International Businesss Machines Corporation & Anor v Web-Sphere Ltd & Ors

Despite the heading, IBM needs no introduction. It is the proprietor of a registered Community trademark. The registered mark is “WEBSPHERE”. It is a word mark rather than a device mark. The classes of goods and services for which it is registered include Class 9 (in particular computer software and computer software for web site development), Class 38 (in particular communications services, communication by computer terminals, electronic mail and computer-aided transmission of data) and Class 42 (in particular services for developing and maintaining web sites and services for providing user access to computers for business management).

In May 1998 IBM launched a software application called WebSphere Application Server Version 1.0. It is what is known as “middleware”. In essence it sits on top of an operating system (such as Microsoft Windows) and “glues together” a computer application program and a network of connected computers. WebSphere has grown in importance to IBM and is now one of its core brands.

The First Defendant began life on 28 May 1997 when it was incorporated under the name Publiweb Ltd. It changed its name on 9 June 1999 to Web-Sphere Ltd. The hyphen is important. Web-Sphere Ltd has three Internet addresses. They are www.web-sphere.com; www.web-sphere.net and www.web-sphere.org. These domain names were allocated on 12 May 1999. It has provided a range of internet-related computer services under the name Web-Sphere since at least June 2000. According to its web site (last updated in January 2001) it offers domain name registration, domain name hosting, domain name forwarding, web site hosting, search engine registration, internet consultancy, email management, rental of server space and web site, design, development and updating. By late 2002 it announced the launch of Web-Sphere software as well.

IBM says that this amounts to an infringement of its Community trademark.

Web-Sphere’s defence involves a number of strands. First, it says that IBM’s trademark should never have been registered. Second, it says that its sign “Web-Sphere” (with the hyphen) is not the same as the registered mark. Third, it says that it has not used its sign in relation to the same goods and services as IBM’s mark. Fourth, it says that even if the sign is similar to IBM’s mark, and even if the goods and services in relation to which it has been used are similar to those in relation to which IBM’s mark has been used, IBM has not shown that there is a likelihood of confusion. Fifth, it says that the dispute is really about the domain names and that the court is not the appropriate place to resolve that dispute.

In addition to the claim of trademark infringement, IBM also sues all three Defendants for malicious falsehood (or trade libel as it is sometimes called). This arises out of a leaflet that was distributed to customers of IBM and others in May 2002.

Mr Iain Purvis appeared on behalf of IBM. Mr De Serville appeared both on his own behalf in person and on behalf of Web-Sphere Limited (of which he is a director). Mr Markson appeared in person.

In 1995 IBM realised that there would be commercial advantages in developing enhanced Web servers to make them more reliable, scalable and high performing. It decided therefore to develop a new form of “middleware” to provide transaction processing. A team to develop the software was created in 1996. In 1997 a group of companies within the IT industry began to develop products using the Java Technology. These companies included IBM.

In March 1998 IBM began to develop a new product name for its application. The idea was that the selected name should have the flexibility to enable IBM to use it for a whole generation of Web based products. Various names were proposed and tested. Ultimately the chosen name was “WebSphere”. IBM’s legal team carried out checks on the availability of the name. The searches included many countries and included registered and unregistered trademarks, company names, Internet sites and domain names. The search did, however, discover a company called “Websphere Inc” in California. IBM entered into a trademark co-existence agreement with that company in May 1998.

By May 1998 the WebSphere Application Server Version 1.0 was ready for launch. The launch took place at IBM’s annual conference for software developers held in Orlando, Florida between 5 and 8 May 1998. It was accompanied by press releases which, in turn, were picked up by the trade press. For example, an article about WebSphere appeared in PC Magazine in July 1998. The press releases were also distributed through Business Wire and Market Wire, each of which distributes to thousand of people. Product shipments began in June 1998 and shipments to Europe began in September 1998. The product earned revenue for IBM amounting to $8.3 million in Europe the Middle East and Africa (of which $1.9 million was earned in the UK). Version 2.0 was launched in December 1998. This represented a major commitment by IBM to the Java technology. Early customers included Charles Schwab, the US based financial and investment advisers, and the Weather Network in the US.

Since December 1998 the WebSphere Application Server has come in three forms: Standard, Advanced and Enterprise. Each is more sophisticated (and hence more expensive) than the next. Standard is regarded by IBM as something of a loss leader and sells for about $1,000 per licence. In January 1999 IBM rebranded its Net.Commerce products as “WebSphere Commerce Suite”.

Version 3.0 of the WebSphere Application Server was launched in September 1999. Also in September 1999 IBM added to the WebSphere product line by rebranding “WebSphere Studio”, and by announcing the integration of all its transactional server technologies into the WebSphere family of products. Since early 1999 IBM has added a product to the WebSphere brand about every six months. Version 3.5 of the WebSphere Applications Server was launched in 2000 and Version 4 in 2001.

From 1998 onwards IBM has done what it can to promote WebSphere. There has been spending on advertising, exposure at trade fairs and the like. Advertisements have appeared in Computing, Computer Weekly, Information Week UK, IT Week, Computer Business Review and other magazines. Advertisements have also been placed on web sites and search engines. The trade press has also covered the various versions of the software. During 1999 there was a lot of interest from potential customers, many of whom were multi-national companies. IBM personnel organised and gave presentations about WebSphere to customers and potential customers. In May 1999, for instance, IBM gave a presentation at a “Java Briefing Day” at Bedfont Lakes in the UK. The briefing was attended by about 230 people, representing some 70 to 100 different companies. IBM also offered training courses in how to use the new software.

Revenues for the WebSphere Application Server have grown considerably. Revenues in the UK grew to $10.26 million in the first half of 1999, $17.87 million in 2000; and by 2003 the revenues for all WebSphere branded products had risen to $71.05 million.

I conclude that by June 1999 IBM had generated a substantial goodwill and reputation in the brand name “WebSphere”; and that that reputation and goodwill has continued to grow.

IBM applied to register its Community trademark on 6 May 1998. It was published for opposition purposes on 6 September 1999 in the Community Trade Marks Bulletin. No opposition proceedings were begun, so the application proceeded to registration on 20 April 2000.

I have already described the registered mark and the classes for which it was registered.

IBM first learned of the existence of Web-Sphere Ltd as a result of a letter from a Mr Roger Taylor of International Internet Investors Corporation dated 30 June 2000. This is a curious letter, because Web-Sphere Ltd has consistently said that Mr Taylor and his company is nothing to do with them. But Mr Taylor gave correct information about the formation, change of name and business of Web-Sphere. He also said that Web-Sphere owned the domain name “Web-Sphere.net”. This information was subsequently checked and verified by Mr Christopher Ling, who works in IBM’s Intellectual Property department. In his letter Mr Taylor said (among other things):

Mr Ling wrote to Web-Sphere at its registered office on 7 July 2000. He referred to IBM’s Community trademark and alleged that Web-Sphere’s activities infringed the mark. He requested that Web-Sphere should change its name and asked it to deregister its domain name.

This prompted a reply from Mr de Serville on 13 October 2000. Mr de Serville said that the allocation of the domain name preceded the registration of IBM’s trademark; and he accused IBM of “reverse domain name hijacking”. He alleged that the application for registration had not been properly publicised. Among the other points that he made in that letter was that IBM’s mark was “Websphere” (all one word) whereas Web-Sphere was hyphenated. He said:

Mr Ling replied on 3 November 2000. He disagreed with the assertion that the domain names had been allocated before registration of the trademark. He refuted the allegation that the application for registration had not been properly publicised; and reasserted IBM’s claim that its registered mark had been infringed.

In Mr de Serville’s reply of 1 December 2000 he again asserted that IBM’s application had not been properly publicised and again asserted that whereas IBM’s mark was one word, Web-Sphere’s name and domain name was hyphenated. He also said that there had been no instance of actual confusion between the two.

There the matter rested for many months. On 26 July 2001 Mr de Serville wrote again. He pointed out that he had had no reply to his letter of 1 December and asked whether he was correct in assuming that IBM accepted the points he had raised. He also asked for confirmation that IBM had no objection to Web-Sphere Ltd applying for the registration of “web-sphere.net” as a Community trade mark and also as a trademark in the USA. Mr de Serville received no immediate reply to that letter either.

On 13 September 2001 IBM held an event for customers at the Marriott Hotel in the former County Hall. Mr de Serville and Mr Markson booked to attend. They booked under their own names, but as representatives of COM Solutions Ltd, a company which owned most of the shares in Web-Sphere Ltd. At the start of the meeting Mr Markson made his way to the front of the room and made a short speech critical of IBM. Mr Markson and Mr de Serville then handed out leaflets to the assembled company. They drafted the text of the leaflet jointly. I quote some extracts from it:

This incident prompted a response on 4 October 2001 from IBM’s solicitors, Bristows, who reasserted IBM’s claim for trademark infringement. Correspondence (both open and without prejudice) followed.

IBM’s response did not satisfy Mr de Serville or Mr Markson. They decided to attract more attention by redistributing the leaflet that they had handed out in September. Together they wrote some new text and photocopied the existing leaflet on the back. About 1000 leaflets were distributed by post to The Times top companies. In addition on 30 May 2002 Mr Markson went to IBM’s Customer Education Centre at Farnborough, where he placed about 300 leaflets under the windscreen wipers of cars parked in the car park. I have already quoted the relevant parts of the old leaflet. The new text read (in part) as follows:

It is the distribution of this double-sided leaflet that IBM says amounts to a trade libel.

In about November 2002 Web-Sphere Ltd produced about 5,000 copies of a glossy leaflet. This announced Web-Sphere Software launching in February 2003. It invited registration by e-mail, and it also gave an address and telephone number for Web-Sphere Ltd. Mr Markson distributed some copies of this leaflet by again going to Farnborough and placing them under the windscreen wipers of cars in the car park.

Community trademarks are governed by Council Regulation 40/94, which is directly applicable in the United Kingdom. A Community trademark is a trade mark for goods or services registered in accordance with the Regulation. A registered Community trademark gives the proprietor certain exclusive rights. These are described in Article 9 (1), which says:

Thus Article 9 (1) (a) gives absolute protection, but it depends on identity between the mark and the sign and identity of goods and services. Article 9 (1) (b) only requires similarity, rather than identity, but it also requires a likelihood of confusion. Although Article 9 (3) says that these rights prevail against third parties from the date of publication of the registration, Article 46 says that a Community trade mark must be registered for a period of ten years from the date of filing of the application. In effect, therefore, the registration is retrospective to the date of filing.

In LTJ Diffusion v. Sadas Vertbaudet[2003] FSR 34 the ECJ considered the criteria to apply in deciding whether a mark and a sign are “identical”. They said:

In Reed Executive plc v. Reed Business Information Ltd[2004] EWCA (Civ) 159 Jacob LJ described this guidance as “opaque”. He went on, however, to illuminate its opacity. Jacob LJ pointed out that:

There is a tension between para 50, which requires a strict interpretation of identity and para 52 which requires the court to allow for the imperfect picture of the average consumer. Although, as Jacob LJ said, the latter is a concept more aptly related to the question whether there is a likelihood of confusion, it is clear, in my judgment, that the ECJ held that same concept applies to the absolute protection given by Article 9 (a).

An addition in the sign (as compared with the mark) may take the case outside one of “identity”. However, the addition (or difference) may be so minute as to leave the sign and the mark identical.

As a matter of policy there is no reason to suppose that the ECJ meant to soften the edges of “identity” very far; because, even where a mark and a sign are not identical, the proprietor of the mark will be protected if there is likelihood of confusion.

It will be observed that in para 52 of their judgment the ECJ referred to para 26 of their previous decision in the Lloyd Schuhfabrik case [2000] FSR 77. That was a case which discussed the criteria to be applied in deciding whether there was a likelihood of confusion between a mark and a sign. By referring explicitly to that paragraph it seems to me that the ECJ must have held that the same criteria are to be applied in deciding whether a mark and a sign are identical. In the Lloyd Schuhfabrik case the ECJ said:

Thus it follows that in deciding whether a mark and a sign are identical, I must base a decision on the overall impression created by them, including their visual, aural and conceptual similarities.

There is one important exception to the rights given by a Community trade mark. A Community trade mark does not entitle the proprietor to prohibit a third party from using in the course of trade:

It is clear that the proviso applies to all three sub-paragraphs of Article 12 (and to this extent the layout of the Regulation as reprinted in Kerly’s Law of Trade Marks and Trade Names 13^th ed. p. 1039 is misleading).

The boundaries of the “own name” exception are not entirely clear. First there is the question whether the “own name” exception applies only to natural persons or to artificial persons, such as companies, as well. This is a question that has been referred to the ECJ, but the provisional view of the English courts is that a company can, in principle, rely on the exception: Scandecor Developments v. Scandecor Marketing [2001] 2 CMLR 30. I shall assume that this is so. However, as Peter Gibson LJ said in Asprey and Garrard Ltd v. WRA (Guns) Ltd[2002] FSR 31:

This is, no doubt, because the adoption of a name by a new company would not be use in accordance with honest practices in industrial or commercial matters: see also Reed Executive plc v. Reed Business Information Ltd para 116 (iii). What goes for the formation of a new company must apply equally to a change of name by an existing company.

In considering whether an alleged infringer has used his own name, it does not matter that the word “Ltd” or other indication of incorporation is omitted: Reed Executive plc v. Reed Business Information Ltd para. 115.

The condition of honest practice constitutes in substance the expression of a duty to act fairly in relation to the legitimate interests of the trade mark owner: Gerolsteiner Brunnen GmbH & Co v. Putsch GmbH (Case C-100/02) para 24; Reed Executive plc v. Reed Business Information Ltd para 126.

The fact that a trader’s own name is used as a trade mark does not necessarily take the case out of the “own name” exception: Reed Executive plc v. Reed Business Information Ltd para 125. Even if there is some actual confusion with a registered mark, the exception may still apply. The amount of confusion which can be tolerated is a question of degree. Only if what the trader does, viewed objectively, amounts to unfair competition will there be infringement: Reed Executive plc v. Reed Business Information Ltd para. 129. The test is both objective and one of reasonable foreseeability. The fact that a trader may honestly believe himself to be entitled to use his sign is not relevant: Asprey and Garrard Ltd v. WRA (Guns) Ltd para 49. In considering what is reasonably foreseeable, the court will expect a trader, before starting to trade under a particular name, to search the national and European registers of trademarks, and if alerted to a potentially conflicting mark, to have made reasonable investigation as to whether it has been used enough to have acquired a reputation or goodwill: Reed Executive plc v. Reed Business Information Ltd para 131 (ii).

Although a national court dealing with an action relating to a Community trade mark must normally treat the trade mark as valid (Article 103) this does not apply in an action for infringement (Article 92). Thus it is a defence to an action for infringement to establish that the mark should not have been registered in the first place.

Of the various grounds upon which the registration of a mark may be refused, Web-Sphere has asserted the following:

The mark is “devoid of any distinctive character”: Article 7 (1)(b);

The mark consists “exclusively of signs or indications which may serve, in trade, to designate the kind, quality, quantity, intended purpose, value, geographical origin or the time of production of the goods or the rendering of the service, or other characteristics of the goods or service” Article 7 (1)(c); and

The mark consists “exclusively of signs or indications which have become customary in the current language or in the bona fide and established practices of the trade”: Article 7 (1)(d).

There is clearly a large degree of overlap between Article 7 (1)(b) and Article 7 (1)(c). Questions arising under these Articles have been considered by the ECJ in a number of recent cases. First, in The Procter & Gamble Co v. Office for Harmonisation in the Internal Market[2002] Ch. 82 the ECJ considered an application to register the word “BABY-DRY” as a trademark for disposable nappies. OHIM argued that the word was not registrable because it indicated the intended function of the product; namely to keep babies dry. The ECJ disagreed. They held:

This might have been thought to have permitted registration of a mark if one of its possible meanings to an average consumer would not have regarded the word in question as being a “normal” way to refer to the function of the product in question. Not so. In Office for Harmonisation in the Internal Market v. Wm. Wrigley Jr Co[2004] ETMR 9 the ECJ considered an application to register the word “DOUBLEMINT” as a trademark for chewing gum. They held:

This seems to represent a narrowing of the test laid down by Procter & Gamble. One should not consider simply the “normal” meaning of the sign, but all its possible meanings. If any one of them is (or could be) descriptive of the goods or services, then it should not be registered as a trademark.

The last word on the subject (at least for the time being) is in the last of this trilogy of cases: Campina Melkunie BV v. Benelux-MerkenbureauCase C-265/00. This concerned an application to register the word “BIOMILD” as a trademark for various foodstuffs, principally yoghurt. The Benelux Trade Mark Office refused to register on the ground that the sign conveyed solely that the products were “biological and mild”. On a preliminary reference the ECJ held:

I draw the following conclusions from these cases.

In considering a neologism, it is necessary to consider both the individual components of the word and also the overall impression created by the word that they form;

If considered both individually and as a whole, the components or the word could be used to describe characteristics of the goods or services in relation to which registration is sought, registration should be refused, even if the words (or the components) are not actually used in that way at the time of the application;

Even if the individual components are (or could be) descriptive of the goods or services in question, the whole may not be if the overall impression created is more than the sum of its parts;

In the case of a word mark both the visual and aural impression must be considered.

Logically, this seems to me to be the place to begin, since if “WEBSPHERE” should never have been registered as a Community trade mark, that is an end to the claim for infringement. I can, I think, take Articles 7 (1)(b) and 7 (1)(c) together.

“WEBSPHERE” is not a word that exists in the English language. It is a neologism, composed of two elements “web” and “sphere”. Taken as a whole, it has no recognisable meaning. If broken down into its constituent parts, one part (“web”) undoubtedly alludes to the Internet (the world wide web), but that is as far as it goes. The other particle (“sphere”) does not, except, perhaps, that it might allude to the world. Neither particle actually describes the goods or services provided in the classes for which the mark was registered. But in any event, the whole is, in my judgment, undoubtedly more than the sum of its parts. As such, it has distinctiveness. I conclude that an objection based on Article 7 (1)(b) or 7 (1)(c) must fail.

So far as Article 7 (1)(d) is concerned, whether a word has acquired a special trade meaning is a matter of evidence. No evidence to this effect has been adduced. The objection on this ground fails too. I conclude that “WEBSPHERE” was validly registered as a Community trade mark.

The only difference between the mark and the sign is the hyphen in the sign. There is not, therefore, complete identity. But in the light of the guidance given by the ECJ, the question is whether the existence of the hyphen in the sign is so insignificant that it would go unnoticed by the average consumer.

Take, first, the aural assessment. Plainly “WEBSPHERE” and “Web-Sphere” sound exactly the same.

Take, next, the visual assessment. If the careful consumer placed the mark and the sign side by side, he would notice the presence of the hyphen in the sign. But the notional consumer does not do that. He has to rely on his imperfect memory. Even the users of the mark and the sign themselves made mistakes in this respect. On Web-Sphere’s side, the hand out distributed at Farnborough included the line:

“WebSphere” was written without the hyphen. On IBM’s side, the witness statement made by Dr West (who was in charge of the development of WebSphere products) said:

“Web-Sphere” was written with the hyphen. From the visual perspective, I regard the presence of the hyphen in the sign as an insignificant difference which would go unnoticed by the average consumer.

However, the question of identity must be assessed globally. Here Web-Sphere has a more substantial argument. The argument goes like this. The average consumer for the goods and services with which this case is concerned is a person who is a sophisticated computer user, and who uses the Internet. In accessing sites on the Internet, he types an address into his computer. What he types is a string of alphanumeric characters. The computer converts what is typed into a string of numbers. The string of numbers is then used by a router to connect the computer to the relevant web site. Each string of numbers is a Uniform (or Unique) Resource Locator (“URL”) which is, in effect, a separate address. In converting the typed input into the URL the computer treats the hyphen in exactly the same way as it would treat any other character. Thus the hyphen serves an important function, namely to connect a computer to Web-Sphere, rather than to IBM. Since the average consumer is a sophisticated computer user and a user of the Internet, he will appreciate the necessity of absolute accuracy in typing an Internet address. Consequently the hyphen is not insignificant.

I see the force of the argument, but I do not consider that it is correct. First, it assumes that the only communication of the mark and the sign is via the Internet. But IBM promotes its brand by more conventional visual means (principally by advertising in magazines). Web-Sphere itself produced a glossy leaflet to promote its intended launch of Web-Sphere software. Second, it assumes that at the point at which the consumer types the address into his computer, he has accurately remembered the full extent of the sign, including the hyphen (or alternatively has remembered that the mark does not have the hyphen). In my judgment this is only likely if the mark and the sign are compared side by side. Third, it gives no weight to the visual or aural assessment of the resemblances between the mark and the sign.

I conclude that the mark and the sign are identical for the purposes of Article 9 (1) (a).

It is admitted on the pleadings that the sign has or will be used in relation to computer software, computer software for website development, communication services by computer terminals, electronic mail, computer-aided transmission of ideas, computer programming, design, updating and maintenance of computer software and services for developing and maintaining websites.

It is also, I think, admitted on the pleadings (and, if it is not, I find) that since May 1988 IBM has used the mark in relation to a suite of software products and services which enable users to develop maintain and run websites and e-business applications.

I find that there is sufficient identity of goods and services.

It follows, therefore, that, subject to the “own name” exception, infringement has been established under Article 9 (1)(a).

If I am wrong about identity, the mark and the sign and the goods and services to which they are respectively applied are plainly similar. But in order to establish infringement on the basis of similarity, IBM must establish a likelihood of confusion.

Mr de Serville submitted that no actual confusion had been shown, and that no customer of IBM had complained about being diverted to Web-Sphere. He is factually correct, but that is not the point. The point is not whether there has been actual confusion, but whether there is a likelihood of confusion. Web-Sphere itself has asserted that there is. The leaflet distributed at the Marriott hotel asserted that IBM was “passing off”, which is an implicit assertion of confusion or the risk of confusion. That was made more explicit by the leaflet distributed at Farnborough, which said in terms that “this confusion is hurting us”.

In my judgment there is a real likelihood of confusion. It follows therefore, that, subject to the “own name” exception, infringement has been established under Article 9 (1)(b).

Neither Mr Markson nor Mr de Serville were associated with Web-Sphere before it changed its name from Publiweb Ltd. Neither could give direct evidence about the circumstances in which it did so.

Although the use of Web-Sphere (even if used in a trademark sense) is no more than the use of Web-Sphere’s own name, I do not consider that it can survive the proviso to Article 12. The reasons for this are as follows:

The name did not exist before the launch of IBM’s WebSphere products, which was itself well-publicised, or before IBM’s application to register the Community trademark;

There is no evidence that, before changing its name, Publiweb conducted any search of trademark registers;

The name is being used as a brand for software products, and Web-Sphere has not taken steps to minimise the risk of confusion;

It is difficult to avoid the inference that the name was deliberately chosen to take advantage of IBM’s reputation and goodwill.

I conclude that Web-Sphere is not entitled to rely on the “own name” exception. It follows that infringement has been established.

The circumstances in which an individual can be jointly liable with a company were considered by the Court of Appeal in MCA Records Inc. v. Charly Records[2002] FSR 401. Chadwick LJ identified a number of principles that can be derived from the case law. The relevant principles, for present purposes, may be summarised as follows:

A director will not be treated as liable with the company as a joint tortfeasor if he does no more than carry out his constitutional role in the governance of the company – that is to say by voting at board meetings.

There is no reason why a person who happens to be a director or controlling shareholder of a company should not be liable with the company as a joint tortfeasor if he is not exercising control through the constitutional organs of the company and the circumstances are such that he would be so liable if he were not a director or controlling shareholder.

Liability as a joint tortfeasor may arise where the individual intends and procures and shares a common design with the company that the tort shall take place.

Mr de Serville was the sole director of Web-Sphere. Mr Markson was not; indeed he had been disqualified from being a director or concerned in the management of a company. His period of disqualification still has a year or so to run. He was, however, described as the “sales Manager” of Web-Sphere on the leaflet distributed at the Marriott hotel, and his name appears on the leaflets. There is no evidence of control of the company being exercised through its constitutional organs. In reality, Web-Sphere did whatever Mr de Serville and Mr Markson wanted it to do. In my judgment, therefore, this is a case in which Mr de Serville, Mr Markson and Web-Sphere had a common design. The conditions of personal liability are satisfied; and I conclude that Mr de Serville and Mr Markson are jointly liable for the infringement.

The essentials of this tort were described by Glidewell LJ in Kaye v. Robertson[1991] FSR 62 as follows:

The word “malicious” is used in a number of different senses. Sometimes it is used to describe the making of a false statement that the maker knows to be false. Sometimes it is used to describe the making of a false statement where the maker of the statement does not care whether it is true or false. Sometimes it is used to describe the making of a false statement which the maker believes is true, but where he makes the statement in order to give vent to his personal spite or ill-will. Sometimes it is used to describe the making of a statement whose purpose is to injure. The various meanings are described by Lord Diplock in Horrocks v. Lowe[1975] AC 139 at 149. The same meanings apply to an action based on malicious falsehood: Spring v. Guardian Assurance plc[1993] 2 All E.R. 273. However, in considering the second sense in which the word “malicious” is used, Lord Diplock warned:

The last ingredient of the tort is either that the statements caused special damage, or that they were calculated to cause pecuniary damage to IBM. No special damage has been alleged or proved. IBM relies on section 3 (1) of the Defamation Act 1952, namely that the words were “calculated to cause pecuniary damage to the plaintiff and are published in writing.” In Ferguson v. Associated Newspapers Ltd (3 December 2001) Gray J held that, in the light of Article 10 of the European Convention on Human Rights which requires any restriction on the right of freedom of expression to be strictly justified as necessary in a democratic society, the word “calculated” should be interpreted as meaning “likely” or “probable”, in an objective sense, rather than something which might well happen or was a possibility. Although he recognised that the purpose of section 3 (1) of the Defamation Act 1952 is to relieve a claimant of having to shoulder the evidential difficulties of proving actual damage, he said that “it is not designed to enable a claimant to advance a speculative case on damage.” I propose to follow that guidance.

The statements relied on are those contained in the leaflet distributed at Farnborough, which reproduced earlier statements that had been distributed at the Marriott hotel.

In particular IBM complains of the following:

“we consider that the US and EC trademarks websphere and WEBSPHERE were improperly registered, inter alia, in that the publish to opposition procedures required by US and EC laws [were] not complied with.”

“The brand websphere is not duly and lawfully trademarked”

“[IBM] UK has been silent upon these matters since November 2000 realising that we are the lawful owners of this trade name and good faith broadcasters and holders of our various domain names.”

“Further, we hereby give notice that customers of IBM’s websphere products may be liable to pay damages if so ordered by a court.”

The first of these statements is in part an expression of opinion (“we consider that”), but also contains a factual assertion that the publish to opposition procedures had not been complied with. That factual assertion was untrue. Mr Ling had told Mr de Serville in his letter of 3 November 2000 that the procedures had been followed, although he was short on the details. The second statement is also untrue, although grounds of opposition to the registration were pleaded by Counsel who settled the Defence (and I have already considered and rejected them).

The third statement was partially true when the leaflet was first distributed in September 2001, but had become false by the time that the leaflet was redistributed in May 2002.

The fourth statement is literally true. If a court ordered IBM’s customers to pay damages, then they would have to pay. But it is a very big “if”. However, the meaning alleged (and not denied in the Defence) is that the statement meant that IBM’s customers could be made liable to pay damages if they used IBM’s WebSphere products. In my judgment this is what it means. It is not true, because IBM’s customers would incur no legal liability to pay damages merely by buying or using IBM’s WebSphere products.

These false statements about IBM were published by the distribution of the leaflets at Farnborough and by mailshot. The distribution was authorised by Mr de Serville and actually carried out by Mr Markson. Both of them, therefore, published the falsehoods.

Were they malicious? In my judgment Mr de Serville and Mr Markson did believe in the truth of the first, second and fourth statements. I do not think that they had reasonable grounds for their belief, but that does not matter. The third statement had been falsified by events since it was first written, and Mr Markson accepted that he had made a mistake in reprinting the leaflet in May 2002. But he said, and I accept, that that was carelessness.

However, in the Defence it is pleaded that:

In his oral evidence Mr Markson said that in distributing the leaflet he did not want “to over-embarrass IBM”. But I infer from that answer that he did want to embarrass IBM at least to some extent. It is perfectly true, as Mr Markson said, that IBM is a giant corporation and Web-Sphere is tiny; but as Goliath and Gulliver both discovered, giants can be injured by small opponents. In my judgment the reason why the leaflet was distributed was to sting IBM into action. The sting would have been useless unless it was designed to injure. I find that it was; and I also find that it was motivated by ill-will. In those senses, the statements were made maliciously.

The last ingredient is whether the statements were likely to cause pecuniary damage to IBM. It seems to me that the question under this head is whether customers were likely to be put off buying IBM’s WebSphere products. I bear in mind the large advertising spend that IBM committed to the promotion of its WebSphere products, and the high level of turnover in them that it had achieved by May 2002. I consider next the very limited distribution of the leaflet and, if Mr Markson and Mr de Serville will forgive me for saying so, its amateurish quality. None of the IBM witnesses were aware of any complaints from any of its customers. Although I bear in mind that the purpose of section 3 of the 1952 Act was to give the claimant a remedy despite the difficulty of proving actual loss, it does not seem to me to be likely (as opposed to possible) that any of IBM’s customers were actually diverted by the leaflet. I conclude therefore that the third ingredient of the tort is not established. It follows that the action based on malicious falsehood fails. I should add that if “calculated” meant “designed or intended irrespective of likelihood”, I would have found the third ingredient established.

The claim for infringement of trademark succeeds, and the claim for malicious falsehood fails. I am satisfied that, having found in favour of IBM on the question of trademark infringement, I have jurisdiction to require Web-Sphere Ltd to change its name, and to cease using the domain names www.web-sphere.com; www.web-sphere.net and www.web-sphere.org, or even to assign them to IBM: British Telecommunications plc v. One in a Million Ltd[1999] 1 W.L.R. 903. I will hear the parties on whether I should exercise that jurisdiction, and what other heads of relief are appropriate to give effect to my judgment.