Mars U.K. Ltd. v. Teknowledge Ltd.
High Court of Justice – Chancery Division
 F.S.R. 138
Mr Justice Jacob
1. Mars are leaders in the design and manufacture of coin receiving and changing mechanisms. Such machines include discriminators whose function is to determine the authenticity and denomination of a coin fed in to the machine. Once upon a time discriminators of this general sort operated purely mechanically. They now work by using sensors consisting of coils which take a series of electrical measurements of a coin as it passes through the discriminator. There are sensors which measure the thickness, diameter, electrical resistivity and, most recently, inductance. The signals from the coils are compared with pre-determined sets of data for valid coins. This data is recorded in an electronic memory on a chip. Such a chip may be a PROM (“programmable read only memory”) or an EEPROM (“electronically erasable programmable read only memory”). A PROM once programmed cannot be reprogrammed. An EEPROM can be reprogrammed to contain new data.
2. The discrimination process includes the use of algorithms (mathematical recipes) which combine the outputs of the sensors in an effective manner to ensure that each valid coin gives an ultimate set of outputs which can be distinguished from duds, foreign coins or other non-valid discs (collectively called “slugs” in the trade). Coins, even of a particular denomination, are not of course all identical. There may be variants arising from the mint (e.g. minor differences between different sorts of £1 coin or arising from a change of alloy by a mint). And coins, of course, wear. Each parameter measured therefore has to have a “window” of acceptable values. If two coins of a particular set of coinage have similar parameters then the window for each will have to be narrow. So also if there is a common form of slug which is close to a particular genuine coin. A very great deal of experimental work as well as skill and judgment is involved in determining appropriate windows for each coin of a given coinage. Mars have a vast collection of sets of coins for many countries of the world and have to keep under constant review the windows for the coins of any particular set.
3. Until about 1992 the Mars electronic coin discriminator, known as the Classic, used a PROM. So if a change of coinage were to occur and it was desired that the device should accept a new coin it was necessary to replace the PROM with a new PROM containing the data for the new coin, and perhaps new data for old coins (for instance when a narrower window was*142 needed). This could also be done by writing to previously unused areas of an existing PROM.
4. In 1990 Mars set about designing a new discriminator which has become known as the Cashflow. The Cashflow has the ability to be reprogrammed for new coin data. It also can do other things with which I am not concerned. It uses an EEPROM. It is necessary to be able to communicate i.e. to read from and write to the EEPROM. This was not necessary or possible with a PROM. It was commercially necessary to make access difficult. For one thing, if it were possible easily to change what coins or slugs would pass the machine as genuine, that would be an open invitation to fraud. In addition to quality control considerations (which of course could affect the reputation of the Cashflow generally) Mars for commercial reasons also wanted to reserve to themselves and their approved agents the re-programming of Cashflow machines to take new kinds of coin or new variants of existing coins. I use the word “re-programming” but in the trade this activity is called “re-calibration”.
5. Mars, for the Cashflow, therefore developed a data layout, a serial communications protocol and an encryption system. None of these are published directly by Mars. Mars recognised and intended that what they had done would be difficult but not impossible to “crack”, i.e. that it would be difficult for some third party to work out how to communicate with the EEPROM of a Cashflow or to find out what windows were used by Mars for particular coins. Similarly, the precise algorithms used by Mars are unpublished and are held within the microprocessor used in the discriminator. Mars would have liked to make it impossible for third parties to find out by reverse engineering how to re-calibrate their Cashflow discriminators, but this could not be done within a reasonable cost.
6. Mars have arranged for the alteration of discriminators in the field to be updated by a number of authorised service companies. These are independent of Mars. What they do is to hire from Mars a tool (at an ultimate cost of £99) which provides access to the EEPROMs of Cashflow machines. There is a fee of £9.50 per discriminator reprogrammed, though in some cases other amounts of about the same order are changed. A new discriminator costs about £30.00. The operators are not given all the unpublished information contained in the Cashflow machines.
7. The defendant is a small company based in Leicester which had its beginnings in the early 90s. Its moving spirits are Mr Hogan and Mr Rawding. They have developed great skills in the field of discriminators, as is clearly demonstrated by their history. They are approved agents for a number of companies, including Coin Controls who are also industry leaders. Last year they were given an award by BT for their work in discriminator re-programming.
8. Some time ago Teknowledge, by reverse engineering, learnt how to write to new PROMS or to write to new channels of old PROMs for updating the Mars Classic discriminators so that they can accept changes in the coinage. No complaint is made of that activity. It is also done by other companies. It is Teknowledge’s activities in relation to the Cashflow which*143 gives rise to this litigation. Only Teknowledge have, thus far, succeeded in reverse engineering the Cashflow.
9. Mars claimed infringement of copyright database right. It seemed clear from the witness statements that the real defences were not a denial of acts which would amount to infringement but a defence in relation to the copyright and data protection claims of a common law right to repair along the lines of that found by the House of Lords in British Leyland Motor Corporation Ltd v. Armstrong Patents Co. Ltd.,  A.C. 577, and a denial of breach of confidence as a matter of law rather than fact. The concessions by Teknowledge were that, subject to a British Leyland defence, it has infringed:
1. Mars’ copyright and database right in the Coin Set Data by (a) reproducing lower limits and window widths in the course of developing its reprogramming software and (b) reproducing window widths in its reprogramming software and in customers’ mechanisms which it has reprogrammed.
2. Mars’ copyright in the Discrimination Algorithms and in the Program Code which implements the Discrimination Algorithms by reproducing the Discrimination Algorithms in its reprogramming software.
3. Mars’ copyright in the Program Code which implements the HII serial communications protocol by reproducing the HII message structures and contents in its reprogramming software.
4. Mars’ copyright in the Program Code by making transient copies of the Code during the course of developing and operating its reprogramming software.
Teknowledge accept that if someone were to make a poor job of reverse engineering that could, in principle, lead to badly re-programmed discriminators and hence to security problems of the sort whereby machines would pass slugs. This point was said to be*144 relevant to Mars’ attack on the applicability of the British Leyland defence, but in the end I think it is not.
10. Teknowledge also were prepared to admit for the purposes of this action that it knew and understood that the Claimant had included encryption within the Cashflow device and that by so introducing such encryption the Claimant was seeking to restrict persons from assessing and analysing how the Cashflow mechanism operated, but notwithstanding such encryption the Defendant believed at all material times that it was entitled to assess and analyse the Cashflows the subject of this action.
11. In the result, I have to decide whether there is a common law defence to the otherwise conceded claims for copyright infringement.
The British Leyland defence: Is “re-calibration” within it?
12. The only activity of Teknowledge of which complaint is made is the “re-calibration” of Cashflow discriminators so that they will accept new, or new versions of old, coins. Complaint is not made of re-calibration done where, for some reason, a machine will no longer accept a coin it should accept and once did accept – where, for instance, somehow something has “drifted” from its intended and originally set value. I mention this because at one point a bit of Mr Vanhegan’s cross-examination seemed to relate to this latter sort of “re-calibration” and Mr Silverleaf was minded to ask for leave to adduce further evidence relating to the point. On the agreed basis that the case was confined to the issue I have such stated, the further proposed evidence was unnecessary and it was not admitted.
13. Thus the question is whether the common law can override statutory rights by way of copyright and database rights to enable ultimate “consumers” and hence agents appointed by them or their behalf to adjust the discriminators to take new coins. The leading authorities on this area of law are British Leyland and Canon Kabushiki Kaisha v. Green Cartridge Company  A.C. 728, both decided under the Copyright Act 1956.
In broad terms, Leyland recognised a “spare parts” defence to copyright infringement and Canon held that any such defence should not be extended and in particular should not be extended to “consumables” such as copier cartridges.
14. Before considering the point in more detail I must deal with Mars’*145 contention that the Leyland defence has not survived the passing of the Copyright Designs and Patents Act 1988 or the amendments to that Act arising from Community Directives. It is first necessary to identify what the 1988 Act did in relation to copyright and industrial articles. It set out the position for the future in section 51 – essentially abolishing industrial copyright as it had become recognised under the 1956 Act. Accordingly, for the future, in relation to industrial designs there is no need for a spare part exception — there is no right from which exception is needed. The Act also introduced, by Part III, unregistered design right. In relation to that section 213(3)(b) excludes from protection:
features of shape or configuration of an article which –
(i) enable the article to be connected to, or placed in, around or against another article so that either article may perform its function, or
(ii) are dependent upon the appearance of another article of which the article is intended by the designer to form an integral part.
This “must fit or must match” exception will cover some spare parts but not others, depending on how necessary it is to copy.
15. As regards pre-Act works, there are transitional provisions specifically provided by paragraph 19(9): ‘ Nothing in this paragraph affects the operation of any rule of law preventing or restricting the enforcement of copyright in relation to a design.’ That has the effect of retaining the Leyland defence, which of course remained necessary in respect of pre-Act drawings for a period of 10 years. Were this not done, Armstrong and others would have had to stop making spares.
16. Thus in relation to industrial designs Parliament has specifically catered for the position in the Act. But this case is not concerned with industrial designs. The rights relied upon are computer programs, database rights and literary copyright (in the algorithms, copyright in which was not contested).
20. Accordingly I conclude that in relation to the rights relied upon here, there is no spare-parts or analogous defence. Even if that were wrong, however, I further conclude that the re-calibration activities here are not within the scope of any such defence. I turn to explain why.
21. In Leyland, the House of Lords permitted the manufacture of spare exhaust systems for motor cars, notwithstanding the fact that by so doing the defendant’s were indirectly copying the plaintiff’s copyright drawings. It is unnecessary to go to the leading speeches in that case because they have been authoritatively analysed by Lord Hoffmann in Canon. In the latter case, the Privy Council held that the “spare parts exception” applied only where it was plain and obvious that the replacement was analogous to a repair which an ordinary purchaser of an article would assume he could do for himself without infringing the manufacturer’s rights, or that the exercise of monopoly power by means of copyright would be against consumers’ interests.
22. Lord Hoffmann held that the spare parts exception:
Cannot be regarded as truly founded upon any principle of the law of contract or property. It is instead an expression of what the House perceived as overriding*147 public policy, namely the need to prevent a manufacturer from using copyright (as opposed to patents or design right) in order to control the after market in spare parts.
23. This is important. It means that the “spare parts” exception is founded on public policy and is yet another example of the general public policy defence to copyright claims. Mr Vanhegan boldly suggested that I should not follow the authoritative analysis in Canon of the speeches in Leyland.He had to so contend because such economic evidence as there is in this case (to which I refer more below) is far too sketchy to show that public policy considerations overwhelmingly override Mars’ statutory rights.
24. One advantage of Lord Hoffmann’s analysis of the spare parts defence is that it leads to a more rational result even in the case of spares. No issues of quality arose in Leyland, but suppose the part concerned were of vital importance to safety – inferior brake pads for commercial aeroplanes, for example (a real example, as I recall from my practice at the Bar). It is difficult to suppose that the result would have been the same, yet if one goes by property or contract concepts, one cannot distinguish between cases where it is in the public interest for the spare to be made available and where it definitely is not.
25. I therefore conclude that there is no overwhelming public policy reason entitling those who purchase machines with discriminators to use Mars copyright and database rights to convert those machines for new coins. In so holding I also bear in mind what Lord Hoffmann also said in Canon:
It is of course a strong thing (not to say constitutionally questionable) for a judicially-declared head of public policy to be treated as overriding or qualifying an express statutory right. Their Lordships therefore think that the prospect of any extension of the British Leyland exception should be treated with some caution.
26. Lord Hoffmann pointed out that the kind of economic question involved is not really “one to be solved by broad generalisation and that the Courts are ill equipped to pronounce upon such matters”. I mention that here because I heard some evidence about costs and prices. Teknowledge’s prices are perhaps somewhat less than that of Mars, though it is not clear by how much in the case of Cashflow re-calibration. On the other hand, Mars have considerably greater overhead in creating and maintaining the various copyright works and databases concerned. None of these figures are precisely quantified, nor are any figures available for Mars’ ongoing costs for research into even better discriminators. Nor do I have any direct evidence from “consumers” (i.e. ultimate purchasers) as to their expectations concerning reprogramming for new coins and in particular as to*148 whether, when buying, they “factor in” a need to go to Mars for such re-programming. Such consumers might know that earlier machines could be re-programmed by third parties, but it does not follow that they would expect this to be possible for any new machine – indeed it was not possible for some time.
27. What is clear is that the purchasers of discriminators are not the kind the ordinary consumer: they are not the ordinary man who bought an ordinary article. The example of repair by a blacksmith fastened on to by Lord Bridge in British Leyland has no compelling analogy with this case. This is so for two reasons. Firstly those who buy, or have, sophisticated devices operated and controlled by computer programs normally look to the original manufacturer for repair and maintenance, and updating of the programs involved. When they buy the article they have no expectation of a handy “computer blacksmith” for updating the program, rather they expect updates to come from the manufacturer. Secondly, altering the machine so that it will respond to different coins is, to my mind, even further from the concept of repair than the supply of consumables such as the cartridges considered in Canon. Consumables run out quickly and one expects to have to buy more. Changes in coinage have nothing to do with the operation of the machine at all. One expects that something may have to be done about accepting new coins, but I see no reason why one should expect to be able freely to use the manufacturer’s intellectual property rights if one wants one’s machine altered.
28. Thus I conclude that there is no common law defence available in this case.