LZLabs Gmbh & Ors v IBM UK Limited

Grounds 1-6 all seek to rely, in one way or another, on the Software Directive and make a raft of points in respect of claims for copyright infringement. IBM contend that this is seeking to turn the case on its head, because this was a claim not for copyright infringement but breach of contract. Some of the other difficulties with these grounds are that they are regularly untethered to paragraphs of the judgment and, on analysis, appear to be attempts to unpick the judge’s findings of fact.

I note that at [74(i)] of her consequentials judgment, the judge said about the early version of these grounds:

“i)

Grounds 1 to 7 in Section (A) raise alleged errors in the Court’s understanding or application of the Software Directive but fail to identify specific errors of law or principle in the Judgment, which sets out the relevant legislation, case law and principles at paragraphs [214] to [270]. The reference to the recent Sony v Datel judgment of the CJEU does not assist because it is not binding, was not relied on by the parties, and does not directly address the issues before the Court.”

I agree with the judge’s reasons for refusing PTA. I have concluded that Grounds 1-3 are particularly egregious examples of many of the difficulties I have already identified. Grounds 4-6 are perhaps more in line with the case at first instance, because the appellants seek to rely on Articles 5 and 6 of the Software Directive in a similar way to that which they advanced to the judge, although they fail to address the detailed reasons how and why the judge rejected these submissions. I consider that in addition, Grounds 1-6 do not begin to get over the particularly high permission hurdle in this case.

This complaint is set out at [1]-[8] of the appellants’ grounds of appeal and [25]-[41] of their skeleton argument. It contains a number of general criticisms of the judgment although there are very few references to the text of the judgment itself. The most focussed submission concerns [258] of the judgment, where the judge said that copyright protection for computer programs “extends to expression of the program but not to ideas and principles underlying functionality of the program”. It is said that this is “an incomplete oversimplification”. The appellants then give various examples of where the judge found examples of “protected expression” in circumstances where the information in question was not a literal copy of any of the source or object code of any IBM computer program.

The ‘literal copy’ point appears to be one of the key elements of this complaint. The appellants say the judge assumed that any such expression was that of IBM rather than the customer who had authored it or made the arrangements necessary for its generation by the compiler. It is said that the judge did not find that any specific code of any specific IBM program had been reproduced in the results of Winsopia’s analysis for Items 3-4, 6-15, 17-30 or 49. As Mr Stewart put it at the hearing, “anything that fell short of reproduction cannot amount to expression” and was, by implication, not a breach.

It is also said at [40] of the skeleton that, if the judge had applied the correct legal test, she would have found either that the ICA on its proper construction did not purport to prohibit the acts performed by Winsopia in relation to those Items or alternatively that, to the extent that the clauses of the ICA did purport to do so, they were void pursuant to s.296A CDPA. This was not apparently an argument advanced below. It is a very general submission, with no analysis and no reference to the individual parts of the judgment apparently being criticised.

In response, IBM rely on clauses 4.1.2 and 4.1.3(a) of the ICA, the latter of which stated that Winsopia would not “reverse assemble, reverse compile, otherwise translate or reverse engineer the ICA Program unless expressly permitted by applicable law without the possibility of contractual waiver”. They say that that is what the judge found that Winsopia had done: picking apart the ICA programs to access the elements they wanted from them. IBM say that it was a breach of contract to reverse assemble/compile/otherwise translate/reverse engineer in order to gain unauthorised insight into the internal implementation of their programs, irrespective of whether there was any reproduction of IBM’s code. They reiterate that this issue was purely a matter of contractual interpretation and that copyright law was irrelevant. They say that the ICA did not need to be glossed as requiring “literal reproduction of the text of any IBM computer program” for there to be a breach.

Further, IBM say that the appellants have failed to explain how the judge’s supposed error in relation to the idea/expression dichotomy offers a defence to IBM’s specific claims in contract (or in procuring a breach of contract or an unlawful means conspiracy).

In his short reply at the hearing on 20 June, Mr Stewart recognised the significance of Ground 1. He said that Article 5(3) of the Software Directive allowed Winsopia to observe/study/test. He said that was (i.e. was the same as) reverse assemble/reverse compile/otherwise translate/reverse engineer. In this way, the picking apart of the IBM programs which the judge found was entirely lawful.

In my view, ground 1 has no real prospect of success. That is primarily for the reasons advanced by IBM. In particular:

For the avoidance of doubt, it is plain to me that observe/test/study is entirely different to reverse assemble/reverse compile/otherwise translate/reverse engineer. The judge’s findings of fact demonstrated that, by picking the ICA Programs apart, the appellants had not done the former, but had done the latter. As noted already, the appellants cannot open up those findings of fact/evaluation.

The complaint is set out at [9]-[12] of the appellants’ grounds of appeal, and in the skeleton argument at [42]-[52]. It is said to follow “as a result of Ground 1”. It identifies three “fundamental problems of logic” in the judge’s assumption that elements appearing in the object code of a compiled customer application, which had been generated by an IBM compiler or translator, necessarily formed part of the protectable expression of the code of that IBM compiler or translator. The only paragraphs of the judgment to which reference is made at [43] of the grounds, which are simply said to be examples of where the judge treated the output of the compiler as part of the compiler, simply because that output consisted of code. Again it is said that this approach was far too simplistic.

The appellants submit that the right analysis is that when a licensee pays to use a commercial compiler or another similar tool, they obtain the right to use that compiler to produce object versions of any source code that the licensee may author. The appellants argue that the licensor has thereby given up any right to object to further use of the code emitted by the compiler at the customer’s instruction “which solely reflects the intellectual creation of the author of the source code of the customer application.” It said that this applies to most of the Items of breach: see [51] of the skeleton argument.

IBM maintain that, since this is said to follow from Ground 1, and Ground 1 must fail, then so too should Ground 2. Their principal substantive response is that the judge made detailed findings about the Programs and how and why they were picked apart by the appellants in breach of the ICA, so that this ground is essentially an attack on the judge’s findings of fact.

In my view, Ground 2 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

The appellants’ case is set out in the grounds at [13]-[15] and in the skeleton argument from [53]-[59]. The appellants submit that in general terms interfaces were inherently unprotectable ideas. Various authorities are cited in support of that submission. It is said that this is a further dimension of Ground 1 (see [54]). It raises the Article 5(3) defence – a reference again to the right to use a copy of a computer program “to observe, study or test the functioning of the program” – so as to say that what the appellants did, in order to determine the ideas and principles which underlie any element of the program, was to perform “acts of loading, displaying, running, transmitting or storing the program” which they were entitled to do under Article 5(3).

In this way, the appellants submit that therefore the Article 5(3) defence should have succeeded in relation to many of the Items which were found to be breaches.

IBM note that the judge found that the ideas and principles underlying any element of a computer program, including its interface, were not protectable through copyright: [258] and [269]. They say that that was entirely correct. But they say that, in the present case, the appellants are seeking to argue that the object code implementing an interface was also ‘not protectable’. That is a very different thing and there was no support for that contention. In particular, the respondent notes that the appellants’ assertion that the object code implementing an interface is “dictated by technical function” and “distinguishable from the underlying idea” are entirely new assertions which were not the subject of any evidence at trial. It is said it is too late for these points to be run now.

I consider that ground 3 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

This argument is set out in the grounds of appeal at [16]-[21] and the appellants’ skeleton from [60] to [69]. The argument is that the judge adopted an erroneously narrow interpretation of Article 5(3) (set out above, containing the authorisation to observe/study/test the functioning of the program). The judge is criticised for concluding that this exception to the protection afforded by the Software Directive was to be construed narrowly, as set out at [240] and [259]. The argument is that the judge should have found that each of the Items which she said constituted a breach of the ICA fell within this permitted exception.

IBM say that the judge was right in law to construe copyright exceptions narrowly, and relies on InfoPaq (as the judge did) for that proposition. Moreover, they say that this ground disguises the main point that the appellants argued below, namely that Article 5(3) permitted disassembly/decompilation/translation of source or object code, an argument that the judge roundly rejected. Accordingly, IBM say that the appellants are now attempting to argue that this exception applied even where, on the facts as found by the judge, it was only possible to study how the program produced outputs etc by disassembling/decompiling the object code of the program and studying its expression and not just its functioning.

I have concluded that Ground 4 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

This argument is set out in the appellants’ grounds at [22]-[26] and in the skeleton argument at [70]-[81]. Article 6(1) provides that “the authorisation of the right holder shall not be required where reproduction of the code and translation of its form…are indispensable to obtain the information necessary to achieve the interoperability of an independently created computer program with other programs…”.

The complaint is that the judge applied an unduly high standard of necessity and presupposed that “indispensable” referred to the goal of interoperability, rather than the acts of reproduction and translation that were permitted by Article 6(1). The complaint is most clearly articulated at [72] of the grounds: that, because the judge reasoned that Winsopia and LZLabs had an alternative migration path, they could avoid the need to migrate the customers application’s object code in the first place, and instead work out a new object code program compiled from the customer’s source code. In this way, the judge reasoned, none of the information was necessary to achieve interoperability, and relied on that reasoning to defeat reliance on Article 6(1) in relation to all of the Items in respect of which it was raised. It is said that that approach was circular and involved asking the wrong question.

IBM maintains that the judge was right as a matter of law, but that in any event it is not explained how some different standard(s) would have made a difference to any of the judge’s relevant findings. Indeed, the judge concluded that the appellants did not in fact engage Article 6 at all, because these operations were not performed for the required purposes of obtaining interoperability information. IBM say that the judge applied the test of indispensability in any event in evaluating the acts employed by the appellants to obtain the information and applied the test of necessity (not indispensability) when seeing whether the information obtained through such acts was necessary to achieve interoperability (see for example [400]). This was, they say, all entirely orthodox.

In my view, Ground 5 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

This is set out in the appellants’ grounds at [27]-[30] and in the skeleton argument from [82]-[92] and concerns Article 5(1), another example of the protection afforded by the Software Directive, this time in respect of error correction. It is said that the judge took an unduly narrow approach to the right to correct errors and failed to have regard to Top System. If she had properly understood the scope of Article 5(1) she would have held that Winsopia’s acts in Item 1 fell within the scope of error correction because they were not specifically prohibited by any clause in the ICA.

In response, IBM say that the complaint ignores the judge’s factual findings in relation to Item 1 that the disassembly was done, not to correct any error, but to improve the performance of the SDM (see [289] of the judgment). That renders Ground 6 moot.

I have concluded that Ground 6 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

Grounds 7-13: Arguments As To The Construction Of The ICA

The judge’s view of these grounds (making allowance for the slight changes in numbering) was set out in [74(ii)] of her consequentials judgment as follows:

“ii)

Grounds 8 to 15 in Section (B) seek to challenge points of construction of the ICA but do not identify any error in the Court’s approach or analysis. The material terms of the ICA are at paragraphs [64] to [75] and the analysis is at paragraphs [147] to [272] of the Judgment.”

I agree with the judge’s reasons for refusing PTA. The appellants challenge the judge’s construction of almost every clause in the ICA of which they were found to be in breach. The basis for the challenge appears to be that the appellants’ arguments at first instance, which were rejected by the judge for a mixture of reasons of law and fact, should have been accepted. There is no analysis of the judge’s reasoning as to why those arguments were rejected. This unsophisticated approach is not a proper basis for an appeal. At no point in relation to Grounds 7-13 is any clear error of law identified by the appellants. Furthermore, these grounds fail to clear the particularly high permission hurdle in this case.

This is dealt with in the appellants’ Grounds of Appeal at [31]-[36] and in the skeleton argument at [93]-[119]. The allegation is that the judge was wrong to interpret “ICA Program” as including all the Items identified at [31] of the grounds; the code fragments identified at [32]; and the macros and copybooks at [33]. It is then said that, if the judge had correctly interpreted “ICA Program” as not including output from running a licenced program, compiled programs and IBM programs subject to different licencing regimes, almost all of the Items that the judge found to comprise breaches of the ICA would not have been so found.

It is separately asserted that, if the judge had interpreted “ICA Program” as one of the 15 named IBM programs licenced by IBM UK to Winsopia, none of those items would have been a breach of the ICA. There is a yet further allegation at [36] of the grounds that, even if the ICA Program included its component parts, the judge should have construed this as subject “to a minimum threshold of substantiality”.

IBM reject this ground of appeal. They point to the definition of “Program” which refers to “the original and all whole or partial copies”. They rely on the judge’s careful analysis of the different constituent elements of an application compiled and link–edited to be run on one or more of the z/OS runtimes at [176]-[193] of the judgment. That included a close analysis of Licence Program Specifications at [185]-[193]. IBM submit that the appellants have simply not engaged with that analysis or identify how it is wrong.

IBM also submit that the appellants’ arguments proceed from the false premise that a compiled and bound customer load module should be treated as a single and undifferentiated object, the entirety of which was created at the behest of the application developer. They say that that was not what the evidence showed or what the court found.

At the hearing on 20 June, I asked Mr Stewart where the “substantiality” provision came from, given that it cannot be found in either the Software Directive or, more pertinently, in the ICA. He said it was a question of context. That was not developed further. To the extent that this was some sort of implied term, the basis for it was never set out.

I consider that Ground 7 has no real prospect of success. I agree with IBM’s response and consider the contrary is unarguable. In particular:

Clause 4.1.3(a) has been set out above. The appellants agreed not to reverse assemble, reverse compile otherwise translate or reverse engineer the ICA Program. The judge found that numerous of the Items comprised a breach of that obligation.

The relevant paragraphs of the grounds of appeal are [37]-[42] and the skeleton argument at [120] – [127]. Again there is a complaint that the judge did not accept the appellants’ arguments below. It is said that the judge “should have applied ordinary principles of contractual construction”; should not have had regard to the expert evidence; and misapprehended that evidence in any event. In consequence it is said that she was wrong to find that the appellants were in repeated breach of clause 4.1.3(a).

There is a separate point at [41] of the grounds in respect of Item 5, where it is said that reverse assembly had to require more than merely a mental act by an experienced human programmer. In addition, there is a reference made to the Software Directive and sections 50BA and 50B(2) of the CDPA and the prohibition or restriction of the right to observe, study and test.

IBM say that not only did the judge construe the ICA in accordance with ordinary contract principles but any alleged breach of those principles has not been identified. The judge rightly had regard to expert evidence as to the meaning of “reverse engineer” etc and that was an entirely proper use of expert evidence, even though it was of limited assistance: see [210]. IBM say that the judge rightly took into account the Software Directive at [213]. As to the Item 5 point and the mental acts, they point out that this is based on evidence of fact from Mr Lynch and was properly set out in the judgment at [394]-[397].

I consider that this ground has no real prospect of success. I again agree with IBM’s response. In particular:

Clause 4.1.3(b) comprises a prohibition on transferring the ICA Program outside Winsopia’s enterprise.

The complaint is at [43]-[45] of the grounds of appeal and [128]-[133] of the skeleton. It is said that the judge was wrong to interpret “transfer” as extending to the distribution by Winsopia to a third party of any copy of any component part of an ICA Program, irrespective of how small or insubstantial that component was. Instead it is said that transfer should have meant the ICA Program itself, alternatively a named component or part. It is said that various Items would not have been found to be in breach of 4.1.3(b) if the judge had interpreted the clause correctly.

IBM’s first point is that this is a rerun of the “small or insubstantial part” point that has already been raised – and which I have already rejected – at Ground 7. Further they say that it could not have affected the outcome because, for the Items identified, clause 4.1.3(b) was far from being the only provision to have been breached.

I consider that Ground 9 has no real prospect of success. I agree with IBM’s response. In particular:

On analysis these are two separate complaints here. Paragraphs [46]-[47] of the grounds complain about the interpretation of the word “use” in clause 4.1.1(a). Paragraphs [48]-[49] comprise a separate complaint about “authorised use” in clause 4.1.1(b). The relevant parts of the skeleton argument are at [134] – [139].

The first complaint is that the judge was wrong to say that “use” included mere receipt of a computer programme or receipt of information derived from analysis which did not itself contain any protectable expression for a computer programme. A series of Items were identified which, it is said, should not have been found to be breaches as a result.

The second complaint is that the judge failed to have regard to clause 4.1.1(c) when construing clause 4.1.1(b) as conferring a limited authorised use of the ICA Programs and therefore approached the question of what was “authorised use” on too narrow a basis.

In response to the first point, IBM say that there is no explanation of what alternative construction should have been adopted and why that alternative construction would not have given rise to a breach. They note that the Items in question were all breaches of numerous clauses, not just 4.1.1. As to the second point, they repeat the absence of any alternative construction which would have permitted a separate company to have access to ICA Programs which were licenced only to Winsopia. They also make the point that the issue is moot because of the court’s anterior findings about the nature and scope of ICA Programs.

I consider that this ground has no real prospect of success. I agree with IBM’s response. In particular:

This complaint is set out at [50]-[54] of the grounds of appeal and [140] – [147] of the skeleton. Again the complaint is that the judge’s approach was too narrow. At [50] of the grounds it is said that she was wrong to construe the language of the Licence Programme Specifications as imposing any limitation upon the class of potential recipients of such redistributed materials. At [51] it is said that she should have held that it granted a wide permission to the person redistributing a bound load module to do so without any enquiry to the identity or licence status of the recipient. It is said that it imposed no restriction on how, where or in what execution environment, recipients may use the redistributed materials. At [52] it is said that the judge made the same errors in relation to the terms of the Sample Program Licences.

At [53] and [54] of the grounds, it is said that the judge took a circular and erroneous approach to the proviso that “any reference to an IBM product, program or service should include functionally equivalent product, program and service that does not infringe any IBM intellectual property right”. It is said that the judge should have held that this language meant that references to z/OS or any other IBM product, program or service had to extend to a functionally equivalent product, such as the SDM. The argument is that since it is not suggested that the SDM infringed any IBM intellectual property right, the proviso did not exclude the SDM. The suggestion at [54] is that any restriction and/or clause 4.1.3(b) of the ICA “were void”. It is assumed this is a reference back to the Software Directive again.

IBM say that the judge’s interpretation of the permission notice at [190] was correct. They say that the appellants’ case that it imposed no restrictions is unsustainable because, amongst other things, that argument gave no effect to the proviso that all use under the notice must be “in connection with [the customer’s] authorised use of z/OS”, or the requirement that all relevant IBM copyright statements must be retained in any distributed materials, with the result that all recipients must be licensed to use the materials. There was also a separate requirement that the distributed materials must conform to the documented program interfaces with z/OS.

IBM note that no relevant error in the judge’s reasoning at [192]-[193] has ever been identified. The judge was correct to note the terms of the proviso. As to the SDM, the judge adopted no assumption as to whether the SDM infringed IBM intellectual property rights. The Software Directive point is rejected on the basis that the acts in issue were not acts of decompilation directed to interoperability, but instead a transfer out of enterprise of often large numbers of IBM CSECTs and other ICA programs.

I consider that ground 11 has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

The relevant part of the provision reads as follows:

“Unless otherwise required by applicable law without the possibility of contractual waiver or limitation, (i) neither party will bring a legal action, regardless of from, arising out of or related to this Agreement or any transaction under it more than two years after the cause of action arose; and (ii) after such time limit, any legal action arising out of this Agreement or any transaction under it and all respective rights related to any such action lapse.”

The judge considered various authorities and applied them at [996]-[997]. In addition, at [1000]-[1004] the judge noted that the authorities demonstrated that express words were needed to bar a claim that had been dishonestly or deliberately concealed.

The appellants’ complaints are at [55]-[58] of the grounds of appeal and [148]-[157] of the skeleton argument. It is said that the judge was wrong to interpret this clause as not applying to the claims in this case for unlawful procurement and unlawful means conspiracy against the immediate and ultimate owners of the counter party. In addition, the appellants say that the judge was also wrong to interpret the clause as meaning that there had been deliberate concealment until IBM discovered the concealment, in circumstances where the relevant fact – that LZLabs owned Winsopia - was not secret but in the public domain. At [56] it is said that the judge was wrong not to find that limb (ii) of clause 1.11.4 precluded IBM from bringing any secondary liability claims for procurement. [57] suggests that the judge was wrong not to determine (i.e. uphold) the appellants’ case that Winsopia had a sufficient interest to enforce clause 1.11.4 on behalf of LZLabs and Mr Moores.

IBM say that the judge’s construction was correct. They say that clause 1.11.4 could only affect the rights of IBM and Winsopia and therefore could not affect the position of third parties (LZLabs and Mr Moores), who were strangers to the ICA. They argue that that was supported by the authorities to which the judge referred. They also suggest that it would be contrary to common sense for a party who had deliberately concealed a wrong to be able to take advantage of clause 1.11.4.

I consider that Ground 12 has no real prospect of success. I agree with IBM’s response. In particular:

The argument is at [59] of the grounds of appeal and [158] of the skeleton argument. It is said that, in relation to each of the clauses, “the judge failed to identify the relevant ambiguity and thereby failed to apply the contra proferentem principle”.

I consider that this ground of appeal to be hopeless. The contra proferentem argument is incontestably bad. The appellants do not identify the necessary ambiguity in any of the provisions which would allow the principle even to be raised. In my view, in relation to each of these clauses, there was no such ambiguity: they were simple and straightforward provisions which the judge properly construed. There is therefore no basis on which the contra proferentem principle could be applied. Furthermore, in the 21^st Century, cases are now no longer decided solely by reference to this outdated rule of construction.

Grounds 14-17: Judge’s Specific Findings of Breach

At paragraph 74(3) of the consequentials judgment, the judge dealt with these grounds as follows:

“iii)

Grounds 16 to 35 in Sections (C), (D), (E), (F) and (G) are an attempt to challenge findings of fact on the technical breaches made by the Court having assessed the credibility and reliability of the witnesses who gave evidence against the contemporaneous documents deployed during the trial and the expert evidence. The technical breaches are considered and determined at paragraphs [273] to [844] of the Judgment.”

It is fair to note that, on the face of it, the number of grounds in relation to the findings of breach have been reduced since the application was made to the judge.

I agree with the judge’s reasons for refusing PTA. These grounds regularly seek to invoke some of the earlier grounds of appeal which I have already rejected. Much of the remainder comprise attacks on the judge’s findings of fact. These attacks are based on the assertion that the judge’s findings are “unsupportable”/“irrational”. However, there is no attempt to engage with the judge’s careful analysis of each of these Items by reference to the terms of the ICA, the expert evidence and in particular, the factual evidence. I consider that these grounds fail to clear the particularly high permission hurdle in this case. For all those reasons I consider that these grounds are, on analysis, hopeless.

The judge’s analysis of Item 1 is at [274]-[310] of the judgment.

These complaints are set out at [60]-[64] of the grounds of appeal and [160]-[167] of the skeleton argument. [60] of the grounds relies on the Ground 7 arguments and [61] relies on the Ground 6 arguments. [62] asserts that the judge was wrong to find that any breach of the ICA by Winsopia in relation to Item 1 had been procured by LZLabs. [63] asserts that the judge made unsupportable and irrational findings in relation to Mr Moores’ evidence. [64] suggests that the judge’s assessment of the changes to the SDM source code was irrational, unsupportable and contrary to the facts and expert evidence.

At the hearing on 20 June, I expressed a certain amount of surprise to Mr Stewart that this attack on the findings of fact of an experienced TCC judge was being maintained. He told me in robust terms that it was because the judge had misunderstood the error. Mr Stewart said that it was a loop on the IBM mainframe which kept running. Mr Moores was told about it and instructed it to be fixed.

IBM contend that there is no basis for an appeal. They say that this ground is an impermissible attack on the judge’s factual findings. They go on to say that the appellants’ case, that the IGZCUST was disassembled and the disassembled code shared among a wide group of individuals outside Winsopia (without access to its main frame for the purpose of correcting an error on that same main frame) was preposterous, and that judge was right to reject it on the facts.

I regard that Ground 14 as hopeless. I agree with IBM’s response. In particular:

The judgment addresses Item 2 at [311]-[345]. The analysis makes extensive reference to the expert and factual evidence.

The appellants’ case is set out at [65]-[67] of the grounds of appeal and [168]-[171] of the skeleton argument. [65] of the grounds relies on Ground 7. [66]-[67] criticise the judge’s conclusion on the evidence that it was “highly likely that the LMD processed IBM CSECTs and other code fragments”. It suggests that the judge’s conclusion was “rationally unsupportable, contrary to the fact and expert evidence before her, and wrong”. It is also said that this was inconsistent with a subsequent finding (which is dealt with under Ground 17).

IBM say that this is an impermissible attack on the judge’s factual findings and that she was justified in reaching the conclusions that she did across 34 paragraphs and 6 pages of the judgment.

Again, in my view, Ground 15 is hopeless. I agree with IBM. In particular:

Item 49 is dealt with in the judgment from [811]-[825].

The relevant grounds of appeal are set at [68]-[72] and expanded upon at [172]-[178] of the skeleton argument. [68] of the grounds is an assertion that the judge was wrong to find that any of the allegations concerned with Mr Taylor’s use of the Winsopia main frame related to ICA Programs. [69] suggests that the judge was wrong to reject the appellants’ case that his use of Winsopia’s main frame during his secondment meant he was within Winsopia’s enterprise. Grounds [70]-[72] seek to reactivate arguments about the Software Directive/CDPA.

IBM say that the judge was entitled to come to the factual conclusions that she did, largely because of Mr Taylor’s own evidence. The detailed reasons why the judge found against the appellants on this Item are encapsulated at [824], which emerged from cross-examination. The judge was entitled to rely on it. The Software Directive arguments must fail for the reasons given previously.

Again I consider this ground to be hopeless. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

The assertion is that, because the judge concluded that Item 39 was not a breach because “[it] was inadvertent and the problem was swiftly resolved”, she should have reached the same conclusion in respect of Items 34, 37, 38, 40, 41, 45 and 46. This is in the grounds at [73] and the skeleton argument at [179]

I again consider that Ground 17 is hopeless. In particular:

Grounds 18-23: Errors Relating to Unlawful Procurement.

In her consequentials judgment the judge dealt with these grounds (again making allowances for the renumbering):

“(iv)

Grounds 36 to 38 in Section (H) are a further attempt to challenge findings of fact as to knowledge and intention of the Defendants, based on the Court’s assessment of the credibility and reliability of the witnesses who gave evidence at trial. The case on unlawful procurement of breach is considered and determined at paragraphs [845] to [936] of the Judgment.”

I agree with the judge’s reasons for refusing PTA. One of the most important parts of this case were the allegations against LZ Labs and Mr Moores (and three other defendants), that they procured Winsopia’s breaches of the ICA. In section VI of the judgment, from [845]-[936], the judge rejected that case against the other three defendants but upheld it against LZLabs and Mr Moores. That section (of 90 paragraphs) contains a summary of the law and an assessment of the wrongful procurement case against each defendant by reference to all the relevant factual evidence. Much of that evidence came from the internal documentation of LZLabs and Mr Moores themselves, and in particular the latter’s repeated attempts to water down and weaken the code of conduct, which had originally been designed to avoid the very breaches that occurred.

No attack on those clear findings of fact is permissible. So none of these grounds of appeal have any real prospect of success. The judge’s analysis when rejecting these grounds was correct. None of them come close to clearing the particularly high permission hurdle in this case.

The appellants’ case is that, if there were no breaches of the ICA, neither LZLabs nor Mr Moores could be liable for procurement. That of course depends on the suggestion that any one of Grounds 1-17 above are arguable. Since they are not, this ground automatically fails.

The part of the judgment that is highlighted in particular by the appellants was [884]-[886]. These said:

“884.

Fourthly, the Software Directive arguments, rejected by this court, would not be sufficient to displace knowledge of the wrongfulness of Winsopia’s activities. As explained by Lord Leggatt in Lifestyle Equities, liability does not depend on whether the defendant knew that the act done by the primary wrongdoer was against the law; all that is required is knowledge of the essential facts which make the act unlawful. In this case, the essential facts are knowledge that the terms of the ICA prohibited the acts of reverse engineering and other misuse of ICA Programs that LzLabs directed Winsopia to carry out.

885.

Fifthly, Mr Rockmann is not entitled to rely on any belief that the Software Directive would provide a defence to the ICA breaches because he conceded that he did not have an understanding of the Software Directive. Further, no legal advice pursuant to which he might have acted has been produced in these proceedings.

886.

Although that would be sufficient to establish the requisite knowledge attributable to LzLabs, I find that Mr Cresswell also had knowledge of the essential facts which made the acts unlawful. He was CEO of LzLabs between 18 May 2015 and November 2020. Thereafter, his role was Executive Chairman, followed by Non-Executive Chairman. He was also a director at Winsopia between 2 December 2015 and 28 February 2022. As an officer of both LzLabs and Winsopia, he must have known the terms of the ICA, including the prohibitions on reverse engineering and other misuse of the ICA Programs. Indeed, in cross-examination he accepted this:

“Q. So, at least at a high level, you appreciated that the ICA prohibited reverse-engineering?

A. Yes, I think I mentioned that in my witness statement.

Q. And you understood that it prohibited individuals outside of Winsopia being given access to ICA Programs or using them?

A.

Yes. There was -- absolutely, yes.””

The appellants’ complaint is confined to [75] of the Grounds of Appeal. It is not addressed separately in the skeleton but worked in with all Grounds 18-26 at [180]-[196]. The argument is that the judge wrongly relied on Lifestyle Equities to decouple the alleged knowledge and belief LZLabs and Mr Moores had about the Software Directive from their knowledge or belief about the ICA; wrongly treated the essential facts as simply being an awareness of certain provisions of the ICA; and wrongly disregarded the principle that LZLabs and Mr Moores “could [not] be liable for procuring a breach of contract if they honestly believed the acts being carried out would not result in a breach.”

At the hearing on 20 June, Mr Stewart’s principal attack was that the judge was wrong to find the concession at [885].

IBM summarise the complaint as being essentially the rejection of the appellants’ case below, which was to the effect that i) they could not be liable in procurement because they believed that the Software Directive meant that Winsopia did not breach the ICA, and ii) it followed that they could not be liable in unlawful procurement.

IBM submit that i) was a matter of fact and the judge was entirely correct to conclude, on the evidence, that the relevant individuals had failed to establish that belief in the Software Directive as a matter of fact: see the whole of [880]-[993] of the judgment, not just the parts picked out by the appellants. Further and in any event, their belief in the Software Directive was irrelevant to their liability in procurement. Knowledge of the law is not relevant. The essential facts which LZLabs and Mr Moores knew were: a) the conduct on the part of Winsopia which they were procuring (such as decompilation or reverse assembly); and b) The terms of the ICA.

IBM say that whether or not LZLabs or Mr Moores believed that Winsopia was in breach of the ICA in the light of their understanding of the Software Directive was irrelevant. The judge correctly held that at [884].

In my view, Ground 19 has no real prospect of success. I agree with IBM’s response and consider that the contrary is unarguable. In particular:

The judge’s assessment is at [863]-[877] and [880]-[887] of the judgment.

This Ground is set out at [76]-[78] of the grounds and within [180]-[196] of the skeleton. The suggestion is that the judge was wrong to find that Mr Rockmann and Mr Cresswell (the directors of LZLabs) had the requisite knowledge and intention and that she should have found that they did not.

IBM note that LZLabs in fact procured Winsopia to commit acts that amounted to breaches of the ICA: see [863]-[877]. They note that that important finding is not challenged by the appellants. Thus, IBM say that this is all about the judge’s finding at [880]-[887] that Winsopia (acting through Rockmann and Cresswell) had sufficient knowledge of the terms of the ICA such that it knew that the essential facts which made the acts which it was procuring unlawful. That was a pure matter of fact and there was plenty of evidence to support the judge’s finding of awareness.

I consider that this ground has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

The judge’s assessment of this issue is at [917]-[936].

The appellants say at [79] of the grounds, and principally at [191] of the skeleton, that the judge was wrong to find that Mr Moores persuaded, encouraged or assisted Winsopia in its breaches of the ICA. IBM say that this is a matter of fact and that the central plank of the judge’s reasoning was that Mr Moores used his power and control over LZLabs – as its ultimate owner – to exert pressure on them to dismantle the protective barriers/code of conduct that had been put in place by the Clean Room process.

I consider this ground of appeal to be hopeless. I agree with IBM’s response and find the contrary unarguable. In particular:

This is set out in the grounds of appeal at [80]-[81] and the skeleton argument, primarily at [190]. It is a carbon copy of Ground 20 in respect of LZLabs. It fails for the same reasons that Ground 20 fails. In addition, the judge found that Mr Moores knew that the ICA prohibited decompilation, disassembly and translation and that his decision to disregard those terms was (on its own) sufficient to fix him with sufficient knowledge.

This is set out at [82] of the grounds of appeal and [191]-[195] of the skeleton. It is suggested that the judge should have found that Mr Moores, as the indirect beneficial owner of Winsopia, was entitled to rely on the principle in Said v Butt.

In response, IBM say that the Said v Butt argument was expressly limited to Mr Cresswell and Mr Rockmann and the appellants never suggested that it applied to Mr Moores as well. It is much too late for that point to be taken because it would have then necessitated a factual inquiry and cross examination at trial.

I refuse permission to the appellants to take this point. Mr Stewart rightly accepted that it was a new argument. It seems to me that there would be clear prejudice if the appellants were permitted to take the point now. First, the inevitable factual enquiry about the nature, scope and extent of Mr Moores’ precise interest (which would have been necessitated by such a plea) simply never took place at trial. Secondly, it would have been necessary for Mr Moores to demonstrate that he was acting in good faith (see [851] in respect of Mr Cresswell and Mr Rockmann). That would have also necessitated cross-examination (again see [913] in respect of the directors). In the absence of such evidence, the prejudice to IBM if the point was allowed in now, would be significant and irredeemable.

There was no suggestion before the judge that the Said v Butt point even arguably applied to Mr Moores. In any event, since it has never been suggested that Mr Moores was an agent of Winsopia, it seems to me that the Said v Butt point could not run in his case in any event.

For these reasons, Ground 23 has no real prospect of success.

Grounds 24-26: Unlawful Means Conspiracy

At paragraph [74(v)] of her consequentials judgment, the judge said this:

“v)

Grounds 39 to 41 in Section (I) are a further attempt to challenge findings of fact as to the combination, knowledge and intention of the Defendants, based on the Court’s assessment of the credibility and reliability of the witnesses who gave evidence at trial. The case on unlawful means conspiracy is considered and determined at paragraphs [937] to [960] of the Judgment.”

I agree with the judge’s reasons for refusing PTA. Since the judge found that the same two appellants (LZLabs and Mr Moores) were liable for unlawful means conspiracy as were liable for the unlawful procurement, the unlawful means conspiracy part of the case adds nothing. Thus this part of the appeal is academic. It also suffers from the same fundamental flaws as Grounds 18-23, and fails to clear the particularly high permission hurdle in this case.

As made plain by [83] of the grounds of appeal, this ground is parasitic upon success on the previous 23 Grounds. If they fail, so does Ground 24. I have found that they disclose no real prospect of success. That therefore is the end of this ground too.

This is set out in the grounds of appeal at [84]-[87] and the skeleton (along with the related arguments about unlawful procurement) at [180]-[196]. The essential argument advanced by the appellants is similar to their procurement case, namely that they believed that their conduct was not unlawful because they believed the Software Directive applied and so it followed that they were not liable in unlawful means conspiracy.

IBM say that these are matters of fact, in particular the judge’s factual finding that neither LZLabs nor Mr Moores established the belief that the Software Directive applied. They also say that even if that belief could be established that would not be relevant to their liability in unlawful means conspiracy as a matter of law, as explained by the judge at [949]-[952].

I consider that Ground 25 has no real prospect of success. I agree with IBM’s response. In particular:

It is suggested at [88]-[89] of the grounds of appeal and [196] of the skeleton argument that the judge was wrong to find liability for these torts in circumstances where there was no damage. It is said that the evidence from Mr Wallin as to damage was extremely thin and the best he could do was identify lost management time.

IBM’s response is to identify the judge’s findings of fact at [959] that there had been some damage and that its quantum was to be investigated at stage two of the trial. They point to the fact that, in preparation for the second stage of trial, IBM’s particulars of claim identify a claim in the order of £1.6 million. Management time is one, but only one, of the heads of loss.

I agree with IBM’s response. The judge concluded as a matter of fact that, on the evidence, there had been damage. In my view, that is sufficient to make Ground 26 unarguable. Further and in any event, the appellants knew that the precise quantification of damage remained for the second part of the trial.

Grounds 27-29: Errors in Relation to Limitation and Concealment

At paragraph [74(vii)(viii)] the judge said:

“vii)

Ground 43 in Section (K) seeks to challenge the Court’s construction of clause 1.11.4 of the ICA but the arguments made were considered in the Judgment and rejected at paragraphs [991] to [1007]. The Defendants have not identified any authority in which a court at first instance, or on appeal, reached a different conclusion on a similar issue of construction.

viii)

Grounds 44 and 45 in Section (K) seek to challenge the Court’s findings of fact on deliberate concealment and knowledge, considered at paragraphs [1013] to [1118] of the Judgment, but do not identify any error in the legal principles set out and fail to engage with the detailed evidence set out in that part of the Judgment in support of the findings.”

Again, making allowances for some change between the draft grounds before the judge and the grounds with which we are concerned, these paragraphs apply to these grounds. Of course, I have dealt with the point about clause 1.11.4 of the ICA above.

I agree with the judge’s reasoning for refusing PTA. It seems to me that all the grounds in this group are, in one way or another, an impermissible attack on the judge’s findings of fact. They must all fail for that reason alone. The particularly high permission hurdle in this case is again of direct relevance; I can see no way in which the appellants can get close to surmounting that hurdle.

At [90] of the grounds of appeal it is said that there can have been no deliberate concealment of the fact that LZLabs owned Winsopia because that was in the public domain. This is developed in the skeleton at [198] – [206]. At [91] of the grounds, it is said that the judge was wrong to find that the breaches were deliberate, in circumstances in which they were unlikely to be discovered for some time. In his oral submissions at the hearing at 20 June, Mr Stewart relied on [96]-[99] of the judgment in Canada Square, to say that one ingredient of deliberate concealment was that there had to be an intention to keep something secret, an initial hiding.

IBM accept that they could have discovered that LZLabs owned Winsopia earlier, but only if they had any reason to make those enquiries. They say they had no such reason, so the point does not run. They rely on the judge’s detailed analysis of concealment at [1022]-[1068] to say that this argument is not open to the appellants on appeal. As to the point that the breaches were deliberate because they were unlikely to be discovered for some time, IBM make the straightforward point that the judge was plainly right because, in fact, it took seven years for the breaches to be detected.

I consider that Ground 27 has no real prospect of success. I agree with IBM’s response and I consider the contrary to be unarguable. In particular:

At [92] of the grounds of appeal and [201]-[201] of the skeleton, the appellants say that the judge should have found Mr Anzani’s knowledge should be attributed to IBM. At [93] they say that the judge should not have decided that IBM was unaware of the connection between LZLabs and Winsopia prior to August 2020 and should have instead drawn adverse inferences against IBM and/or found that they had not discharged the burden of proof. At [94] the appellants contend that the judge erred in finding that IBM could not have found out about the concealment of the unlawful means conspiracy prior to June 2023. It is said that the judge’s reasons were inadequate so her finding was unsupportable.

In response, IBM say that Mr Anzani was not an employee and not an agent and that it would therefore have been wrong in principle for the judge to find that the knowledge of a parent company should be attributed to the subsidiary. As to the point at [93], IBM say that the judge accepted IBM’s evidence as to when they became aware of the connection, and that the documents supported that evidence (as the judge found). They say there is no basis to redo that factual inquiry. They make the same point in relation to the finding about unlawful means conspiracy prior to June 2023.

I consider this Ground 28 has no real prospect of success. I agree with IBM’s response. In particular:

The judge’s judgment on this is set out at [1119] as follows:

“1119.

There is a dispute between the parties as to the date on which any damage occurred so as to give rise to a cause of action in the tort of unlawful means conspiracy. The competing dates are 2016 (the defendants) and 2021 (IBM). The evidence before the court on the date of damage is sparse, not least because issues of quantum were not included in this trial. I do not consider that it is necessary to resolve this matter because on the facts it does not affect any limitation defence or the key findings on liability. The only claim that might be affected would be IBM’s claim against Winsopia for unlawful means conspiracy, which might be caught by the contractual limitation period of two years, subject to the issue of dishonest or deliberate concealment. However, based on my Judgment dated 29 November 2023, when giving permission to IBM to amend to plead the unlawful conspiracy claim, I am satisfied that, as a result of late disclosure, IBM could not with reasonable diligence have discovered the concealment regarding the unlawful conspiracy claim prior to June 2023, when that claim was formulated.”

In the grounds of appeal at [95] and the skeleton at [207], this unremarkable statement is attacked. It is said that the judge should have determined this issue and should have found that there was either no damage or that the cause of action arose no later than March 2016.

I consider that Ground 29 is hopeless. Indeed, I regard it as going beyond the boundaries of a properly considered ground of appeal. The judge had produced an enormous judgment covering a huge range of interleaved issues. She declined to deal with this particular issue because it was academic. To criticise that conclusion perhaps gives some clue as to the true nature and purpose of this ‘kitchen sink’ PTA application. In any event, the damage point is not open to the appellants because it is an attack on the judge’s findings of fact. The argument that the cause of action accrued in March 2016 is untenable in the light of the judge’s explanation of concealment and knowledge.

Grounds 30-31 Errors Relating to Audit and Termination.

At [74(ix)] of the consequentials judgment, the judge said:

“ix)

Grounds 47 to 49 in Section (L) seek to challenge the Court’s findings of fact on the audit and termination issues considered at paragraphs [961] to [987] of the Judgment but do not identify any error of principle or law.”

I agree with the judge’s reasons for refusing PTA. At trial, IBM’s case was that Winsopia refused their reasonable request for an audit. Winsopia’s answer was to say that, because there was confidential material involved, some sort of separate agreement was necessary. The judge dealt with all of the arguments in the judgment at [961]-[987]. She concluded that the request for an audit was reasonable and that in the circumstances the termination was lawful. In addition, IBM had a separate claim, based on the breaches of contract found by the judge, which either justified termination under the contract or comprised repudiatory breaches of contract.

Accordingly, it follows that, as with most cases concerned with termination/repudiation, the result turned on the judge’s evaluation and findings of fact. Again that means that these grounds of appeal must fail, principally because they are an illegitimate attempt to reopen those findings of fact and because they do not clear the particularly high permission hurdle in this case.

This is dealt with in grounds of appeal document at [96] and the skeleton at [208]-[212]. There are two complaints. The first is that the judge was wrong to find breach because of the ‘confidential information’ point. The second is that IBM won on an unpleaded case because their pleaded reliance was on IBM’s letter of 3 December 2010 and so they should not have been able to rely on their letter of 12 January 2011.

In response, IBM say that the judge was entitled to reach the conclusion she did about confidential information, so that this is an attack on the judge’s findings of fact. She found that “much of the material sought was not confidential” so that was a complete answer to the defence in any event. As to the pleading point, they say that the letter of 12 January was pleaded at paragraph 38 of the RRRAPOC and that in any event the letter of January only repeated what was in the letter of 3 December.

In my view, Ground 30 has no real prospect of success. I agree with IBM’s response. In particular:

The relevant paragraphs in the Grounds of Appeal are [97]-[99] and the skeleton at [213]-[217]. [97] is parasitic upon success of all of the earlier grounds. [98] asserts that the judge was wrong to say that Winsopia had been given reasonable notice and an opportunity to comply. [99] says that the judge was wrong to find that IBM was entitled to terminate at common law.

IBM make the point that [97] must fail because the other grounds on which it relies have also failed. They say that they were entitled to terminate under the contract or by way of repudiatory breach, just as the judge found. That latter claim was not capable of remediation because the ICA Programs had been transferred to a third party. As to the audit request, they rely on the fact that the judge concluded that Winsopia had more than enough time to comply with that request.

I consider this ground of appeal has no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

Grounds 32-34: Consequential Judgment and Relief

The injunction and other relief granted by the judge in her consequentials judgment seem to me to be the inevitable consequence of the appellant’s failure in the main action. The injunctions granted were all a matter of discretion for the judge who spent many weeks hearing and subsequently evaluating the evidence, and then another three days at the consequentials hearing. Her conclusions are not wrong in principle, so this court would not interfere with the exercise of her discretion. The particularly high permission hurdle noted above applies with even more force because this is an attack on a pure exercise of the specialist judge’s discretion.

These are dealt with in the appellants grounds of appeal at [100]-[106] and the skeleton at [218]-[228]. Ground 32 appears to rely on a number of grounds which have already been rejected and a number of arguments which I have already concluded were unarguable such as the complaint about no literal copy of any source code or object code having been made.

Ground 33 complains that it was wrong to grant an injunction to restrain the use or dissemination of pure information which was not established to be protected by any intellectual property rights.

IBM repeat their principal point that it is irrelevant whether any component of an ICA Program was actually copied into the SDM. They maintain that what was relevant was that the appellant’s systematic breaches were used to “develop or improve performance of the SDM” (as the judge found in the consequential judgment at [22]). In consequence the SDM was tainted. Any parts of the SDM that were not tainted would not be subject to the injunction but the appellants provided no proper evidence to demonstrate whether the tainted parts of the SDM could be separated from untainted parts. The appellants’ proposals were described by the judge as “woefully inadequate” and “not credible” at [23]. As to Ground 33, IBM’s point it that the property rights are immaterial and all that mattered was the use of information unlawfully harvested from ICA Programs.

Grounds 32 and 33 have no real prospect of success. I agree with IBM’s response and consider the contrary to be unarguable. In particular:

However, it emerged during the hearing on 20 June that neither Lord Justice Males nor I were very enthusiastic about the formulation of the injunction in respect of derived information. That is because it referenced information “which derives from breaches of the ICA”. That seems to us to be open to all sorts of argument, given the bitter litigation thus far. We suggested a change to reflect this concern, and IBM in their email of 3 July 2025 agreed to that change. We therefore amend that injunction to refer to information “which derives from breaches of the ICA as found by the judge in her judgment”.

This is set out in the Grounds of Appeal at [107]-[108] and the skeleton argument at [229]. It is said that it was wrong to grant the injunction against LZLabs and Mr Moores, or if it was, it should have been more limited. IBM say that the judge was fully entitled to require delivery up against the background of widespread breaches in order to protect against further misuse. There was no justification for the appellants continuing to hold ICA Programs.

In my view, this was plainly a matter that the judge was entitled to consider and rule on when ordering delivery up of the ICA Programs. Any other result would have allowed the appellants to keep the product of their unlawful conduct. Ground 34 is therefore unarguable.

The judge concluded that the appellants were liable for the costs of the proceedings as a whole including the costs incurred by IBM in pursuing the claims against LZLabs Limited (the UK company), Mr Creswell and Mr Rockmann. The appellants complain about that conclusion.

There can be nothing in this complaint. Costs are a paradigm example of the exercise of a judge’s discretion. It is common for defendants who have taken every point and lost every point to be ordered to pay the claimant’s successful costs of pursuing all the defendants, including in this case three defendants who, although closely related to the relevant parties and events, were found not to be liable. It is a form of Sanderson order: see White Book, para 44.2.28.

In consequence, this ground of appeal is refused. It has no real prospect of success.

Conclusions

For all these reasons, this application for permission to appeal is refused. Notwithstanding the length of all the documents (including these reasons) the ultimate position is a simple one. There are no free-standing points of principle that the judge got even arguably wrong. Instead, as she herself said, she applied the settled law to the facts as she found them to be. It appears that the appellants were aware from the outset of the risk of IBM pursuing them when they discovered what they were doing. After many years, that risk eventuated before the judge. None of the arguments in the grounds of appeal document, or the long accompanying skeleton can, in my view, put off the consequences of that risk-taking any longer.

I agree that permission to appeal must be refused for the reasons given by Lord Justice Coulson.

When I first read the judgment in this case it struck me as a thorough and careful piece of work by a specialist judge with a firm command of the evidence given during a trial which lasted 38 days (including her five reading days) and which must have required extensive further reading out of court. It was therefore surprising to see an appellant’s notice which insisted that the judge got almost everything wrong. The appellants’ 35 grounds of appeal and 80 pages of skeleton argument have confirmed, rather than shaken, my initial view. They illustrate powerfully the truth of the old adage that less is more and its converse, which could usefully be borne in mind by those seeking permission to appeal in this kind of case, that very often more is less.