Printed circuit boards (PCBs) are ubiquitous in electronic equipment. Typically a PCB is a non-conductive substrate on which there are metal traces carrying electrical signals. They may be on one side of the substrate or on both. In many cases it is necessary to protect the PCB against damage and against interference with the circuitry by, for instance, condensation which might cause it to short-circuit or by corrosion. This is done by a coating. However, once the PCB has been protected, it may become necessary to alter or repair the circuitry. Typically this will require the removal of the protective coating or part of it.

European Patent (UK) No EP 0 577 094 BI claims a “Protective Covering for printed circuit boards and method for their manufacture”. The novel feature of the covering, inspired by packaging of consumables, was that it contained added wax and was fused to the board by the action of heat. It claimed that the addition of wax produced certain benefits, particularly in carrying out repairs. The patent is now owned by the Respondent, Siemens Schweiz AG (“Siemens”), following a series of mergers. It claims that a coated PCB manufactured by the Appellant, Thorn Security Ltd (“Thorn”), and installed in a smoke alarm, infringes the patent. Thorn’s PCB is coated with a coating that contains 5 per cent wax. Following a trial of the action Mann J held that the patent was valid and infringed, and awarded damages or, at Siemens’ election, an account of profits.

There is no challenge to the judge’s decision on the validity of the patent. However, with the permission of Jacob and Wilson LJJ, Thorn appeal on the question whether it has been infringed (“the Infringement Issue”). Thorn also appeal, with the permission of the judge, on the question whether Siemens are entitled to recover damages in respect of a period of infringement which pre-dated their registration as proprietor of the patent (“the Assignment Issue”). We deal first with the Infringement Issue.

We can take the description of the patent in suit largely from the judge’s judgment. The patent describes the invention as follows:

It goes on to describe the existence of existing coatings, designed to protect from moisture and other environmental effects, and says that protective coatings were usually manufactured from a suitable plastic compound and created by dipping, potting, varnishing, spraying or coating, all of which have disadvantages in presenting difficulties or impossibilities of covering uniformly and in coating sharp edges and points. Coatings made in this way are either liquid or particles. The patent next describes an item of prior art (Toshiba) where a strip of thermoplastic material is applied to the front and rear sides of a PCB and bonded by fusing to form an integrated circuit card. This process has the disadvantage that the circuits are no longer accessible after sealing. The patent goes on to refer to certain plastic films made from "thermoplastic material to which a certain amount of wax is added and which has been used to package certain goods, such as consumables". It continues:

To achieve these objects, the printed circuit board with a protective coating according to the invention is characterised in that the protective coating consists of a plastic film made of electrically insulating, thermoplastic material to which a wax is added to improve its flow properties, said wax at the same time facilitating through-plating and soldering operations through the protective coating, and that the geometrical shape of the plastic film corresponds to the area of the printed circuit board to be protected, and the film is fused onto the printed circuit board by the action of heat, so that it covers the printed circuit board in a virtually air-tight and moisture proof manner. Although the claim itself does not specify any particular proportion of thermoplastic to wax, the specification says that according to a preferred embodiment the thermoplastic material will consist of between 70 and 90 per cent of the compound product and the wax will consist of between 30 and 10 per cent. The preferable percentage of wax is 30.

A product known as "Jet-Melt 3748", manufactured by 3M, is referred to as being a starting material. This was identified as being known as a hot-melt adhesive, that is to say an adhesive which is heated and melted before being used to form the bond required. Jet-Melt 3748 itself contains wax. A second component consisting of wax is then added to the Jet-Melt 3748. Lunarit 3286 is given as an example. A manufacturing process is briefly described:

At a late stage in the described process:

Claim 1 of the patent, divided into convenient integers, claims:

Although the claim begins as a product claim, it changes towards its end to a process. It is, therefore, a claim to a product produced by a particular process. The italicised parts of the claim are integers which Thorn say are not found in their coated PCB. Consequently they say that their product does not infringe. There was some debate at trial about what was meant by “through-plating”. Plating is a procedure conducted prior to soldering to ensure a wettable surface, that is to say one which provides a better surface to achieve the bond. It would be necessary to carry out this operation prior to replacing or repairing a module on the board, and can be done by applying solder or by chemical means. The judge held that “through-plating” meant carrying out plating operations through the coating. This was the construction for which Siemens contended. There is no appeal against that.

Thorn’s product, and in particular how the coating was applied to the PCB, was described in a Product and Process Description (a PPD). The judge identified the following relevant features from the PPD:

Material known as Easybond 90-795C is used. It is purchased as a standard polyolefin polymer which is known to contain 5% wax.

It has a softening point defined by the "ring and ball" test as being between 96° and 115° C.

Raw pellets of the material are melted and turned into a strip of film 300 microns thick. These are mounted on a backing.

It is then "kiss-cut" to shape.

Having been brought to the PCB manufacturing facility, the shapes are peeled from the backing "and loaded to the PCB, ensuring that the PCB contour matches that of the patch".

The PCB and patch are loaded into the bottom half of a clam-shell press incorporating a silicone moulded pressure pad designed to match the contour of the PCB.

The pressure pad is maintained at a temperature of between 50° and 80°C "to ensure that the polyolefin sealant patch material is at an optimal condition to achieve good adhesion to the printed circuit board and adopt the contour of the various SM [surface mounted] components.."

The shell is clamped shut and vacuum is applied to the closed press to prevent the formation of unwanted air bubbles.

The silicone moulded pad presses the polyolefin sealant patch into intimate contact with the underside [sic] of the PCB so that it adheres to its contour.

When the pressure differential drops sufficiently the shell is opened and the now sealed PCB is removed.

A repair process is described. During this process the polyolefin patch is partially removed by melting it down with a hot air gun. It disperses, leaving a hole down to the component(s) solder pads on the PCB. Components can be removed and new components fitted through any remaining thin layer of polyolefin material without the need for totally cleaning the area to be repaired down to the surface of the PCB. Soldering and de-soldering of components uses a soldering iron set at around 380 degrees centigrade.

The judge heard evidence from a number of experts: Prof Ehrenstein and Dr Pecht for Siemens and Mr Godfrey for Thorn. He found Dr Pecht to be a conscientious, thorough and experienced man; and in general he preferred Dr Pecht’s evidence to that of Mr Godfrey, where they conflicted. There were two additional questions of construction that he had to decide. The first was what the claim meant by “facilitating through-plating and soldering operations through the protective coating”. He held that the concept of “facilitating” was that there were practical benefits in making something easier in a way that could be appreciated by a practical user. A scientifically measurable difference without practical consequences would not qualify. Thus, so far as the question of infringement was concerned, he held that Siemens were entitled to rely on any means by which the wax in Thorn’s product made through-plating and soldering through the coating easier in a practical sense. Siemens challenge this construction by Respondents’ Notice.

The second question of construction was what the claim meant by “the film is fused onto the printed circuit board by the action of heat”. The judge construed this part of the claim as meaning that the fusion was caused by the action of heat alone. Siemens also challenge this construction by Respondents’ Notice.

It is convenient to deal first with the questions of construction. There was no dispute about the general approach to claims construction. Section 125 (1) of the Patents Act 1977 (“the 1977 Act”) provides:

Section 125 (2) provides that the Protocol on the Interpretation of Article 69 of the EPC applies to the construction of section 125 (1). That Protocol eschews a strict literal meaning of the words used in the claim; and equally eschews treating the claims as mere guidelines. Rather it adopts a position between these two extremes which combines a fair protection of the patentee with a reasonable degree of certainty for third parties.

However, the question is always what the person skilled in the art would understand the patentee to be using the language of the claim to mean. In approaching that question the skilled reader of the patent will consider the language of the claims in the light of the specification and drawings. Sometimes there is a limitation in the claim to the extent of the monopoly claimed. As Lord Hoffmann pointed out in Kirin-Amgen Inc v Hoechst Marion Roussel Ltd [2005] RPC 9 such a limitation may appear inexplicable. He continued:

Mr Miller QC, for Siemens, says that the judge was wrong to hold that the notion of “facilitating” carries with it any element of practical benefit or that the practical benefits must be capable of appreciation by a practical user. He relies on a dictionary definition of “facilitate” as meaning “to assist the progress of” or “to promote” or “to help”; and reasons from that that there is no connotation of practical benefit to a user. Since at least the decision of the House of Lords in Catnic Components Ltd v Hill & Smith Ltd [1982] RPC 183 the courts have consistently stressed that a patent specification is a practical document addressed to a readership with a practical interest in what it describes. Even if we were to adopt the dictionary definition on which Mr Miller relies, the question would still remain: is progress assisted or helped if there is no practical difference? We find it difficult to accept that a patentee would include in his claim an integer which claimed that something was “facilitated” if it had no practical effect at all. In addition the facilitation to which the claim refers is the facilitation of identified practical operations: through-plating and soldering operations. It is not merely a reference to a change in the physical properties of the coating material, which had already been done by the preceding integer (“to which wax is added to improve its flow properties”). Nor are we able to accept that the patentee was doing no more than identifying an inevitable consequence of adding wax to the coating material. Whether this was the consequence of adding wax might depend on the quantity of wax that had been added.

The soundness of the judge’s conclusion is reinforced by a consideration of the purpose of the integer. The specification acknowledges that prior art had already proposed the use of thermoplastic material to encase a PCB. It also acknowledges that prior art had already proposed plastic films made from thermoplastic material to which wax had been added. Thus if the claim had merely claimed a protective coating made from thermoplastic material to which wax had been added, it might have been vulnerable to an attack on its validity. The claim does not specify the quantity of wax contained in the coating. But as we have said the example of thermoplastic material (before the addition of wax) given in the specification itself contains wax. The patentee cannot therefore have intended to claim a thermoplastic material containing very small quantities of wax. The inventive concept claimed in the patent was that the addition of wax in sufficient quantities facilitated the practical operations of through-plating and soldering through the protective coating. Thus it was in order to distinguish the invention claimed by the patent from the prior art that the patentee added the integers claiming that the wax facilitated through-plating and soldering operations through the protective coating. In our judgment, this integer operates as a performance requirement of the claim. In other words the protective coating must contain sufficient wax in order to achieve the facilitating effect. As Mr Purvis QC, for Thorn, put it: the patentee must have been talking about something that matters. We consider that the judge was right on this point.

On the second question of construction Mr Miller says that the judge wrongly introduced a limitation into the claim that was not there on a fair construction of it. Although the claim required the film to be fused onto the board “by heat”, it did not require it to be fused by heat alone.

Mr Miller says that in the patent in suit all that the claim does is to exclude methods of fusing that do not involve heat. He says that the patent will be infringed if the fusion process in the allegedly infringing product is promoted by heating, even if the actual fusion is achieved by the application of pressure or the creation of a vacuum (the inverse of pressure). As long as heat is one of the causative factors of fusing, there will be fusion “by” heat. The judge was wrong to hold that the fusion of the coating to the board must be by heat alone.

In order to interpret the claim we must look at the specification and drawings to see what the patentee means by the claim. In so doing we must have in mind that claim 1 is a claim to a product defined (at least in part) by the process by which it is made. So in order to understand claim 1 it is necessary to look at the way that the specification describes the process. Having made the film and cut it to the desired shape, the film is heated to a temperature above the softening point but without melting it (process step D). It is then pressed over the electronic components (process step E). At this stage, therefore, the film has been heated and had pressure applied to it. The specification continues by saying that in a “final process step” the coating is fused to the board by further heating (process step F). Thus it is not until process step F, which consists only of further heating, that the coating is fused to the board. Having regard to the way in which the process is described in the specification we consider that the skilled reader would conclude that the fusion takes place by the action of heat alone. In our judgment, therefore, the judge was also right in his construction of this integer of the claim.

In Biogen Inc v Medeva plc [1997] RPC 1 Lord Hoffmann said:

Clarke LJ summarised the position in Assicurazioni Generali SpA v Arab Insurance Group (BSC) [2003] 1 WLR 577:

This passage was approved by the House of Lords in Datec Electronics Holdings Ltd v United Parcels Service Ltd [2007] 1 W.L.R. 1325. These principles have been applied in subsequent decisions of this court: see Meretz Investments NV v ACP Ltd [2008] 2 W.L.R. 904; Henwood v Barlow Clowes International Ltd [2008] EWCA Civ 577. Mr Miller also relied on a number of cases relating to the circumstances in which an appellate court will interfere with a judge’s exercise of discretion. Those cases are not in point. As Lord Mance said in Datec (in a speech with which the majority agreed):

Lord Mance also commented that the approach of an appellate court to the exercise of discretion or an evaluation of judgment analogous to the exercise of a discretion is a “steeper appellate hurdle”. The present appeal is not a case of discretion. Nor is it a case where the judge was applying a legal standard such as reasonable care or obviousness. This is a case in which the judge had to reach factual technical conclusions based on the evidence that he heard. The challenged conclusions to which the judge came concern narrow questions of technical fact; and we have been taken by counsel to all the evidence that bears on those conclusions.

The question on this appeal is whether, giving full weight to the advantages enjoyed by the judge who has heard the oral evidence, his finding of fact has been shown to be wrong.

We begin with the first of the two disputed integers.

The judge held that the presence of the wax improved the flow properties of the plastic film used by Thorn at the temperatures at which Thorn carried out its process, and at higher temperatures at which the initial experiments took place. He also found that wax in the film would improve the flow properties of the film generally (para 51). He found, further, that film containing wax would have a lower softening point, reduced viscosity, increased wetting capability and lower surface tension (para 67). It is important to note at this point that a reduction in viscosity means that the film containing wax is runnier than a film without wax; and that an increased wetting capability means that the film with wax is more adhesive (or stickier) than a film without wax.

The judge then had to decide whether the wax facilitated through-plating and soldering operations through the protective coating. Siemens’ case was that repairs on the board are made easier because the wax makes the plastic softer, makes it melt (in the sense of softening at a lower temperature than would otherwise be the case), improves its flow properties and improves its wetting abilities, with the effect that if one carries out a repair activity on a coated board or component, the coating is more easily penetrated, and flows back more readily to the disturbed area. This is particularly so where the repair activity involves a soldering iron. Thus part of Siemens’ case was that the coating containing wax had improved wetting capabilities (i.e. was stickier than its waxless counterpart). In paragraph 56 of his judgment the judge said that Siemens were entitled to rely on the following features (if demonstrated) as facilitating through-plating:

Speed of penetration of a soldering iron;

The easier flowing away of the coating from the site of the repair; and

The flowing back of the coating once the repair was complete.

He then turned to consider whether these practical benefits had been demonstrated on the evidence. One of the main planks of Siemens’ case was an experiment called “Experiment 3”. This was designed to show that even after penetration by a soldering iron the coating in Thorn’s product (containing 5 per cent wax) continued to protect the underlying PCB, either because less of it had been removed during the repair, or because it had flowed back after the penetration. The judge was unimpressed with Experiment 3. He held that few useful conclusions could be drawn from it; and that the experiment did not establish either that Thorn’s product parted more easily than a waxless product; or that, having been penetrated, the coating flowed back and thus healed the penetration (paragraph 65). However, Experiment 3 was not the only plank in Siemens’ case. Siemens also relied on its other experiments and on the inherent improvement in the flow properties of the coating which is brought about by the presence of wax (paragraph 52; paragraph 66).

Before coming to his conclusions the judge commented on the state of the evidence generally. He said in paragraph 70 of his judgment:

This was not a promising start to a finding of infringement. The judge’s conclusions on the first of the disputed integers are contained in paragraph 71 of his judgment:

The reference to “through-contacting” as opposed to “through-plating” at the beginning of this paragraph seems to have been no more than a slip of the keyboard; and neither side made anything of it. In paragraph 72 the judge said:

The foundation of the judge’s ultimate conclusion on the presence of this integer in Thorn’s product is his finding in paragraph 72 (iii) of the judgment. Thorn complain that the judge’s finding was not supported by the evidence and was, in some respects directly contrary to it. It is important to note that as well as the specific references to particular experts, the judge also relied on the expert evidence about the general effects of introducing wax into a coating, and the results of experiments 1, 2, 4 and 5.

The judge expressed this finding as follows:

Embedded in the sentence are two sub-findings:

The presence of wax means that melting occurs more easily; and

Because melting occurs more easily charring or similar damage to the film is less likely at the high temperature of a normal soldering iron.

It is important to note that the judge’s finding related to charring or similar damage to the coating; not charring or damage to the board.

The judge dealt with the general characteristics of a coating to which wax had been added in paragraphs 66 and 67 of his judgment in the following terms:

This discussion was a discussion at a theoretical level. Siemens’ case was that the addition of wax changed the properties of the coating. The judge accepted that in paragraph 67 of his judgment as a theoretical proposition. Siemens’ case was that the mere fact that the properties of the coating had been altered by the introduction of wax led to the conclusion that it would facilitate through-plating and soldering. Consistently with his construction of the claim, the judge then had to decide whether this alteration in the properties of the coating led to the claimed benefits in a practical sense. The judge subsequently said that “flowing back” had not been established except in the case of penetration of the coating by a very thin probe. But penetration by a very thin probe (probably for the purpose of measurement) as opposed to through-plating or soldering was not part of the claim. So that finding can be put on one side.

The judge had already accepted Mr Godfrey’s evidence that a normal soldering iron operated at something over 300°C. He had also accepted Mr Godfrey’s evidence that in the case of penetration of the coating by a soldering iron at this temperature it made no practical difference to ease of penetration whether the coating contained wax or not. The evidence of Experiments 4, 5 and 6 had shown that Thorn’s coating (with wax) had a lower melting point than the equivalent coating without wax. In the former case the wax began to melt at 12^o C and the bulk began to melt at about 48^o C. According to Prof Ehrenstein at 86^o C the crystalline parts of the thermoplastic material were melting. In the latter case the material began to melt at about 80^o C. Again according to Prof Ehrenstein the maximum of the melting process took place at 97^o C. Thus the proven difference in melting point between the waxless coating and the coating with wax was 11^o C.

It was the temperature gradient between a normal soldering iron and the coating (with or without wax) that had led the judge to conclude that there was no practical difference in ease of penetration. We are unable to see how, in the light of that finding which seems to us to be entirely correct, the judge could have concluded that the a difference in the melting point of 11^o C meant that the coating with wax melted more easily in a practical sense when in contact with a soldering iron operating at a normal temperature. In our judgment the two findings cannot co-exist.

Mr Purvis submits, and we agree, that in order to conclude that the greater ease of melting reduced the likelihood of burning or charring, it would have to be established:

That in the absence of wax burning or charring was a real possibility and

That the addition of wax materially reduced that risk.

The principal basis for the judge’s finding on this point was the evidence of Prof Ehrenstein. Although the judge does not specifically say in that paragraph which part of Prof Ehrenstein’s evidence he was referring to, it is common ground that he must have been referring to paragraph 38 of Prof Ehrenstein’s first report. The judge quoted part of that paragraph in paragraph 63 (c) of his own judgment:

Making all allowances for the fact that English is not Prof Ehrenstein’s native tongue we find this sentence extremely obscure. It may mean (as Siemens submitted in opening the trial below) that the wax helps avoid burning and other degradation of the coating material. However, in commenting on this passage the judge himself said that Prof Ehrenstein was operating at a theoretical level, because no comparative tests had been carried out on a waxless coating. Moreover, as Mr Purvis pointed out, Prof Ehrenstein did not say that it was the 5 per cent wax in the Thorn coating which prevented “burning and other degradations”; and did not say that the heating of the coating would, in any practical sense, have been less rapid in a waxless coating. Nor was the point put to Mr Godfrey in cross-examination.

In addition, none of the experts had mentioned “charring” of the coating as a hazard to be avoided.

In the light of the judge’s construction of the patent (which, as we have said, we consider to have been correct) we do not, with respect, agree with him that the evidence supported the finding that this possible theoretical benefit of a coating with wax had been proven to be a practical benefit present in Thorn’s coating.

The judge expressed this finding as follows:

Although the judge did not explicitly say that the noticeable difference he had in mind was a difference in ease of penetration of the coating, it is common ground that that is what he meant. The principal basis for this finding was the evidence of Mr Godfrey. This claimed advantage based on the use of a lower temperature soldering iron was not a claimed advantage that had featured in the way that Siemens had opened its case. Neither of the experts called by Siemens suggested that the possibility of repair using low temperature soldering irons was a real benefit of the invention. So this finding was based entirely on answers extracted from Mr Godfrey in cross-examination. The judge had accepted Mr Godfrey’s evidence that there was no material practical difference between the waxed product and the waxless product in relation to a soldering iron at a normal temperature (paragraph 71 (ii)). What, then did Mr Godfrey say about lower temperature soldering irons? His first reference in cross-examination was to repair. He said that in order to repair a PCB you would remove the coating “by using heat, a hot air gun or thermostatically-controlled soldering iron, so you can get at the faulty component.” (Day 5 p. 143). However, this part of his evidence did not refer to the temperature at which the hot air gun or thermostatically controlled soldering iron might operate. It does not, therefore, bear on the question whether a lower temperature soldering iron would be used. On p. 170 there was the following exchange:

A few minutes later there was the following exchange:

Mr Godfrey repeated his evidence that at normal soldering iron temperatures there was no practical difference between a waxless coating and a coating containing wax. But he did refer to a lower temperature soldering iron in the following exchange:

The question for the judge was whether the claimed practical benefits of the patent were present in Thorn’s product. It seems to us, therefore, that unless there was a practical possibility of a lower temperature soldering iron being used to carry out soldering operations on the Thorn product, no practical benefit would have been demonstrated. Mr Purvis complains that in so far as Mr Godfrey said anything about lower temperature soldering irons, it was purely hypothetical. It has not been shown that lower temperature soldering irons exist in the real world. Mr Purvis further argues that it was not shown that the solder used in the Thorn product was low melting point solder. That is true. He says that without such evidence (and a concomitant finding of fact by the judge) infringement has not been proved. We agree. Mr Godfrey’s evidence, which we have quoted, was a theoretical and generalised discussion; not targeted at Thorn’s product containing no more than 5 per cent wax, and with a melting point of less than 100^o C. Mr Godfrey’s evidence was that whether there was any practical benefit was a question of degree; but whether the requisite degree existed in Thorn’s product was not explored further. Whether or not lower temperature soldering irons exist in the real world (and we have no reason to doubt that they do) the question was whether the existence of a lower temperature soldering iron meant that the ease of penetration of the coating by such a soldering iron was a practical benefit that existed in the Thorn product. In our judgment Mr Godfrey’s evidence does not support the conclusion that it does. Indeed the description of a repair of the Thorn product in the PPD specifically states that the soldering iron is set at 380^o C. There was no challenge to the PPD in this respect. Mr Miller argued that a repair might be carried out using a low temperature soldering iron to penetrate the coating, before switching to a normal temperature soldering iron to carry out the repair itself. But this, too, is in our judgment, a purely theoretical scenario, which would (at least) have had to have been put to the experts before it could have led to a finding of a practical benefit. Mr Miller also argued that the PPD described removing the coating with a hot air gun; and that this might support the judge’s finding. However, the judge’s finding specifically related to a lower temperature soldering iron (rather than a hot air gun); and in any event although a hot air gun was mentioned, there was no evidence about the temperature at which a hot air gun would operate.

We have been persuaded by Mr Purvis that this finding cannot stand.

The judge expressed the third finding as follows:

The judge said that this finding was not directly addressed in the evidence but was a sensible inference to draw from it.

Mr Miller argues that the judge may have underestimated the extent of the evidence on this point; and in particular the evidence of Mr Godfrey in cross-examination. We have already quoted part of the exchange, but it is convenient to repeat a bit more of it:

The judge did not rely on this exchange in coming to his conclusion on the third of the findings. In fact reading more of Mr Godfrey’s cross-examination immediately following the quoted passage, it is by no means clear precisely what he was saying. On one entirely plausible view of his answers taken as a whole, he was saying that the presence of wax in the coating made it more difficult to remove the coating. This is an instance where the judge had the advantage of hearing the evidence, and assessing the import of Mr Godfrey’s answers on this point taken as a whole. We do not consider that we should interfere with the judge’s conclusion that there was no direct evidence on the point. The question is, therefore, was the judge wrong in drawing the inference he did?

Mr Purvis complains that this inference drawn by the judge was actually contradicted by other findings of fact that he made. The judge had found, at a theoretical level, that the addition of wax to the coating both reduced its viscosity and reduced its surface tension thus increasing it wettability. The former characteristic made it runnier and the latter made it stickier. The extent to which these two countervailing characteristics of the coating with wax balanced each other would have been a matter for expert evidence; but the experts did not address the question.

The drawing of inferences is, of course, a familiar technique in judicial decision making. It enables a judge to conclude that, on the basis of proven facts A and B, a third fact, C, was more probable than not. In drawing inferences in this way the judge applies his or her knowledge as a person of ordinary awareness of the way that people behave, or the way that things work. However, once the subject-matter of the inference is a matter of technical or scientific evidence, for which the evidence of experts is necessary, a judge should be cautious before drawing any inference that the experts themselves have neither drawn nor commented on. It is precisely because the subject-matter is outside the ordinary experience of the tribunal that expert evidence is necessary in the first place. So ordinary experience is unlikely to be a reliable guide to the drawing of technical or scientific inferences. It may well be that a judge is entitled to draw a necessary inference from scientific or technical evidence. But the judge in the present case did not regard the inference he drew as a necessary one. He regarded it as no more than a sensible one.

In our judgment the inference that the judge drew was not, with respect, one that he should have drawn. There were two characteristics of a coating with wax that pointed in opposite directions: lower viscosity and increased wettability. What overall effect those competing factors had upon the coating was, in our judgment, a question for direct expert evidence and was not a matter upon which an inference could safely have been drawn.

Accordingly, in our judgment, Siemens failed to prove that this integer was present in Thorn’s product.

The second disputed integer was that part of the claim which said that the film:

Based on his construction of this part of the claim, the judge went on to consider whether Thorn’s product contained this integer. He held that it did. His conclusion was not based on evaluating the evidence that he heard; but was based on his interpretation of the PPD. The relevant parts of the PPD are:

The judge expressed his conclusion as follows:

The critical finding in that paragraph was that Siemens had not proved that Thorn’s product:

Thus the judge found that Siemens had not proved that the coating was fused onto the board “by the action of heat”. Why, then, did he hold that this integer was present in Thorn’s product? Direct evidence having failed, again it was a question of the judge’s inferences from the general properties of the product. He supplied the gap by reference to the PPD. This in itself was not objectionable. He reasoned as follows:

However, the quotation from paragraph (i) of the PPD on which the judge based his conclusion is not a full quotation of that paragraph. Selectively quoted it gives the impression that the heating of the sealant patch material is done in order to enable it to adhere to the PCB without more. But that is not what the PPD says. The heating is applied in order to ensure that the sealant patch material is in optimal condition to adhere to the PCB. In other words the heating produces a condition; not the adhesion or fusion itself. The adhesion itself is produced by the subsequent step of pressure “so that it adheres” to the contour of the PCB. The purpose of the pressure is to cause the adhesion. As the judge correctly pointed out, claim 10 of the patent in suit describes a process in which the softened film is pressed onto the modules; but as he also correctly pointed out, even in that process the final adhesion is caused by a second heating. When the judge said that it was necessary to put the process together in the “proper sequence” it seems to us that what he was in fact doing was altering the process or sequence that was described in the PPD; and thus attributing to the function of heating described in the PPD a consequence or effect that it did not have. Although the judge said that it was the heat “that does the final job”, it was in fact, according to the PPD, the application of pressure that did the final job. The fusion thus took place by the application of pressure; not by the action of heat. Without the application of pressure, on the judge’s own findings, there would have been no adhesion and hence no fusion. It may well be that the fact that Thorn’s product does not fuse by the action of heat is attributable to the fact that its wax content is only 5 per cent, whereas the patent specification envisages a range of 10 to 30 per cent wax, with a preference for 30 per cent. However, we need not speculate about the reason. For the reasons given, we have been persuaded by Mr Purvis that on the judge’s construction of the patent this integer was not proved to be present in Thorn’s product either.

It follows in our judgment that Thorn’s product did not infringe claim 1. In case, however, a higher court is of a different view on the Infringement Issue, we shall deal with the Assignment Issue. It was fully argued before us and it arose on the judge’s finding of infringement. It is, moreover, a point which has not previously been considered by this court.

We now turn to the Assignment Issue.

The application for the patent was made in the name of Cerberus AG, a company incorporated under Swiss law. On 30 January 2002, Cerberus AG filed with the Patent Office a form showing that the new owner of the patent was Siemens Building Technologies AG, which was also incorporated under Swiss law. On 11 March 2005, Siemens Building Technologies AG gave notice to the Patent Office that that company had been merged with the claimant in the action and the respondent to this appeal, Siemens Schwiez AG.

The alleged infringements occurred from 1998 onwards. The question that the judge had to consider was whether Siemens Schweiz could sue for infringements before 25 February 2002. That depended on the effect of section 68 of the 1977 Act which, as originally enacted, provided:

The words in italics have now been replaced by the words “in proceedings for such an infringement, the court or comptroller shall not award him costs or expenses” (Intellectual Property (Enforcement etc) Regulation 2006), which were designed to implement art 13 of the Enforcement Directive 2004/48/EC. It is not suggested that anything on this appeal turns on this amendment. The 1977 Act does not define the term "proprietor", but, by inference he is the person who owns the patent.

Section 68 refers back to section 33. That section deals with priorities between various holders of patents. S 33 provides:

Where an application for the registration of a transaction, instrument or event has been made, but the transaction, instrument or event has not been registered, then, for the purposes of subsection (1)(a) above, registration of the application shall be treated as registration of the transaction, instrument or event.”

Not all of “the transactions, instruments or events”, described in s 33(3), lead to a person becoming the proprietor or an exclusive licensee of a patent. For example, a person may not become a proprietor by becoming the person in whose favour security is granted. The incorporation of s33(3), by reference into s 68 is, therefore, imperfect.

We need also to set out s 30 of the 1977 Act, which played a crucial part in the judge’s reasoning. S 30 provides:

The judge described the mergers which had taken place in the Siemens group at [81] and [91] to [97] of his judgment. It is not necessary for us to repeat the details. The material elements of the transactions were that on the occasion of each merger there were the following steps: (1) the merging parties entered into an agreement to transfer all the assets and liabilities of the merging company to the acquiring company in accordance with art 748 of the Swiss Code of Obligations Governing Corporations (“the Swiss Corporations Code”); (2) that agreement was approved at general meetings of (at least) the merging company; (3) the relevant documents were then filed with the Swiss Commercial Registry; and (4) upon registration of the filing of the relevant documentation, the assets and liabilities were as a matter of Swiss law vested in the acquiring company. It then manages the assets of the merging company. (It is not clear on the evidence whether the beneficial ownership passed to the receiving company immediately on registration of the documentation, or whether it only did so when the creditors of the merger company had been paid or provided for. But that point goes only to timing and is of no consequence in this case.) The essential point is that in the words of Professor von der Crone, the expert on Swiss law whose evidence the judge accepted:

Professor von der Crone explained in cross-examination that the doctrine of universal succession had its origins in the law of inheritance under which the assets and liabilities of the deceased vest in the appropriate person by operation of law. The interests of the creditors are not protected by an order of the court approving the merger (as would be the case under s 900 of the Companies Act 2006) but the members of the board of directors of the acquiring company are personally liable for the separate administration of the assets of the merging company (art 748 of the Swiss Corporations Code).

If the required documents are filed with the registrar under art 748 of the Swiss Corporations Code, the registrar cannot refuse to make the relevant entries on the commercial register. He only exercises a judgment relating to “whether the formalities are correct (or whether all the required information is provided, or whether the applicant has the necessary authority).” Article 748 is not a provision which deals with mergers, but with dissolution, and it has been adapted by interpretation for use in connection with mergers.

Most takeovers in the United Kingdom are by share exchange and so this form of reorganisation is not well known in the United Kingdom. It is, however, very different from a sale of the undertaking and assets of the company. That would not be effective to transfer liabilities, nor would it involve the dissolution of the merging company. Moreover the parties can chose the proper law of a contract for the sale of assets. By contrast, a merger can only take place under the law of the place of incorporation. It involves a change of the status of the merging company: see generally National Bank of Greece and Athens SA v Metliss [1958] AC 509. In this case, there were no appraisal or other rights given to the shareholders because the merging company was a wholly-owned subsidiary of the acquiring company.

The judge rejected an argument by Siemens that the amendment to s 68 was retrospective ([84] to [90]). There is no appeal against his ruling on that point. When he came to consider whether the merger constituted an “assignment” within s 33(3), he proceeded by a series of strictly logical, stand alone steps. He held that the merger agreement did not assign any patent. (He accepted as a matter of fact that this was so notwithstanding a decision of the Swiss Supreme Court, as Professor von der Crone considered that that case was wrongly decided). He held that the merger agreement was analogous to an agreement to transfer the property which was in due course completed by the example a conveyance. He held, following his earlier decision in Tamglass Ltd OY v Lyoyang Glass Technology Co Ltd [2006] EWHC 655 (Ch), that an assignment had to be an express consensual bilateral document transferring a patent. He held that it was not enough that, under the Swiss doctrine of universal succession, a merger agreement was essential. He held that, if the word “assignment” had the wider functional meaning contended for by Thorn, it would include the vesting of property on death or pursuant to a court order. An assignment had to take place under the assignment, not pursuant to it.

The judge also rejected the argument that s 33(3)(e) applied. He held:

The judge's reasoning incorporated his reasoning in Tamglass. The ratio of that case on s 33(3)(a) is that it followed from s 30(3)and (6) of the 1977 Act that there had to be a consensual document in the nature of a consensual bilateral transaction. He also rejected an argument based on s 33(3)(e). On the evidence of Finnish law in that case, the registration authority had to give permission for implementation of the merger but the merger itself resulted merely from notifying that authority of the implementation of the merger within four months after the grant of permission.

The judge recognised that the result was anomalous. The purpose of s 68 is to ensure that “the people who own the monopolies get on the register” per Jacob J in LG Electronics v NCR Financial Solutions Group Ltd [2003] FSR 24 at [18]; see also Molnycke v Proctor & Gamble [1994] RPC 49 at 109 per Morritt J. The effect of the judge’s ruling was that a well-recognised event in corporate terms fell outside the protection intended to be afforded by s 68. His judgment is described in the grounds of appeal as having “opened up a large lacuna in which the purpose of [s 68] is defeated”.

On this appeal, Mr Purvis submits that the judge should have given s 33(3) a more purposive interpretation. He relies on the following passage from the Report of the Committee to Examine the Patent System and Patent Law chaired by M A L Banks (1970) (Cmnd. 4407):

This recommendation applies to any acquisition of ownership, and is not limited to particular kinds of change of ownership. Mr Purvis submits that there is no obvious policy reason for leaving out assignments of the kind in question here. He submits that there is no requirement that the property passes on the date of the signed agreement. There is no policy reason why assignments which are conditional on some further event occurring should be treated differently under s 33 from unconditional assignments. Mr Purvis criticised the judge’s reasoning as over-analytical. In the case of each merger, the transaction viewed as a whole was clearly an assignment. The judge’s reasoning also results in a narrowing of s 33 itself. Mr Purvis also relies on s 33(3)(e),

Regarding s.33(3)(a) Mr Miller relies on Tamglass. He further submits that a distinction is to be drawn between an agreement to assign and an assignment. The former is not an assignment: see per Jacob J in Coflexip Stena Offshore Limited’s Patent [1997] RPC 179. To become an assignment for these purposes, a separate legal document assigning the patent would be required, as in Casey’s Patents, Stewart v Casey (1892) 9 RPC 9. A mere agreement would not be within s 33(3), and if an agreement were conditional there might be difficulties meeting the six months’ time limit in s 68. In our judgment, Mr Miller’s argument here is not a sound one: s 68 (b) of the 1977 Act caters for the situation where the six months’ time limit is exceeded for some good reason, and the assignments may be conditional just as agreements may be conditional. Mr Miller further submits that any person who searched for the owner of the patent in question would discover that the mergers had taken place. (This would be possible by the simple act of searching the Swiss Commercial Register). Hence, the policy of being able to determine the true owner of the patent has been complied with.

Mr Miller submits that s 30(6) lays down an exclusive method of transferring patents and that an assignment therefore cannot exist which falls outside that provision. Furthermore, on his submission, the only form of devolution by operation of law that is covered by s 33(3) is that occurring on death: see s 33(3)(d).

As to s 33(3)(e), Mr Miller submits that there is no official direction or discretion of the kind to which to which s.33(3)(e) applies.

We turn to our conclusions. The term “assignment” is inherently capable of more than one meaning. It may refer to the passing of ownership or, where parties may only transfer property by executing an instrument in a particular form it can mean that instrument as executed in relation to particular property: see, by analogy, Lyle & Scott v Scott’s Trustees [1959] AC 763. This case deals with the different meanings of “transfer” in relation to the articles of association of a company and shares held on a share register. "The context must determine in what sense the word is used.”: per Lord Reid at 778. So too the word “assignment” may refer to the instrument of transfer of a patent, which must follow the minimal requirements of s 30(6), or the vesting of an absolute interest in it in another person. The judge was right to look at other uses of the term “assignment” in the 1977 Act, but where a statutory provision is intended to meet a specific mischief or has a different history from another provision, statute may not use the same term consistently (see, for example, Giles v Rhind (No 3) [2008] 1 BCLC 1 1103 at [50], Gingi v Secretary of State for Works and Pensions [2002] 1 CMLR 587).

A striking feature in the present case is the recommendation of the Banks Committee, set out above. The interpretation of a statute is a deductive exercise based on close examination of the precise words that Parliament has used, but the process still requires a cross-check to be made of the result of a literal wording against any internal or (where admissible) external clues as to the effect of the provision. The judge’s reasoning as a pure exercise in deductive logic cannot be faulted, but in our view he failed to cross-check his answer on a literal approach against those internal and external clues. Those clues have led us down a different trail. The clues all go in the direction of warning against a narrow literal approach.

The first clue is of course the Banks Committee report. That report, which led ultimately to the enactment of the 1977 Act, is admissible on a question of statutory interpretation in relation to any implemented recommendation to show the mischief to which s 68 is directed. It is self-evident that Parliament in adopting the Banks Committee’s recommendation set out above intended s 68 to be, so far as possible, all-embracing and not to have obvious lacunae. The strong recommendation of the Banks Committee which led to s 68 mandates a purposive reading of s 33(3) to give effect to the mischief it identified, if that can be achieved.

There are also significant internal clues. The opening words of s 33(3) are “transactions, instruments or events”. These individual components of this phrase are wide and open-textured and so the combination of them in the opening words can be taken as an indication that, so far as possible, s 33(3) is intended to be comprehensive. There is no reason why an assignment, if given a wide meaning, should not be an instrument in some circumstances and an event in other circumstances, or a combination of an event and an instrument in other circumstances. The opening words of s 33(3) contrast with the words used in, for example, s 30(6), which are limited to assignments, mortgages and assents.

A further interpretative clue is found by considering the effect of the judge’s interpretation on s 33 itself. S 33 addresses the problem of priorities when a person acquires an interest in a patent when a prior interest has been created which has not been registered and of which he is unaware. Section 33 lays down rules for that purpose. There would be no purpose in limiting those rules to assignments which qualified as consensual bilateral documents. Like s 68, s 33 should have as wide a field of operation as is consistent with the words used.

Another interpretative clue of a comparative minor nature is to be found in s 40 of the 1977 Act. This refers to contracts of assignment. This suggests that there is a narrower sub-group of assignments which constitute contracts of assignments.

The judge was strongly influenced by the references to assignment in s 30(6). But this subsection has to be read closely. At first blush, it suggests that an assignment can only be by way of bilateral consensual agreement, but it is indisputable that assignments can also be by operation of law. The facts of this case demonstrate that. Moreover, the references to assignment in s 30(6) are not in fact limited to bilateral consensual documents. They can be deeds poll, as where a person decides to give a patent to another person, or vest it in him in the administration of a deceased’s estate. The other person need not be a party to that document or assent. The word “assignment” in s 30(6)(a) has, therefore, to be read as referring only to those forms of assignment which can take effect by a document complying with the formalities in s 30(6). S 30(6) is accordingly not the silver bullet mandating a narrow interpretation of s 33 (3). It would, moreover, make a nonsense of ss 33 and 68 if the disponor under an assignment taking effect by operation of law could delay the operation of those sections by claiming that those sections operated only when a document falling within s 30(6) had been executed reflecting the effect of an assignment by operation of law which had already taken place. Once the meaning of "assignment" is liberated from its meaning in section 30 (6)(a) and the link with s 33(3) which the judge’s reasoning imposed, there is no reason why the word "assignment" should not receive a wide meaning in s 33(3). In our judgment it is wide enough to include an assignment by operation of law, certainly an assignment by operation of law of the kind involved in universal succession following a merger. There is nothing in s 30(3) to indicate that this conclusion is not correct: on the contrary, it would be odd if the true interpretation is that sub-section refers expressly to the vesting of patents by operation of law (it would appear, by implication, by any means), but deals with such vesting in the remainder of that sub-section and in s 33(3) only when it occurs on the death of an individual. Yet that is the result to which the judge’s interpretation would lead.

It is not necessary for us to deal with all the circumstances in which a patent can be assigned by operation of law. If the proprietor of a patent is an individual who becomes bankrupt under English law, his assets will vest in a trustee in bankruptcy by virtue of s 306 of the Insolvency Act 1986. If a corporation is dissolved its assets will vest by operation of law in the Crown (s 654 of the Companies Act 1985, which is to be repealed and replaced by s 1012 of the Companies Act 2006). In these situations, the position may be that there is no assignment for the purposes of s 33(3) until the trustee in bankruptcy, or the Crown (as the case may be), assigns the assets to third party. But if that were the position, it would be analogous with the position on death under s 33(3)(d). Those questions will have to wait until another day. To take up a separate point made by the judge, we do not consider that s 33(3)(d) stands in the way of the interpretation which we have adopted since its purpose may be to postpone the need to register a change of ownership until the period of administration of the estate has been completed. No disposition of the patent can be made until that date.

The transfer of assets in this case was effectively a return of assets in liquidation. The merger effects a change in the status of the merging company but that does not mean that the merger cannot involve some other legal step, such as a return of assets: see, generally, Sun Alliance Insurance Ltd v IRC [1972] Ch 133. No argument has been addressed to the word “assignation” in s 33(3)(a). This word does not appear in s 30. In our view, the word “assignation” does not stand in the way of the interpretation we consider to be correct. Nothing in this judgment undermines the conclusion in Coflexip that a mere agreement to transfer a patent, which is not (or not yet) followed by a formal assignment of itself does not fall within s 33(3). We accept the argument that to reach a different conclusion on that point might lead to uncertainty in practice.

As to s 33(3)(e), we do not need to decide the meaning of this provision and need only express the following provisional views. The registrar of the Swiss Commercial Register is clearly an official who, dependent on his powers, is capable of being described as a competent authority. However, the fact that the registrar has to decide whether the documents that have been filed are adequate or not is not a direction which results in the transfer of a patent taking place. The transfer takes place as a result of the Swiss Corporations Code. Furthermore, the registrar’s act is of a ministerial nature. As we see it, section 33 (3)(e) was intended to cover the case where the order or directions are dispositive and involve the exercise of some discretion conferred by law. S 33(3)(e) would on this basis include a vesting order made by a court in favour of a shareholder or creditor of a dissolved company following the dissolution of a company, which had caused its assets to be vested in the Crown as bona vacantia (see, for example, Re Strathblaine’s Estates Ltd [1948] Ch 228). There is no duplication between s 33(3)(a), as we have interpreted it, and s 33(3)(e) because the latter refers to a specific type of assignment which is not one which is merely by operation of law but involves, as we have explained, the exercise of some judgment on the part of the court or competent authority.

For the reasons given above, in our judgment the judge was in error and his decision on s 33(3) in this case must be set aside. We also overrule his decision on s 33(3) in Tamglass.

For the above reasons we allow the appeal on the Infringement Issue and the Assignment Issue and dismiss the respondent’s notice.