Environmental Technologies Inc (EPI) & Anor v Symphony Plastic Technologies Plc & Anor

This is an appeal, with my permission, from a judgment of Peter Smith J given on 21^st December 2004. The case concerns an alleged breach of rights in trade secrets. With the parties’ consent the Judge held the whole hearing in private – it being not only impractical to pop in and out of private during the course of the argument but also such a course might have led to the inadvertent mention of an alleged trade secret in open court. For the same reason we also sat in private.

The claimants (collectively “EPI”) are two US companies. The defendants (collectively “Symphony”) are English companies. Nothing turns on the fact that on both sides there are two corporate entities. By a succession of agreements and supplementary agreements, EPI licensed Symphony to use its know-how and trade marks. EPI also agreed to supply Symphony with additives to be used in the manufacture of a range of thin film plastic material. The most important of these in practice was an additive designated by EPI as DCP509. It was a term of the agreements that Symphony were not to analyse the additives.

The Judge gave his judgment in two portions – an open portion ([2004] EWHC 2945 (Ch), [2005] FSR 502) dealing with the law and some general, non-secret, factual matters, and a longer, confidential, portion dealing with the facts in detail. When we came to read the papers it became clear that the Judge’s finding that Symphony had not analysed DCP509 and had made no use of any information derived from EPI was critical. Unless that finding of fact could be displaced, the main claims, for breach of contract and breach of confidence, were bound to fail. Accordingly we invited the parties to deal with this question first. Because it involves alleged trade secrets, it follows that this judgment in its full form should also remain confidential.

This judgment in its full form is for the parties’ eyes only. By the time of its publication portions will have been redacted by the parties so as to exclude alleged trade secrets. Since that will reduce its intelligibility to the general reader and because no question of general interest arises I see no point in setting out background explanatory matter. It is sufficient to delve straight into the points argued, pausing only to remind myself of the requisite standard for appeal on a question of fact.

That standard is set by CPR Part 52.11(1) and (3). So far as is relevant here the appeal is by way of review only and it is essential for the appellant to show that the Judge was wrong. It is not for us to assess the mass of evidence on the transcript to make our own findings – it is the Judge, not us, who saw and heard the witnesses and immersed himself in the evidence for many days. Only if he has clearly made a material mistake (which would include accepting a wholly improbable story) is the judgment vulnerable here. Mr Hobbs acknowledged this in principle, but to my mind, at many points in his argument he was in substance just asking us to re-try the issue on the transcript.

Mr Hobbs submitted at the outset that the burden of proof of independent derivation of Symphony’s product, BD92384, lay on Symphony. He next submitted that Symphony had not satisfied that burden and accordingly the Judge was wrong to hold that independent derivation was established.

First then, was Mr Hobbs right about the burden of proof? I think not. He relied upon the following matters as shifting the evidential burden of proof (the legal burden of proof clearly and admittedly lying on EPI):

Mr Hobbs sought by analogy to say that the reasoning in some copyright cases applied here. That reasoning is not so much a rule of law as one of rational weighing of evidence. It is simply this: if the degree of similarity between the work alleged to have been copied and the alleged piratical work is unlikely to have come about by coincidence, it is for the defendant to prove that it was. This is of course no more than an example of a shifting evidential burden.

Here only the first of Mr Hobbs’ points relates to a comparison of the two “works”. I do not think a comparison of them remotely raises an inference of copying of one from the other. The formulae are set out at [REDACTED] of the confidential judgment. [REDACTED].

Technically the [REDACTED] most important ingredients are [REDACTED]. It is accepted that EPI’s [REDACTED] remains a secret, so Symphony cannot have copied that. That is far from the full significance of this point. For the product to work [REDACTED].

So technically the fact that [REDACTED] are different is alone a strong pointer against an inference of copying. But there is also the question of proportions. The difference for the [REDACTED]. So clearly there is no apparent exact copying. To arrive at the notion that Symphony got [REDACTED] by copying one has to postulate either that the EPI product was mis-assayed or that it was assayed correctly and a deliberate decision was made not to have the same proportion. Neither of these postulates leads to a prima facie conclusion of copying. As to the [REDACTED], again the proportions are wildly different, which again does not suggest copying.

Mr Hobbs suggested that EPI’s expert witness, Dr Wiles, did opine that the similarities between the formulae themselves were so close that there was an inference of copying. I do not think he clearly so stated. The passages relied upon do not spell that out clearly enough, if indeed they say it at all. What he said was:

“The defendants’ BD92384 formulation contains [REDACTED] of the claimants’ DCP509 additive in functionally and numerically similar amounts and proportions”

That is not enough. Of course the components are functionally the same – [REDACTED] for the ultimate plastic film. Functional similarity is not enough to raise an inference of copying.

I note that there was a dispute about the [REDACTED] – Dr Wiles suggesting it had a function whereas Mr Barclay thought not. If Dr Wiles is right, then the difference in quantities would, it seems, matter and be an even stronger pointer against copying. As to “numerically similar” all that Dr Wiles seems to be essentially comparing is the numbers as such. It is true he later says the amounts are “functionally” similar – but that is not saying the differences in amount are so trivial that the fair inference, from the formulae alone, is copying.

Above all Dr Wiles fails adequately to deal with significance of the difference in [REDACTED]. This does not begin to suggest that Symphony either got its [REDACTED] from EPI or deal with the question of how Symphony arrived at [REDACTED]. That could not have been done by copying – a matter upon which Dr Wiles is silent.

By way of conclusion Dr Wiles said this:

“103.

For the reasons given in this report, I consider that the Defendants’ formulation of BD92384 substantially conforms to the content and function of the formulation of DCP509.

104.

I have asked myself whether this, consistently with the information provided by the Defendants and Wells Plastics as to the origin and evolution of BD 92384, could be a coincidence in circumstances where the Defendants had the opportunity and also, apparently, the motive, to find an alternative to DCP 509 [REDACTED].

“105.

For the reasons given in this report, I cannot conclude that this is a coincidence and can only conclude that the degree of replication observable in BD 92384 as compared with DCP 509 is attributable to a process of substantial copying.”

This is not saying that a mere comparison of the formulae alone is so improbable technically that independent derivation is improbable. It takes into account “the reasons given in this report.” These include other things too (e.g. how quickly the formulation was arrived at), motive and so on – non-technical matters and technical all mixed up together. Dr Wiles was here offering an opinion on the ultimate question, the question which was for the judge. He was going beyond his expertise. I am not surprised that the Judge found that Dr Wiles spoke “as an advocate” and treated his evidence with caution.

As to Mr Hobbs’ other points supporting his suggested inference, there is little in them. I agree that Dr Whiteman’s evidence shows some possible intention to copy, but it does not follow that there was copying. Whether there was depends on what was done. The fact that Symphony wanted a “seamless” change loses all significance once one realises that all that was wanted was a product which would satisfy RAPRA tests, not an identical product to DCP 509. Once one appreciates that a successful formulation depends on getting the [REDACTED] there is nothing in the speed and lack of trial point – for neither of these things could have been achieved by copying. And the fact that some of Symphony’s witnesses were disbelieved on some matters, but not on the crucial question of copying from DCP 509 is simply not a pointer towards copying.

So I reject Mr Hobbs’ onus point. That brings me to what his positive case on copying was. I do not intend to go into every detail. The broad outline is enough. He submits that what Mr Stephens did first was to have analysed a product made by Willow Ridge. Having found out what that was, he set about having DCP 509 analysed by Mr Barclay at Wells. To cover his tracks in advance, he dishonestly called the product being analysed “Willow” in every document which refers to the exercise. Mr Hobbs was not explicit as to whether Mr Barclay was “in” on the subterfuge from the outset. He said he did not and could not know. But if he was not at the outset he was later and lied in his evidence to cover his tracks.

It is a bizarre – near fanciful case. It presupposes an enormous degree of cunning and foresight on the part of Mr Stephens. The Judge did not accept it. What the Judge found was that the product which was analysed by Wells and (so far as AA is concerned) by Giltech was not DCP 509. And that it was a Willow product, as the contemporaneous documents say. There were not, as Lord Justice Rix called them in argument, two products, “echt Willow” and “ersatz Willow.”

The sole foundation of Mr Hobbs’ suggestion that there was an initial analysis of a Willow product is Mr Stephens’ notebook. The page upon which Mr Hobbs relies contains a date, 07/02 and three entries, separated by lines. The first says “Evaluation of Willow.” Underneath that there are set out [REDACTED] written in front of them – presumably a proportion. They are the [REDACTED] used in Symphony’s product BD92384, albeit the proportions are [REDACTED]. They cannot have been produced by an analysis of a Willow product since it is common ground that the [REDACTED] cannot be discovered by reverse engineering. And besides analysis will not tell you the [REDACTED]. Mr Hobbs’ point rests on the entry below that: “Willow is [REDACTED]. From this page it is supposed to be inferred that by 2^nd July a Willow product had been analysed.

I just do not see that. The notebook itself demonstrates an extraordinary jumble of dates and jottings. The word “evaluation” is equally consistent with “trying out” as with analysis. The intervening entry of [REDACTED] is not consistent with any analysis of Willow. And the entry relied upon, with its emphasis and exclamation is actually more consistent with Mr Stephens having discovered at some point that, contrary to what he had been told by Willow, a Willow product did contain [REDACTED]. The Judge dealt with the notebook at paragraph 37 and has not been shown to be wrong. For what it is worth, if it had been up to me to decide, I would have decided the same way.

Next Mr Hobbs faces the problem of the AA analysis done by Giltech for Wells. It is not suggested that anyone at Giltech was in on any dishonesty. The witness from Giltech, Mr Healy, their Director of Research and Development was accepted to be an honest witness. Giltech provided all the documents they have about the analysis done for Wells as a favour using their AA machine. The key documents consist of 3 pages of detailed readings from three samples. [REDACTED] of what was being analysed. The agreed expert evidence is that the AA technique of analysis is accurate to at worst ±5% but in practice people take it as accurate to ±2%.

Since DCP509 contains [REDACTED] Mr Prescott submitted that what was analysed by Giltech cannot have been DCP509 – it is well outside the range of possible experimental error. And if that is right, that is an end of EPI’s case. Mr Hobbs sought to rebut this in two ways. Firstly he suggested two other possible sources of error. Secondly he relied upon a passage in Dr Wiles’ evidence.

The first further source of experimental variability was suggested to be that arising from manufacturing tolerances. Mr Barclay referred to these in respect of product manufactured by Wells saying that it was in the region of ±2%. That is of course not EPI’s manufacturing tolerance but we were invited to assume that it would be about the same as Wells – a not unreasonable assumption. This is nowhere near enough to explain the difference between measurement [REDACTED].

Mr Hobbs’ other source for suggesting a discrepancy related to possible variation between samples measured. He particularly pointed to the analysis of a Willow product done by Pyxis in December which suggested that there could be considerable variations – Pyxis giving a variation of between [REDACTED] of the sample analysed by them. The difficulty with that is that Pyxis’ method of analysis was of a small part of the surface of a single granule. Actually Pyxis was not even happy with the [REDACTED] as their report indicated. But Giltech’s method did not involve surface analysis. Three separate half gram quantities of granules were ashed and dissolved in concentrated nitric acid. Variations in the surface, or indeed variations between granules, would all be averaged out and be irrelevant to Giltech’s method of analysis.

Accordingly, in my judgment, although this is not dealt with in detail by him there was ample material for the Judge to say, as he did:

“However, given the difference of percentages, I do not see how Giltech could possibly have analysed DCP 509 and produced the figures they did.”

Actually I do not see on these detailed experimental and technical considerations how the Judge could have concluded otherwise. But that does not matter. The question is whether the Judge has been shown to be wrong. And he was not.

Mr Hobbs submitted that Dr Wiles had opined that the Giltech results were consistent with the product being DCP509 and was not really cross-examined about that. The trouble with this is (a) that Dr Wiles was found to be a partisan expert witness, (b) that he gave no reasons for his opinion and (c) Professor Scott, and other witnesses had not agreed with him. Given the detailed technical and scientific reasons I have set out in detail, I do not think this point carries any weight.

Finally on this [REDACTED] point Mr Hobbs tried to rely upon two handwritten figures on a letter sent by Miss Simmons of Wells to Mr Healy of Giltech. [REDACTED] The average of these is [REDACTED]. Mr Hobbs suggested that this was consistent with the product being DCP509 whose [REDACTED]. This point has formidable difficulties. First the two figures themselves are not consistent with the accuracy of the AA technique. Second, no detailed documents emerged from Giltech (accepted as honest) showing these measurements – which is to be contrasted with the many detailed measurements for the [REDACTED] measurement. Third, Mr Healy, although not himself having conducted the measurements, gave the opinion that the figures were most likely no more than “sighting” figures. The Judge accepted his evidence and particularly quoted from Mr Healy's evidence on the point. The Judge said:

“Mr Healy provided an explanation as to the variance and I accept that”

The Judge was not shown to be wrong. Again, for what it is worth, I would have come to the same conclusion.

That actually concludes the appeal – what was analysed at Giltech cannot have been DCP509. The Judge has not remotely been shown to have been wrong.

Mr Hobbs however, had another point. He suggested that some other measurements made by Wells, not Giltech, show that DCP509 was subject to analysis by Wells. This was analysis by Fourier-Transform Infrared, FTIR. The Judge dealt with this in detail [REDACTED]. He unhesitatingly preferred the evidence of Mr Sidwell to that of Dr Wiles, giving details reasons as to why. He was not shown to be wrong in so doing. I will therefore be brief in dealing with the point.

Wells were asked by Symphony (note not EPI) to provide what documents they had about the work done for Symphony. They did. The documents included 4 graphs showing an FTIR analysis. The first question is whether the samples were anything to do with Symphony at all. For although one is marked in an unknown hand “Symphony” another of them clearly relates to another, Wells, product, BD91971. The Judge accepted Mr Barclay’s supposition that the 4 graphs had simply been misfiled. He did so because he not only had that evidence from Mr Barclay but also he considered that the detailed technical evidence supported it. And of course there was the fact that the graph for BD91971 must have been misfiled because it was in the Symphony file.

EPI concentrated on graph 1 as it was numbered in our papers. It is headed [REDACTED] which was said to mean [REDACTED] – which I will accept. It shows two remarkably similar traces, one in red and the other in blue. The blue, but not the red, shows [REDACTED]. Initially Dr Wiles said no more than the blue trace was consistent with the product being DCP509 or an additive containing an [REDACTED]. “Consistent with” of course can be a long way from “pointing to.” Dr Wiles did not consider the red trace, or why it looked so similar to the blue apart from the [REDACTED]. Nor did he comment on the fact that the title does not suggest a comparison between the traces of two different products was being made, whereas the titles of the other graphs do say that, in two cases by a handwritten addition (not the hand that wrote “Symphony” on one of them). Two of the other graphs repeat the same two traces with additional traces – suggesting indeed that they are comparisons with the single product of trace 1.

Mr Barclay said that graph 1 was in fact of an additive containing [REDACTED]. The two traces were both of the same substance. Mr Hobbs at one point submitted erroneously that it was common ground that neither trace could have been of this substance – but in the end accepted he was mistaken. Apart from the [REDACTED] on the blue trace they are consistent with the product being of [REDACTED]. So what then are the extra blue peaks? The judge had three theories advanced. Dr Wiles (who it seems ignored the red, or assumed it was a trace of a different product from that of the blue) advanced the case that they showed [REDACTED]. The argument then ran this way: the product is assumed to be that supplied by Symphony. Willow itself does not contain [REDACTED]. Giltech had shown the product contained [REDACTED]. So the product must contain [REDACTED] – and the only contender for a possible sample in these circumstances is DCP 509. The second theory was advanced by Mr Barclay who suggested contamination from the hands of the operator. The third was that of Mr Sidwell who suggested machine contamination.

The judge did not accept Dr Wiles’ deductions. He accepted the view of the wholly independent, more experienced, Mr Sidwell of RAPRA. He gave detailed reasons for so doing. Despite Mr Hamer’s valiant efforts to show that the judge was wrong, I was wholly unconvinced that he was.

I would only add this: this point only came into the case at a very late stage (initially Dr Wiles had said: “the quality [of the spectra] makes it impossible to make a definitive statement that a [REDACTED] is present but there is an indication of this.”). It was Symphony, not EPI, which instigated some experiments at RAPRA (which were, in the Judge’s words, “something of a fiasco”) designed to show that trace 1 was indeed of an [REDACTED] based additive of Wells, one called BD92336. That was hardly the action of a group of intending perjurers conspiring to pervert the course of justice.

Mr Hobbs’ reliance on the FTIR evidence fails.

Next and last, on the period during which the Symphony product was devised, I should deal with the evidence of Mr Barclay. The Judge held that he was “truthful in respect of the key issue as to how Symphony’s product was formulated” although, contrary to his assertion, the Giltech results of [REDACTED] got him “started.” Mr Prescott accepts that latter finding and so must we, though I confess that I wondered whether the Judge might have made an error here, confusing for a moment [REDACTED]. He said [REDACTED] that Symphony’s product has [REDACTED]. That must be an error. The product has [REDACTED], which translates to [REDACTED]. This raises the question why, if Mr Barclay knew the figure was [REDACTED] and used that to get started he would depart from it by as much as [REDACTED]. But no matter. What does matter is that the judge carefully examined every aspect of Mr Barclay’s evidence before reaching his conclusion of independent design. It was simply not shown to be wrong.

That leaves the remainder of the evidence. This is about events subsequent to the involvement of Wells. It hardly matters, unless it shows some other alleged analysis of DCP509, which it is not suggested it does. The Judge analysed it in depth. He was critical of a number of the Symphony witnesses, particularly how much they knew about Mr Barclay’s work. I do not propose to go into all this. In my judgment he was not shown to be wrong. It suffices, I think, to deal with the allegedly most significant matters.

First is to note the heart of Mr Hobbs’ case as to these subsequent events. It is that the various witnesses (Mr Stephens, Mr Blacher and Mr Laurier) all knew that DCP509 had been copied and that they knew Mr Barclay had come up with a successful additive. All their subsequent actions, by way of creation of written documents and by way of sending off samples of Willow to Pyxis, were designed to conceal this. It is a near fantastic allegation. I cannot see the beginnings of a basis for it, even though the evidence was unsatisfactory. I will deal therefore, with only some of the major points

Perhaps most important here, is the fact that none of the witnesses could have known that Mr Barclay had come up with a successful formulation. Even supposing they knew he had come up with a formulation at all (only Mr Stephens said he did – the others, he said, were being kept in the dark so as not to raise expectations), they could not have known it would be successful (i.e. pass the RAPRA accelerated tests for degradation of plastic film made with the additive) until RAPRA so reported on about 11^th December.

So, until then, if Symphony were to get away from dependence on EPI, which it clearly wanted to, it would have to get additive from some other supplier if it could not make its own. So the November renegotiation of the EPI contract which enabled Symphony to use non-EPI additive is no indication whatever that anyone at Symphony knew they had their own “in-house” product.

What is one to make of two other matters which featured large in argument? First Mr Blacher’s note of 4^th December and second the fact that samples of a Willow product were sent to Pyxis for analysis?

The Judge dealt with the note [REDACTED]. His analysis seems to me to be about the most rational explanation one can find for it. I do not think his conclusion paragraph [REDACTED] as to Mr Blacher’s degree of knowledge is inconsistent with what he says [REDACTED] – he does not find [REDACTED] that, as of the date of the note, Mr Blacher knew there was a product not only formulated but under test at RAPRA.

What then of the sending of a sample of Willow to Pyxis for analysis? Why should this have been done at all if Symphony already knew it had a successful product? Mr Hobbs said the motive here was to create a paper trail so as to show, falsely, that the Symphony product was created by reverse engineering of Willow whereas it was in fact by reverse engineering of DCP509. This on its face is an absurd allegation – for, unless the Willow product was the same as DCP509 (which it is not) the trail would lead to a wrong formula and be sure of exposure if it were to be examined. Mr Hobbs suggested that the hope was that Pyxis would come up with the same figure as Giltech – but why would any rational man, however dishonest, expect that to be if he thought Giltech was done on DCP509 and Pyxis were testing Willow? And why, when in fact the paper trail did not produce the supposedly hoped for result, would the rational man press on and use the Barclay additive?

Moreover the dates do not fit with Mr Hobbs’ theory. For Pyxis reported on 12^th December – so the products were sent for analysis before anyone at Symphony could have known that the Barclay product “worked.”

I do not delve further, that is not the function of the Court of Appeal. I fear I have already delved more than is called for by a “review” to see whether the Judge was wrong. I think his careful and detailed discussion and weighing of the evidence was unimpeachable.

The Judge’s other reasons for dismissing the action, was that the key features of the DCP509 recipe [REDACTED] were published by EPI [REDACTED], thus destroying anything confidential about them. It followed, the Judge held, that the claim in breach of confidence failed. Moreover he held that the “no-analysis” provision of the licence could not prevent an analysis which yielded only public domain information.

None of these questions begin to arise unless there was an analysis. The Judge held there was none. We have upheld him and so indicated to the parties. So these other points do not arise – they are purely academic. Mr Hobbs asked us to resolve them nonetheless. We declined to do so for a number of reasons.

First it is not the function of the court to resolve purely academic questions. Mr Hobbs suggested the question was not purely academic because, on the Judge’s findings, Symphony were free to modify their product so as to make it more like DCP509. But there was no evidence it intended so to do. And if it were to use information revealed as a result of disclosure (e.g. that [REDACTED] the ingredients of DCP509 were [REDACTED]) there would be a breach of its obligation not to use such information.

Second the question of whether or not and if so how far the law of confidential information can cover material in the public domain is difficult legally. It may well be fact sensitive – for instance it may make a difference if it is the claimant himself who puts the information into the public domain or may involve a question of degree of notoriety of the prior publication. So there are particularly good reasons not to try to decide the question as an academic point.

Thirdly it was far from clear what the point was or what Mr Hobbs’ clients actually wanted or could get by our deciding the point. No draft order or declaration was offered. We could not see what we were being asked to do other than just give an advisory “opinion” – which itself, given our primary findings, would probably not constitute a binding precedent for any actual proposition of law.

Mr Hobbs abandoned his appeal on passing off. He sought to raise one other point – based on clause 14.4 of the Agreement. The clause required Symphony to return to EPI all “know-how” and all documents bearing of EPI’s trade marks. Before the Judge this was treated as an obligation solely concerned with documents. He found that there was a breach but that the failure to return the documents led to no damage and accordingly awarded only a nominal sum of £2.00.

On appeal Mr Hobbs asked for an enquiry as to what damage had been suffered by breach of this clause. He accepted that the Judge was right so far as the non return of documents was concerned, but sought to raise a new point. He submitted that the evidence showed that Symphony had supplied to a Chinese company 5.7 tonnes of additive and that this had not been returned. He argued, for the first time, that the additive itself should be treated as “know-how” because of the definition of that term at the beginning of the Agreement. He said he did not know what had happened to this material and he ought to be able to find out and, if necessary, claim any damage caused by its non return – by implication a suggestion that the Chinese company may have analysed the additive.

Miss Berkeley responded. She pointed out that it was not said the additive had not been properly supplied to the Chinese company: Symphony were, subject to appropriate conditions, permitted to supply to sub-contract manufacturers and the supply had been during the currency of the Agreement. The matter was not fully gone into at trial: the point now sought to be raised was not an issue at the trial. It was established at some point that Symphony could not, in fact, secure the return of the material despite attempting to do so. Quite what happened to it remains unknown. Mr Hobbs was not able to indicate any kind of damage at present. He did not even know whether the Chinese company had consumed the additive before the final termination of the licence.

In my judgment it is too late for Mr Hobbs to raise this point now. There are matters which could have been dealt with in evidence. They may have indicated a prima facie case of possible damage. But as matters stand an enquiry under clause 14.4 would be entirely speculative.

Moreover, it should not be forgotten that the Judge did order an enquiry as to any damage which may have been suffered under clause 14.8. This will cover any manufacture or sale after the date of termination of the licence. EPI must be content with that.

I agree.

The judge had a rich harvest of evidence, both factual and expert, to consider, and in this court we have been asked to say that he came to the wrong conclusion on an allegation of copying, which ultimately he rejected. Such an appeal, on a question of fact, a fortiori in a case where the structure of the argument involved submissions that a number of Symphony’s witnesses were lying, and had agreed to lie, in pursuit of a dishonest course of conduct, is in the nature of things an uphill struggle.

In the present case, Mr Hobbs QC has made the most in support of EPI’s appeal of certain difficulties in Symphony’s case. Thus the judge accepted the evidence of Dr David Whiteman that in 2001 (the year before the events with which this appeal is concerned) he was invited to express a view as to his ability to analyse what was probably EPI’s DCP 509 additive. He was able to exploit Symphony’s embarrassment in putting forward a defence, viz reverse engineering of the Willow Ridge product, which was inconsistent with Mr Barclay’s evidence as to how he had developed his own additive for Symphony and was not believed by the judge. There were the added circumstances that Symphony had little to disclose in the way of its own paper-trail regarding the development of its own additive; that Mr Barclay was found, contrary to his evidence that he had relied on nothing but his own intuition, to have been given a start by the Giltech analysis of the Willow Ridge additive; and that some of the evidence of Symphony personnel was described by the judge as extraordinary and incredible. Nevertheless, the judge carefully considered all aspects of this evidence, but was still not persuaded that Symphony’s additive was reverse engineered from DCP 509.

Lord Justice Jacob has, moreover, explained the technical, documentary, and forensic reefs by which EPI’s case is not merely diverted but destroyed. Of particular moment, it seems to me, are the facts that it would be extremely difficult (and useless) to copy EPI’s [REDACTED] and yet there is no complaint of copying the latter; that the Giltech analysis is not consistent with its subject matter being DCP 509; that the FTIR graphs are not shown to be of DCP 509; that Mr Barclay, who is an employee of an independent company and not of Symphony, was accepted as being in essence a witness of truth; and that the Symphony personnel witnesses, for all the unsatisfactory elements in their evidence, were not found to have lied on the critical question of the copying of DCP 509. Finally, there was the forensic need, and improbability, of a conspiracy to tell lies in court. This is all irrespective of any question of shifting burdens of proof.

When viewed in this context, so many of Mr Hobbs’ or Mr Hamer’s points can be seen to be essentially speculative inferences, some more possible than others, some more improbable than others, but all essentially driven by an a priori need for consistency in the making of a case driven by the result and remedies desired by EPI rather than a search for the untidy facts of life.

The trial judge had to make that search. His findings indicate that to some extent Symphony and its personnel, knowing that they were seeking their own substitute for EPI’s product, did not conduct themselves or the ensuing litigation in a way to allay suspicion. However, he showed care and objectivity in arriving at his findings, which Mr Hobbs’ and Mr Hamer’s sustained advocacy has not in my judgment displaced.

This appeal took a remarkable course. The major part of the argument on paper before the appeal opened, and in the event the whole of the argument before us, related to EPI’s attempt to dislodge the judge’s conclusion that BD92384 had not been copied from DCP509. In pursuit of that issue there were before the court five files that contained the pleadings, judgments, witness statements, and an “abridged” version of the trial bundle which, together with very full skeleton arguments, might have been thought to provide ample material on which to consider the issues. In addition, however, the parties between them contrived to put before the court the whole of the transcript of the trial below, extending to some 1800 pages, and a further eleven files containing exhibits and materials that had been before the trial court. These files were not sequentially paginated, but inspection of them suggests that in total they fell not far short of 5,000 pages. There were also copied or re-copied all of the material that had been used on the application for permission to appeal, again not sequentially paginated, but appearing to extend to some 900 pages.

The appeal simply on the issue of whether the judge’s finding that copying had not taken place should be reversed occupied the court for three whole days. That indulgence was largely in deference to EPI’s strongly expressed contention that the judge had gone seriously wrong in his approach to the case and had thereby committed an injustice. Nevertheless, it was a failing in case management, for which the presiding Lord Justice accepts full responsibility, that an issue of that order, however important to the parties and whatever their complaints, should have been permitted to occupy that length of oral hearing. The court however does not accept responsibility for the exorbitant amount of material that was placed before it, not least because the parties were warned on the permission application that the material before the court on that occasion (from the figures set out above, in itself much shorter than that which was eventually lodged) must itself be severely pruned.

This amount of material was placed before us in support of what was in effect an attempt on the part of EPI to re-open the judge’s findings of fact and exercise of judgement in relation to the extent and nature of the copying that was alleged to have occurred. In so doing, EPI will have had in mind the guidance given to this court by the House of Lords in Designers Guild Ltd v Russell Williams [2000] 1 WLR 2416 at p 2418G: it is not for the Court of Appeal to embark on an issue of substantiality of copying, and therefore a fortori of copying at all (I would respectfully add), unless the judge has misdirected himself. In accordance with that guidance Mr Hobbs did indeed submit that the judge had misdirected himself. The basis of that submission was that it had been established that BD92384 was sufficiently similar to DP509 for an inference to arise that it had been produced by copying. Once that inference was established it fell to Symphony to establish an innocent explanation for the similarities. If the judge had correctly followed that approach it would, or at least might, have become clear that copying had taken place. There are a number of difficulties in this argument, some of them general and some of them specific to this particular case. I will endeavour to set them out in some sort of logical order.

First, to describe this process, as is sometimes done, as involving a shift in the burden of proof conceals that the burden of proof remains on the claimant throughout. Jacob LJ, in his §7, considers that that latter point was indeed understood throughout the appeal. I am less confident on that score, to the extent that I take the liberty of repeating that proof of sufficient similarity merely places on the defendant an obligation to give an explanation of that similarity, which the court then considers together with all the other facts. That the effect of an initial finding of similarity is so limited was firmly established in the case often cited for this point at least in the copyright field, Francis Day & Hunter v Bron [1963] 1 Ch 587, Willmer LJ, at p 612, characterising as quite untenable a wider rule contended for by counsel in that case, that would indeed have placed a burden of proof on the alleged copier. The jurisprudence is the same in other fields, for instance that of medical negligence, where unusual and unexpected damage to the patient calls for an explanation from the doctor but does not place a burden of exculpation on him: see for instance Lillywhite v UCL NHS Trust [2005] EWCA Civ 1466[87].

Second, the issue usually arises in, and the authority cited by Mr Hobbs is drawn from, the field of copyright. It is likely to be much easier for the court to operate a test of substantial similarity when, as in the classic copyright case, comparing two documents or works of art, for instance in Francis Day v Bron two musical scores; than it is for the court to compare two technical products, to see if the one is “similar” to the other. To look upon this picture and on this calls only for the skills or sensitivities taken to be possessed by any educated person, ergo by any judge. But if a judge is to decide that one industrial product or chemical formula sufficiently resembles another he cannot rely on his own experience, but needs expert evidence. Mr Hobbs said that such had been before the court in this case, in the shape of the evidence of Dr Wiles that, in the summary passage set out by Jacob LJ in his §12, said that

“The Defendants’ BD92384 formulation contains [REDACTED] of the Claimants’ DCP509 additive in functionally and numerically similar amounts and proportions.”

Jacob LJ has already set out the ways in which this evidence was inadequate for the role ascribed to it. I would add one further point. It was argued that the defendants’ expert, Professor Scott, had agreed with that analysis in that he had not listed that observation in his reservation from the joint report of the two experts. I fear that that claim only goes to demonstrate that if a scientific finding is to be made the basis of a significant legal argument it has to be made clear that that is what is being done, and the experts on the other side be given a fair chance to comment. That is not how this point developed, nor was it clear that that was the thrust of Dr Wiles’ evidence. To say that the components of the two products were “functionally” similar adds nothing: there was no doubt that the [REDACTED] had the same purpose, but that in no way leads to an inference of copying. To say that the components were numerically “similar” is not enough, without more, to demonstrate that a prima facie inference of copying should be drawn: an inference that, elsewhere in his evidence, Professor Scott did not accept.

Third, however, it is difficult to see that it is likely to make much difference in most cases, and certainly not in the present case, whether an initial inference of substantial similarity is drawn or not. In any case where there is similarity of products, whether or not “substantial” similarity, and there are circumstances in which copying could have occurred, commonsense will require the defendant to give an account of how he produced the goods. If Symphony had simply said nothing, and left it to EPI to prove its case, it is very difficult to think that the judge would not have drawn adverse inferences. As it was, even though the judge did not go through the formal steps that EPI said that he should have done, he received, and certainly did not take without scrutiny, an explanation from Symphony of how the product had been produced. He accordingly and in any event did all that he was required to do by the principle here under consideration.

It was a very uphill task for EPI to seek to offset the explanation that the judge did accept. That is not just because of the great difficulty of disturbing a decision of a judge who has seen and heard witnesses, whose testimony is to assist him not only on simple issues of fact but also in terms of probability and the overall plausibility of one case as against the other. Even more pressingly, as the present case developed before us it became clear that EPI could only disturb the judge’s finding as to Mr Barclay’s origination of the product if it could show that there had been, or at least might have been, what would have amounted to a fraudulent conspiracy between Messrs Stephens and Barclay to copy and then to cover their tracks; and subsequently between them and others to maintain that deception in the face of the court. The judge was well aware that the stakes were as high as that. His very positive findings in respect of Mr Barclay, and his clear finding that, whatever else might be said about his evidence, Mr Stephens was not telling lies about the key steps in the development of the BD92384, were made in that context.

Even if that had not been, as it was, the end of the case, the detailed re-examination that we were taken through of the materials on which the judge based that conclusion demonstrated, to me at least, that far from this being simply a case where we could not interfere with the judge’s finding, any other finding would have strained credulity. I add only one consideration, to emphasise what my Lord has already said on this subject. Much attention was paid to events in July 2002, and to the sending of the sample to Mr Barclay by Mr Stephens. We were asked to accept that that sample was of DCP509, despite being called “Willow” in Mr Stephens’ e-mail to Mr Barclay of 23 July 2002; and that the designation “Willow” was a deliberate misnomer, agreed between the two men to throw subsequent enquirers off the track. This arrangement would have had to have been made between two people who, even on EPI’s case at its highest, had only met each other a month previously; in respect of whom there is no evidence of any sort of detailed contact; and one of whom was an employee of an independent company with no history with EPI and very little to gain from conniving in a hazardous scheme that exposed it to justified attack from that very large player in the market. And the arrangement would have had to be wholly premeditated, down to the production of false e-mails. Very cogent proof would be needed of anything as extreme as that.

These considerations only go to reinforce my agreement with the conclusion reached by my Lord, and with the reasons that he gives for that conclusion.

I finally add one case management point. Mr Hobbs urged that, even though our finding on the copying point entailed the dismissal of the appeal in any event, we should nonetheless go on and review the judge’s findings on what were said to be two important issues of law: first, that EPI’s material, otherwise confidential, had been put into the public domain through publication [REDACTED] referred to by Jacob LJ in §46 above; and, second, that no complaint could be made of breach of confidence in respect of material that was in the public domain.

As a matter of principle the court does not hear issues that are moot unless there is good reason for so doing. EPI are protected, so far DCP509 is concerned, by Symphony’s continuing undertaking not to publish or make use of information that has come to them by reason of the proceedings; and although that is no doubt not all that Mr Hobbs might like in an ideal world, I was not persuaded that there was anything more that his clients could legitimately expect from these proceedings in the form that they took. The more general issues of law undoubtedly raise questions of difficulty. I for my part find it by no means straightforward to reconcile on the one hand the apparently blanket rule that any claim of breach of confidence must fail if the material in question is in the public domain (see e.g. Saltmann v Campbell Engineering (1948) 65 RPC 203 at 215.10 and Mustad v Dosen [1964] 1 WLR 109); and on the other hand the “springboard” cases, that seem to inhibit use of even public domain material if it is conveyed in circumstances that aspire to confidence: a difficulty that, with respect, is not resolved by the observations in the House of Lords in A-G v Guardian Newspapers [1990] 1 AC 109 at 285. These may be questions of considerable general importance that call for an answer in this court or, quite likely, at a higher level. It is that importance, and that difficulty, that make it quite inappropriate to try to decide them on what would be, in the light of the decision that this court has already made, hypothetical facts.