just and convenient to do so ".

In all cases the court must decide whether in all the circumstances it is just to grant the relief sought against the particular defendant. "

Intellectual property cases

101. It is normal for an injunction to be granted in a case involving infringement of intellectual property rights, once it has been established that there has been infringement and the infringement has not completely ceased at the time of trial. Part of the reason that an injunction will normally be granted in such a case is that, at trial, there often remains a dispute as to whether the defendant is or is not entitled to undertake the acts of which complaint is made. The defendant is usually maintaining that it does have the right to do so and may expressly or implicitly be threatening to do so.
102. The approach in copyright cases was set out by the Master of the Rolls, Lord Woolf, giving the judgment of the Court of Appeal in Phonographic Performance Ltd v. Saibal Maitra and others [1998] FSR 749 at 771:

"...where a person establishes infringement of copyright and a threat to continue infringement, an injunction will in the ordinary case be granted without restriction. ...But the court, when granting an injunction, is still required to exercise a discretion and in so doing there could be circumstances where restriction or refusal of an injunction would be warranted. "

103. A similar approach was taken by the Court of Appeal in the patent case, Coflexip S.A v. Stolt Comex Seaway MS Ltd [2001] RPC 182, where Aldous LJ said at paras. 6 and 7:

"...whenever a court at the end of a trial grants permanent injunctive relief, the purpose should be to give effect to its judgment on liability ...The injunction granted should protect the plaintiff from a continuation of the infringements of his rights by the threatened activities of the defendant. But the injunction must also be fair to the defendant. "

"....Normally, when a defendant has infringed, the court will assume it is not a one-off activity and will grant an injunction to stop repetition. This course is not inevitable. In a few cases courts have concluded that even though infringement has occurred, no future threat exists. In such cases, injunctive relief has been refused... "

104. The same principles must surely apply where a person establishes that there has been a breach of contract which prohibits an act akin to infringement of an intellectual property right.

No threat to infringe

105. The court may, however, refuse to grant an injunction and, ordinarily, would refuse to do so, where no threat to infringe existed at the time that the order came to be considered. In Landor & Hawa International Ltd v. Azure Designs Ltd [2007] FSR 9, Lord Justice Neuberger said, at para. 46:

"It seems to me plainly inappropriate in principle to grant an injunction in favour of a claimant who clearly and unequivocally agreed, before the action for the injunction even started, to refrain from taking that very action which the injunction would have forbidden him from taking. "

106. In that case, the defendant had initially offered an unequivocal undertaking before the action was commenced but the undertaking was withdrawn in terms which made it clear that the defendant was again threatening to infringe the claimant's rights. An injunction was held by a unanimous Court of Appeal to have been rightly granted at trial.

More complex cases

107. There are more complex cases in which, although the defendant may have clearly and unequivocally agreed before the trial not to undertake the actions which the injunction would forbid him from taking, it may nonetheless be appropriate to grant an injunction. Examples may be cases in which, for one reason or another, the court considers that it would be appropriate for the assurances given by the defendant to be backed by court sanction.
108. For example, there may be situations where a defendant has previously given contractual undertakings not to undertake certain acts but has acted in breach of those undertakings. Another example may be where there is a dispute over the scope of the undertakings given, with the claimant contending that the defendant is not free to undertake certain acts and the defendant contending that he is. In such situations, I do not understand Lord Justice Neuberger's statement of principle as deciding that the court is invariably bound to accept the defendant's agreement, clear and unequivocal as it may be, without regard to other circumstances which may place that agreement in context.
109. In such a case, the court may conclude that adequate protection for the claimant and its rights requires that the contractual undertaking is backed by an injunction so that compliance is, so far as possible, guaranteed. This may be, in part, because of the impact of an injunction, as compared with a contractual undertaking on the steps that a defendant will have to take in order to avoid serious penalties in the event of a further breach. Because liability for breach of an injunction prohibiting infringement of copyright or breach of contract is, unless specifically qualified, strict, a corporate undertaking subject to such an injunction will be liable for contempt of court, even if the source of the breach is, for example, an employee acting contrary to specific instructions. The nature of any penalty for breach of an injunction may depend, among other things, on the seriousness of the breach and on the steps that have been taken by the undertaking to prevent any breach from happening. A person or undertaking subject to an injunction may take greater care in compliance than one subject only to contractual restriction. So, for example, an injunction may be of greater utility than contractual undertakings where it appears that the defendant's approach to ensuring compliance is more casual than necessary to guarantee protection of the claimant's rights.
110. An illustration of this approach comes from British Telecommunications plc v. Nextcall Telecom plc [2000] FSR 697 where the defendant had given contractual undertakings which it did not fully honour. The claimant sought an injunction in unqualified form to enforce the contract. The defendant contended, in effect, that it was impossible to guarantee compliance with its obligations, because of the possibility that a rogue employee may continue to make the misrepresentations that its contractual undertakings were intended to prevent. Jacob J granted an injunction in unqualified form observing that he was "only enforcing by injunction precisely what the defendants undertook not to do by contract. " Jacob J also pointed out that it did not follow from the absence of qualification of the injunction that the claimant would sensibly bring proceedings for contempt of court for the slightest breach. He said:

"If they were nonetheless so to proceed, then if the defendants had truly taken all reasonable precautions to prevent a breach, it is most unlikely that they would be punished. There may well be no order as to costs... or even an order for costs against [the claimant] if the court thought the application pointless. All would depend on the circumstances. "

Trivial breaches

111. On the other hand, the court may refuse to grant relief where the interference with a claimant's rights is trivial (see Imperial Gas Light & Coke Co v. Broadbent (1859) 74 H.L. Cas. 600). The court has refused an injunction where the dispute was one for "the application of reason, common sense and ordinary forbearance, not for an injunction" (Behrens v. Richards [1905] 2 Ch 614) or where "the violation of the right is so slight, formal, and unsubstantial that the plaintiff can have no ground in conscience to complain of it..." (Harrison v. Good L.R. 11 Eq. 338, 352).
112. Mr Turner for Cantor refers to Insurance Co. v. Lloyd's Syndicate [1995] Lloyds Rep. 272 for the proposition that it is not necessary for a claimant to show that it is likely to suffer actual damage in order for an injunction to be granted. While his submission is correct, it does not follow that the court has no power to withhold injunctive relief where the breach in question is a technical one or is unlikely to occur again, or, if it does, it is unlikely to cause any real damage. That power is reinforced by one of the authorities cited to Mr Justice Colman in that case, Sharp v. Harrison [1922] 1 Ch. 502. An injunction was refused despite a finding of breach and a declaration was made that the defendant had breached the covenant. Nothing in the more recent approaches to the law, whether in the CPR or the EU Enforcement Directive (Directive 2004/48/EC of the European Parliament and of the Council of 29 April 2004 on the enforcement of intellectual property rights) or regulations made thereunder, suggests that the power of the court to refuse relief in the case of technical breaches or trivial matters has been removed. To the contrary, the concept of proportionality has been given greater status in recent times, reinforcing the fact that the court should not encourage costly litigation over very little by granting relief even in trivial cases.

Summary of principles

113. I therefore summarise the applicable principles as follows. First, an injunction may be granted pursuant to s.37(l) of the Supreme Court Act 1981 whenever it is just and convenient to do so. Second, the grant of an injunction involves the exercise of the court's discretion, and the court should, in so doing, take account of all of the circumstances, one factor of which is the importance or triviality of the breach. Third, there are certain kinds of case, of which intellectual property cases are examples, in which an injunction will normally be granted if a claimant has established infringement of its rights and there is a threat to continue (or at least no clear and unequivocal undertaking not to continue). Fourth, where there is no threat to continue acts which have been held to be unlawful, because the defendant has clearly and unequivocally agreed not to do them before the action was brought, it is not right in principle to grant an injunction. Fifth, there may, however, be situations where, even though a defendant may have agreed not to undertake the acts in question, an injunction may be just and convenient, having regard to all the circumstances. This may be, for example, because of the greater incentive for respect of a claimant's rights that an injunction would provide, and which, in particular cases, it may appear just to grant. Sixth, the court may, in appropriate cases, take proportionality into account in granting or refusing injunctive relief.

The relevant facts

114. Applying those principles, I consider the following matters to be of particular relevance.
115. First, it has been admitted that GA's use of GA Golf was an infringement of Cantor's rights. Despite that, GA Golf was used by GA for a considerable period. Although it was not directly put to Mr Lawrence that GA knew that its use of the Golf Game was an infringement of Cantor's rights, Mr Lawrence admitted that he knew in June 2005 that GA was using GA Golf. Mr Lawrence said that his understanding at the time (that is to say in 2005) was that the proportion of graphics was not a substantial part which is why it was there. He later accepted in cross-examination that it was plain that some copying had taken place. He accepted that GA's understanding at the date of the trial was that the parts of Cantor's software contained within the Golf Game constituted a substantial part. It was not however made clear what the basis of GA's previous understanding was that use of GA Golf would not have been wrongful. It was not, for example, positively said (as it might have been) that GA Golf had been used, notwithstanding the fact that it had been clear that parts of it were copied, because GA had been specifically advised that the parts in question were insubstantial.
116. Second, Ms Stratford-Martin on behalf of Cantor wrote to Mr O'Neal of GA on 30th June 2005. The letter expressed concern that the arrangement that GA had entered into with Paddy Power to offer the backgammon game on Paddy Power's web-site was a material breach of the licence agreement. The response from Mr John Jones, GA's Chief Executive, on 11th July 2005 stated that there was no breach of the clause of the agreement since the backgammon software game was developed under a completely different programming language called Java, that there was no breach of the licence agreement since "...none of the Intellectual Property Rights transferred to us in September 2002 are involved." Although that may have been true in relation to the backgammon game, it was not, as it turned out, true in relation to GA Golf. The letter could be read as intending to provide comprehensive reassurance to Cantor that "none" of its intellectual property rights were being used, which GA later admitted was not the case.
117. Third, after receipt of the more formal letter before action dated 16th February 2006 which raised the question of infringement of rights in GA Golf, GA took apparently swift action both as regards GA Golf and GAMoney. GA Golf was withdrawn immediately (subject to the point mentioned below) and GA offered undertakings in its letter of 24^th March 2006. GA arranged for copies of GA Golf which it believed to be available on the web to be taken down (or at least rendered inaccessible). GA also arranged to remove GAMoney and appear to have focused specifically on this because of the threat of litigation. The e-mail from Jesse Easton of GA dated 28^th March 2006 entitled "Removal of GA_Money references" is indicative that a specific request had been made to remove all such references and that, by the 28^th of March 2006, all of the GAMoney references had been removed or commented out in the dev and QA environments (albeit not yet the live environment). That e-mail sought confirmation that doing so would not result in anything odd happening.
118. Mr Lawrence was cross-examined as to whether the removal of the references to GAMoney was because of a belief that Cantor would be likely to complain about it if they found out. He originally said that he never had it in his mind that GAMoney was ever doing anything, so there would not really be a problem with it being there and that there was no pressing technical reason to remove the references. While credit must be given for the rapid removal of the references to GAMoney when the complaint was raised it is in my judgment right also to take into account the fact that removal of GAMoney was both contemplated earlier and could have taken place earlier. In particular, the e-mail from Mr Lawrence to Mr Bennett, dated 9^th August 2004 asks (among other things) "Can't GAMoney disappear now?". Although Mr Lawrence said there were technical reasons why GAMoney could have "disappeared" at that earlier stage, in fact it did not do so. Mr Lawrence gave a credible explanation for why it was not removed before. In addition to this evidence referred to above, he said in his oral evidence that the reason that this was not done was that there was no functional or technical reason to do it and that the database developers did not get round to doing it. His evidence, which I accept, was as follows:

"I think developers are fairly lazy people. If there is not a good technical reason for doing something, then it is not done. I think also it probably required a bit of care. The change itself would have been very easy, but it was changing code that could have made the system unstable. So it was a very low reward for a task which was unnoticed, which I guess is why for a developer it was not very attractive to do. "

119. That specific description of the approach of its developers in this case is supported by the more general opinion of Mr Du Vergier who referred in his cross-examination to the fact that programmers tend to be quite "promiscuous" (by which he meant that they tended to use code without too much cognizance of rights unless they were expressly told not to). He accepted that there was quite a lot of bad practice going on. Understandable though that may be, it is not a defence to infringement of copyright to say that preventing the infringement is a "low reward" task. The approach of the developers described by Mr Lawrence is not calculated to guarantee respect for the intellectual property rights of others.
120. Fourth, on the other side of the argument, GA's use of GAMoney is entirely historic. Cantor does not suggest that its use was important after June 2004 or that its use has at any time caused Cantor significant damage. Those are powerful points in favour of GA's position as regards injunctive relief with respect to GAMoney. Set against that, however, GA has resisted any relief being granted in relation to GAMoney. Its stance has been that such use as there has been of GAMoney has not been wrongful although it has sought to end any such use. That is to say, GA was maintaining at trial that it had a right to use GAMoney in the way in which it had been before the complaint was made, although it was not in fact proposing to do so and it was clear that there would be no benefit in its doing so as of now.
121. Fifth, GA submits that (i) Cantor was quite content with a contractual undertaking and (ii) there is no reason to consider that GA would breach it. GA is right as to the former and that, too, is a strong point in its favour. However, the latter has become less clear because, following the main hearing and a considerable time after undertakings had been offered, one of Cantor's solicitors discovered that a demonstration version of GA Golf was still accessible via a Google search of the web.
122. It is important not to make too much of this material. It has not been suggested by Cantor that this use is of particular significance as such and GA's witnesses (in particular Mr Lawrence and Mr Murray) have provided a reasonable explanation for why this way of accessing GA Golf was not picked up before. That means of access has now been removed and Cantor does not suggest that this is likely to happen again. Moreover, GA says, with some justification, that had the undertakings originally offered been accepted, this continued use would not have happened because the copies would already have been delivered up. It is not, of course, possible to blame Cantor for that but it highlights the fact that the continued use was both unintentional and of limited importance. In this connection, however, GA offered to submit to an injunction if breaches of the offered contractual undertaking came to light. That being so, it is not unreasonable, in the light of the recently discovered uses, for an injunction to be put in place.

Summary

123. The overall picture that emerges from the evidence is that an injunction may have some, albeit probably marginal, utility in enforcing the contractual undertakings previously offered by GA and ensuring that GA was assiduous in removing the software from offending use, should that be necessary.
124. It is therefore, in my judgment, just and convenient to grant an injunction in this case, having regard to all the circumstances, notwithstanding the questionable proportionality of the proceedings to the matters at stake, at least in so far as they relate to GAMoney. It would not be unjust to GA, in particular because the injunction would (as in British Telecommunications) only involve enforcing that which GA had undertaken not to do by contract and has in fact taken steps to ensure is not done in the future. There is no suggestion that it would be inconvenient.
125. Given the way in which matters have progressed and the doubtless chastening experience of this litigation for GA, it may be for question whether the grant of an injunction would make more than an incremental difference. That doubt is not so strong as to justify refusal of the injunction given the fact that such is normal in intellectual property cases. That said, if the right test for grant of an injunction is whether it is really necessary to ensure that Cantor's rights are protected, in my judgment the answer is that it is not necessary but, as the law stands, Cantor does not need to go that far to be entitled to injunctive relief.

Delivery up

126. I do not consider that separate principles apply to delivery up in this case, which is usually undertaken in aid of an injunction, to ensure that there is no stray use.

Conclusion on injunction

127. An injunction will therefore be granted as will delivery up, in terms which it is hoped may be agreed between the parties. The injunction will relate both to the use of GA Golf and GAMoney.
128. It remains for question whether an injunction would have been justified had the only matter in issue been GAMoney. The use of GAMoney, albeit in technical breach and infringing in the manner determined above is accepted to be of limited importance.
129. That matter may need to be taken into account in connection with costs. The British Telecommunications case is illustrative of the proposition that the mere fact that a person is held entitled to relief of a given kind, does not automatically entitle that person to the costs of obtaining it, if the breach by the defendant is of minimal importance or the proceedings pointless. It will be for separate consideration whether the cost of correcting a degree of casualness on the part of GA of limited importance to Cantor, as regards GAMoney, justifies the benefit of doing so.

CONCLUSION

130. For the reasons given, in my judgment:

(a) The use made by GA of GAMoney from 2004 to 2006 constituted a breach of the Licence Agreement and an infringement of Cantor's copyright, notwithstanding its comparatively minor nature and the fact that it could easily have been avoided.

(b) An injunction and delivery up should be granted to restrain further breach in terms to be agreed between the parties or determined by the court in default of agreement.

(c) Breach of contract and infringement of copyright by the use of GAGolf has been admitted. There is no dispute that there should be an inquiry as to damages or an account of profits in respect of that use.

Form of order and costs

131. I will hear further argument on the form of order and costs, if they cannot be agreed. I conclude by saying that both sides, including solicitors, counsel and experts, have prepared this case with intelligence and care and have presented it excellently.