SAS Institute Inc v World Programming Ltd

The Claimant, SAS Institute Inc (“SAS Institute”), is a developer of analytical software known as SAS (referred to in these proceedings as “the SAS System”). The SAS System is an integrated set of programs which enables users to carry out a wide range of data processing and analysis tasks, and in particular statistical analysis. The core component of the SAS System is Base SAS, which enables users to write and run application programs (also known as “scripts”) to manipulate data. Such applications are written in a language known as the SAS Language. The functionality of Base SAS may be extended by the use of additional components, including three which are relevant to these proceedings called SAS/ACCESS, SAS/GRAPH and SAS/STAT (the four components being collectively referred to as “the SAS Components”). The SAS System has been developed over a period of 35 years. It is a very valuable asset, producing revenue for SAS Institute of $2.3 billion in 2009.

Over the years SAS Institute’s customers have written, or had written on their behalf, thousands of application programs in the SAS Language. These can range from fairly short and simple programs to large and complex programs which involve many man years of effort to create. Prior to the events giving rise to this dispute, SAS Institute’s customers had no alternative to continuing to license use of the necessary components in the SAS System in order to be able to run their existing SAS Language application programs, as well to create new ones. While there are many other suppliers of analytical software which compete with SAS Institute, a customer who wanted to change over to another supplier’s software would be faced with re-writing its existing application programs in a different language.

The Defendant, World Programming Ltd (“WPL”), perceived that there would be a market demand for alternative software which would be able to execute application programs written in the SAS Language. WPL therefore created a product called World Programming System or WPS to do this. In developing WPS, WPL sought to emulate much of the functionality of the SAS Components as closely as possible in the sense that, subject to only a few minor exceptions, it tried to ensure that the same inputs would produce the same outputs. This was so as to ensure that WPL’s customers’ application programs executed in the same manner when run on WPS as on the SAS Components. There is no suggestion that in doing so WPL had access to the source code of the SAS Components or that WPL have copied any of the text of the source code of the SAS Components or that WPL have copied any of the structural design of the source code of the SAS Components.

According to two previous decisions of the courts in this country, it is not an infringement of the copyright in the source code of a computer program for a competitor of the copyright owner to study how the program functions and then to write its own program to emulate that functionality. SAS Institute challenges this view of the law. Even if it is correct, however, SAS Institute contends that WPL has both committed a series of infringements of copyright and acted in breach of contract in creating WPS and its accompanying documentation as a result of the way in which WPL has gone about those tasks.

SAS Institute’s principal claims are as follows:

SAS Institute also advances another claim concerning use of the Full Edition of the SAS System which has been ordered to be tried on a later occasion.

The claim was commenced by SAS Institute on 14 September 2009. At that time the only claim advanced by SAS Institute was the third claim listed in paragraph 5 above. Subsequently WPL applied for a speedy trial and SAS Institute applied to expand its claims. On 16 December 2009 Philip Marshall QC sitting as a Deputy High Court Judge gave directions which both gave SAS Institute permission to amend its Claim Form and Particulars of Claim and expedited the trial to the first available date after 28 May 2010. Subsequently the trial was fixed for the second week of June 2010.

On 16 April 2010 Lewison J made a case management order to enable that trial date to be adhered to. Among other things, that order limited SAS Institute to (a) 31 specific examples of similarities between WPS on the one hand and the SAS Manuals or the SAS Components on the other hand and (b) 51 specific examples of similarities between the WPS Manual and the SAS Manuals. By virtue of the order, such similarities are to be taken to be representative samples of the similarities between the respective systems and manuals. Apart from one similarity which had already been identified, the order required SAS Institute to identify the other 30 similarities between WPS and SAS as soon as possible and in any event by 14 May 2010. It also required SAS Institute to serve its expert’s report by 14 May 2010 and WPL to serve its expert’s report by 21 May 2010.

In the event the other 30 similarities were only identified in exhibit PAI11 to Professor Ivey’s report served on 14 May 2010. Counsel for SAS Institute explained that this was because the decision as to which similarities to rely upon was only taken shortly before service. Professor Ivey explained that inspecting the source code of WPS, and identifying similarities to the SAS Components and SAS Manuals, was a substantial task.

I mention these matters for two reasons. The first is that (subject to certain rulings which I gave on the first two days of trial) Lewison J’s order defined the scope of the trial. The second is that it is necessary to be bear this chronology in mind when considering the criticisms made by each side of the other’s expert witness which are discussed below.

SAS Institute called the following factual witnesses: Dr James Goodnight, SAS Institute’s Chief Executive Officer; Kathryn Council, Vice President of SAS Institute’s Publications Division; Alan Eaton, Senior R&D Director at SAS Institute; and Sonia Sparkes, Senior Contracts Officer for SAS Software Ltd, the UK subsidiary of SAS Institute. None of these witnesses’ evidence was seriously challenged. In addition, SAS Institute served witness statements from two witnesses under hearsay notices, namely Dr David DeLong, a retired academic who worked for SAS Institute from 1979 to 1993, and Cynthia Puryear, a Marketing Operations Specialist employed by SAS Institute.

WPL called the following factual witnesses: Anthony Brent, Graham Clemow, Martin Jupp, Dr Peter Quarendon, Thomas Quarendon, Oliver Robinson, Benjamin Scurr, Declan Vibert and Kevin Weekes. Apart from Mr Clemow, who is a contractor engaged by WPL via his own company, these are all employees of WPL. Many are also shareholders in WPL, and some are directors. With the exception of Mr Robinson, who is WPL’s operational manager, and Mr Brent, who has helped write the WPS Manual, they are the principal programmers who have developed WPS. They were all frank and straightforward witnesses. The only areas of WPL’s witnesses’ evidence which were seriously challenged were as to (i) the use made of the SAS Manuals in writing the WPS Manuals and (ii) the use made of the Learning Edition for performance testing. I shall consider this evidence below. In addition, WPL served a witness statement from one witness under a hearsay notice, namely Professor Frank Harrell, an academic who contributed to early versions of parts of Base SAS.

SAS Institute’s expert witness was Professor Peter Ivey. From 1989 to 2001 he was Professor and Head of the Electronics Systems Group in the Department of Electronic and Electrical Engineering at the University of Sheffield. He continues to work at Sheffield on a part-time basis and is also an Honorary Professor at the University of Durham. Since 2001 he has been a director of Innotec Ltd, a consulting firm. He has over 35 years’ experience in the ICT industry.

WPL’s expert witness was Dr Robert Worden. From 1978 to 1997 he was employed in various different capacities by Logica, including four years as managing director of Logica Cambridge. Since 1997 he has been employed by Charteris plc, a consulting firm. He too has some 35 years’ experience in the ICT industry. He also has considerable experience as an expert witness in ICT cases.

Both experts were required to prepare lengthy and detailed reports on complex subject matter within an extremely short timescale. As a result, as they both frankly acknowledged, they were both forced to take certain shortcuts. Nevertheless, I found the evidence of both experts to be of assistance. The fact that they took rather different approaches to their task contributed to this, since it gave me the benefit of differing perspectives on the issues.

Counsel for WPL attacked Professor Ivey as having “exhibited strong and unremitting bias in his written and oral evidence”. This accusation was not put to Professor Ivey in cross-examination, as it ought to have been. Even disregarding that, I do not consider that the accusation is justified.

Counsel for WPL made three main submissions in support of the accusation. The first was based on the following criticism made by Pumfrey J (as he then was) in Navitaire Inc v easyJet Airline Co Ltd [2004] EWHC 1725 (Ch), [2006] RPC 3 at [9]:

“Procedurally, there has been one important problem with the case. By the order of Master Moncaster made February 18, 2003, Navitaire were required to provide complete particulars of each and every similarity between OpenRes and eRes relied upon in support of its allegations of infringement. Rather than supply such particulars, Navitaire's advisers required Navitaire's expert witness, Dr Gillian Hunt, to prepare an expert report that concerned itself only with the similarities between the systems. It said nothing about the differences and it did not provide alternative explanations for features alleged to be copied. It was in effect the indictment. This was not compliance with the Master's order, and it was not fair to Dr Hunt, who prepared an entirely one-sided document, to the extent even that if a help message contained some text similar and some text different from that in the Navitaire system, the similar part only was referred to. The report turned out to be an embarrassment, as it contained material that was not placed in context by any other report (Dr Hunt ended up making seven reports), and its nature ensured that it contained no nuance.”

Although this paragraph was directed more at Dr Hunt’s instructions than at Dr Hunt, elsewhere in the judgment Pumfrey J was more directly critical of Dr Hunt herself: see [139], [170]-[171], [203] and especially [208]-[210].

Counsel for WPL submitted that, just as in Navitaire, Professor Ivey’s report had focussed exclusively on similarities rather than differences and was an indictment. Counsel for SAS Institute responded that, while it was true that Professor Ivey had focussed on the 31 similarities, this was in accordance with his instructions and with Lewison J’s order of 16 April 2010. In addition, he pointed out that there was a difference between this case and Navitaire in that derivation was not in issue in this case, whereas it was in issue in relation to some of the claims made in Navitaire.

I have to say that I do not think Professor Ivey should have been placed in the position he was. As Pumfrey J made clear, it is not the function of an expert’s report to set out particulars of similarities. Particulars of similarities should be set out in a statement of case, and then considered by the respective expert witnesses. In the circumstances of the present case, however, I do not consider that either Professor Ivey or those instructing him can be criticised for proceeding in the way that they did. Professor Ivey did as he was instructed. Those instructing him had little practical alternative. As a result of WPL’s insistence upon the case being tried speedily, Lewison J made an order which required those representing SAS Institute to identify the similarities relied upon within 4 weeks and to serve an expert’s report within the same period. The only people with sufficient technical knowledge to conduct the exercise of identifying similarities who had access to the WPS source code were Dr Ivey and his assistants, and they were only permitted to inspect the source code at the premises of WPL’s solicitors. Professor Ivey explained that this exercise required a great deal of work which had to be curtailed due to the court-imposed deadline. Not only was he instructed to focus on finding similarities, but also he did not have time to do more.

The second main submission made by counsel for WPL was that Professor Ivey had failed to differentiate between emulating the functions of the SAS System on the one hand and copying which went beyond emulating the functions of the SAS System on the other hand, and had used legal terminology such as “adaptation” in a way which positively obscured the distinction. In this connection, he relied on what Pumfrey J had said in another case about computer software, Cantor Fitzgerald International v Tradition (UK) Ltd [2000] RPC 95 at [70]:

“A number of judges have protested at the willingness of some experts to exploit the ignorance of the court on matters within their special expertise. … Where the subject matter of the action lies in a highly technical area it is of particular importance that the expert is scrupulous in putting forward all relevant considerations which occur to him or her as being relevant to the issue to be decided. The court has no points of reference other than those provided by the expert. It is reprehensible for the expert to hold back relevant information. The danger is manifest. If both experts lack objectivity the court is deprived of any proper basis to arrive at a decision. …”

In my opinion Professor Ivey did not distinguish as clearly as he might have done between emulating the functions of the SAS System and copying which went beyond emulating the functions of the SAS System, but I see no reason to criticise Professor Ivey for this. In my view it is clear that he approached the matter in accordance with his instructions, and while this distinction is central to WPL’s case, it is not one which SAS Institute accept, as I shall discuss below. As to the language which Professor Ivey used in his report, he explained that he used words like “adaptation” in their ordinary English meanings, and not as terms of legal art. I accept that explanation, but I think it is unfortunate that Professor Ivey was not instructed to avoid such terminology. (It is also regrettable that he was instructed to provide a technical explanation of terminology used in the EU Software Directive, evidence that in my opinion was neither admissible nor helpful.)

The third main submission made by counsel for WPL was that Professor Ivey had expressed opinions in support of SAS Institute’s case without any proper technical basis and in some cases which fell outside the scope of his expertise. It is true to say that cross-examination revealed that Professor Ivey did not have a full understanding of certain matters addressed in his reports. I accept his explanation, however, that he had done his best in the limited time available. Moreover, he did not claim to be an expert in statistics.

Although I am not persuaded that Professor Ivey was biased or acted as an advocate for SAS Institute, I do consider that some of his evidence was not as careful as it should have been, even making every allowance for the time pressure he was under. The aspect which most concerned me was his evidence in relation to what he described in paragraph 237 of his report as “decompilation (reverse engineering) of SAS proprietary files”. I think that Professor Ivey should have known better than to use the expression “decompilation” in this context even without instruction, and should have appreciated that it was incumbent on him to explain precisely what he meant by “reverse engineering” in this context. Furthermore, I think that Professor Ivey should have known better that to assert, as he did in paragraph 250 of his report:

“The SAS file formats must be detailed in the SAS source code, in order for the SAS System to be able to generate files in those formats. The whole point of the reverse engineering exercise undertaken by WPL is to replicate the operation of that SAS source code in WPS.”

That evidence was capable of seriously misleading a non-technical judge. There is in fact no evidence that the SAS source code does “detail” the relevant SAS data file formats, as opposed to reading and writing files in those formats. Professor Ivey not only did not exhibit the relevant SAS source code to his report, but also had not even inspected it. I do not consider, however, that this means that the remainder of Professor Ivey’s evidence should be rejected.

For his part, counsel for SAS Institute criticised Dr Worden’s evidence on a number of disparate bases. The first was that Dr Worden was a generalist with no prior experience of the SAS System. In my view prior experience of the SAS System was not a pre-requisite for giving expert evidence in this case, and Dr Worden did have a good general knowledge of statistics and statistical software.

The second criticism was that Dr Worden’s report was based upon testing a “null hypothesis” and contrasting it with a “counter-hypothesis” which he defined as follows:

“Null Hypothesis: In their design, development, and testing, WPL were building a system to emulate parts of the external behaviour of SAS as closely as they could; and they employed only competent modern software engineering techniques to do so. They used the SAS materials at their disposal solely to try to determine the external behaviour of SAS. The resemblances between WPS and SAS material were caused solely by WPS employing modern software engineering techniques to build software that performs some of the same functions as the SAS software.

Counter-Hypothesis: Rather than employ competent modern software engineering techniques, WPL sometimes preferred to make copies in some form of the SAS materials available to them and were able by some means to deduce the structure of the SAS source code from those materials.”

Counsel for SAS Institute submitted that it was unclear what the null hypothesis and the counter-hypothesis actually were. I disagree. In my view the definitions set out in Dr Worden’s report are tolerably clear. It is fair to say that, as counsel’s cross-examination of Dr Worden demonstrated, there is considerable room for debate as to where one draws the line between the two, and in particular as to what amounts to “competent modern software engineering techniques” and what amounts to copying which goes beyond emulating external behaviour. Nevertheless, I consider that the distinction Dr Worden was attempting to draw is in principle a relevant one.

The third criticism was that Dr Worden had made some attempts to quantify the extent of certain similarities which were flawed and subjective. Dr Worden himself pointed out in his report that there was an element of subjectivity involved, and he readily accepted the flaws in his analysis which were pointed out to him in cross-examination. While I accept that this means that Dr Worden’s attempts at quantification have to be treated with caution, I do not consider that this devalues the remainder of his evidence.

The fourth criticism was that Dr Worden’s understanding of the facts was at variance with what the evidence of the factual witnesses showed to be the case. This criticism overlooks the fact that Dr Worden’s report was prepared under considerable time pressure and before both SAS Institute’s and WPL’s factual witnesses had served a number of additional witness statements, let alone their oral evidence. In these circumstances, it is not in the least surprising that Dr Worden’s understanding sometimes varied from what the factual evidence shows.

SAS Institute was founded by Dr Goodnight and other colleagues from North Carolina State University in 1976. It has grown to have over 11,000 employees in 55 countries. It has over 45,000 customers to whom it supplies over 200 software products. It has approximately 9.2% of the business analytics software market. Its current R&D investment is 23% of revenue.

Many of SAS Institute’s developers have expertise in statistics as well as in computer software. Eight of its employees are Fellows of the American Statistical Association, and a Senior Director of R&D was recently elected as President of the ASA.

I have briefly described the SAS System above. It is available for use on a variety of different platforms ranging from IBM mainframe computers to personal computers running Microsoft Windows. SAS Institute market both a “Full Edition” and a “Learning Edition”. The latter, as its name implies, is an edition of the software which enables users to learn how to use the SAS System. The Full Edition is generally licensed on an annual basis. The Learning Edition is licensed for use until a fixed expiry date.

As is common with software, SAS Institute periodically releases new versions of the SAS System. Each new version typically contains new or enhanced functionality compared to its predecessor. New functions may be included as a result of internal research and development or in response to customer requests or in response to competitors’ offerings. It follows that the functionality of each of the SAS Components has grown substantially over the years.

Although many of the statistical analyses and computations performed by the SAS System are based on methods published by others in the academic statistics literature, some have been devised by SAS Institute’s own employees.

As noted above, there are four SAS Components in issue:

As I have said, SAS application programs or scripts are written in the SAS Language. This consists of statements, expressions, options, formats and functions. These elements are expressed in tokens, that is to say strings of characters used in accordance with certain conventions. One of the main types of token is names, which may be elements of the SAS Language such as procedures, e.g. LOGISTIC and UNIVARIATE, or user-defined. Certain words and symbols are reserved to the SAS Language. Reserved words are known as “keywords”.

A SAS script consists of two main types of program steps, known as DATA steps and PROC steps. The DATA step is the primary method for creating or manipulating a SAS data set. A DATA step is a group of SAS Language statements that begins with a DATA statement and contains programming statements that create SAS data sets from raw data files or manipulate existing SAS data sets. SAS data sets consist of data stored in particular formats recognised by the SAS System. A SAS data set consists of two key parts, descriptor information and data values. The descriptor information describes the contents of the SAS data set to the SAS System, and the data values are data which have been collected or calculated. The data are organised into a table, which consists of rows, called observations, and columns, called variables. An observation is a collection of data values that usually relate to a single object, and a variable is the set of data values that describe a given characteristic. This is illustrated below:

DATA step statements may be either executable or declarative statements. Executable statements result in some action during individual iterations of the DATA step, while declarative statements supply information to SAS and take effect when the system compiles program statements. Executable statements include CALL, DO, IF and IF-THEN/ELSE. CALL routines alter variable values or perform other system functions. Declarative statements include FORMAT and INFORMAT. FORMAT statements associate formats (instructions to control the written appearance of data values) with variables. INFORMAT statements associate informats (instructions to read data values into variables) with variables. In addition, there are various SAS functions which can be used either in DATA step statements or in other ways.

A PROC step is a SAS Language statement beginning with a PROC statement. In the PROC step the user can invoke various procedures or “PROCs” by name. These PROCs are software tools written by SAS Institute to perform a wide variety of particular types of data analysis and reporting. PROCs process and analyse data in SAS data sets to produce statistics, tables, reports, charts, plots and so on. There are well over 300 PROCs. Each PROC represents a significant body of work in its own right, including research into the best approach to statistical analysis. Choices need to be made as to what features to introduce, what syntax to use, what statements and options are appropriate for a particular procedure, and what output should be generated by the new procedure.

Two procedures which featured quite extensively in the evidence are PROC LOGISTIC and PROC UNIVARIATE. It is convenient to take these in reverse order.

PROC UNIVARIATE carries out statistical calculations on distributions related to single variables, such as, for example, the age distribution of the onset of a particular disease. PROC UNIVARIATE can plot the distribution, calculate the mean, plot histograms and so on. The variety and sophistication of the options available to the user with this procedure are illustrated by the fact that the description of the operation and use of PROC UNIVARIATE in version 9.2 of the SAS System occupies 248 pages of detailed description in the relevant SAS Manual.

PROC LOGISTIC is a rather more sophisticated procedure than PROC UNIVARIATE. In simple terms, it fits a linear model to all the variables in a set of data. To take an example given by Dr Worden, suppose one has a data set for a group of people that gives their alcohol consumption, food consumption and whether or not they have gout. The purpose of PROC LOGISTIC is to take the data set and to produce a rule which says: if this is your alcohol consumption and this is your food consumption, here is your probability of getting gout. It infers that rule from the data by an iterative fitting process using one of two methods. PROC LOGISTIC not only calculates the best fit model, but it also gives confidence limits on all the parameters. Again, there is a huge range of options available to the user in this procedure.

A simple SAS script looks like this:

This script consists of one DATA step and one PROC step. The DATA step begins with a DATA statement which creates a SAS dataset named “distance”. It assigns the value 26.22 to “Miles” and calculates “Kilometers” as 1.61 times Miles. There is then a PROC step beginning with a PROC statement. This invokes PROC PRINT, a procedure which enables various forms of output to be printed. The PROC step ends with the RUN statement, which causes the operations in the PROC to be executed.

In addition to DATA steps and PROC steps, the SAS Language comprises three components called Output Delivery System (“ODS”), SAS Macro Language and SAS SQL steps. ODS is used to output files in a variety of formats. SAS Macro Language is used to program repeated steps. SAS SQL steps are simple combinations of Structured Query Language commands, which are used for manipulating databases.

The manner in which a SAS script is processed is illustrated in the following diagram in which there is a single DATA step and a single PROC step:

The executable forms of DATA steps and PROC steps are passively resident in the system and require a context in which to execute. A program called the SAS runtime environment, a part of Base SAS, executes continually while the SAS System is active on a machine. It is the SAS runtime environment that transfers the flow of execution from itself to a DATA step or a PROC step. When a DATA step is submitted for execution, it is parsed, compiled into SAS's proprietary form of object code and then executed. The PROC steps are not compiled by the SAS System. Instead, the SAS System parses each PROC step to produce statement structures containing the relevant parameters which are then used to run pre-compiled C language programs.

In paragraph 6 of its Particulars of Claim, Amended Particulars of Claim and Re-Amended Particulars of Claim, each of which was verified by a statement of truth, SAS Institute pleaded that:

“SAS Programs [i.e. scripts] are written in a programming language known as the SAS Language.”

This averment is entirely consistent with the way in which SAS Institute describes the SAS Language in both its marketing and technical literature. For example, the SAS website says of Base SAS that it “includes a fourth-generation programming language”. Similarly, the SAS 9.1.3 Language Reference: Concepts (Third Edition, 2005) manual states at pages 4 and 6:

“Overview of Base SAS Software

The core of the SAS System is Base SAS software, which consists of the following:

SAS language a programming language that you use to manage your data.

…

SAS Language Elements

The SAS language consists of statements, expressions, options, formats, and functions similar to those of many other programming languages.”

(I note in passing that, in a more recent edition of this manual, SAS 9.2 Language Reference: Concepts (2009), exhibited by Professor Ivey to his report, the second passage quoted is unaltered, while the first passage has been changed to read:

“Overview of Base SAS Software

The core of the SAS System is Base SAS software, which consists of the following:

DATA step a programming language that you use to manage your data.

…”)

Despite this, SAS Institute adduced evidence from Professor Ivey to the effect that the SAS Language is not a programming language. In paragraphs 110-111 of his report, Professor Ivey expressed the opinion that the SAS DATA steps had many of the characteristics of a conventional programming language, but that the PROCs were “entirely unlike a conventional programming language”.

In paragraphs 126 and 128 of his report Professor Ivey quoted the following definitions from the IEEE Standard Glossary of Software Engineering Terminology, commenting that the first definition was “all-encompassing” and the second “very widely drawn”:

“Programming language. A language used to express computer programs. See also: assembly language; high order language; machine language. Contrast with: query language; specification language.

Computer program. A combination of computer instructions and data definitions that enable computer hardware to perform computational or control functions. See also: software.”

In paragraphs 133-136 of his report, Professor Ivey expressed the opinion that the SAS Language was a domain specific computer language with many of the attributes of a command language, rather than a programming language, in accordance with the following sub-ontology of computer languages proposed in “Towards A General Ontology Of Computer Programs” by Pascal Lando, Anne Lapujade, Gilles Kassel and Frédéric Fürst (ICSOFT 2007: July 22-27, Barcelona, Spain):

In cross-examination, Professor Ivey re-iterated and expanded on what he had said in paragraphs 110-111 of his report. He accepted that SAS DATA steps had most of the characteristics of a programming language and satisfied the requirements for Turing Completeness developed by the British mathematician Alan Turing, but said that the PROC steps did not.

In paragraphs 233-236 of his report Dr Worden expressed the opinion that the sub-ontology relied on by Professor Ivey had no scientific basis and was wrong, and that the SAS Language was a programming language. In a supplementary report he expressed the opinion that Turing Completeness was not a practical criterion for assessing whether something constituted a programming language.

In cross-examination Dr Worden said that DATA steps had most of the characteristics of a programming language. He accepted that this was less true of the PROC step, but said:

“My feeling is that it is all a matter of commands issued to the computer. It is all a matter of commands issued to the computer, whether it is simple commands in the DATA step or complex commands in the PROC step. What you are doing with this language is telling the computer what to do.”

Based on this evidence, counsel for SAS Institute accepted that the DATA steps were a programming language, but submitted that the PROCS were not since the procedures were complex programs. In my view this submission confuses the PROC steps in the SAS Language with the programs in the C language which are invoked by them. The PROC steps are commands issued by the user to the computer expressed in a particular language which is parsed by the SAS System. Accordingly, the SAS Language as a whole, including the PROC steps, satisfies the IEEE definition of a programming language.

In my judgment the SAS Language is a programming language. I am comforted by the fact that this assessment coincides not only with SAS Institute’s own assessment in paragraph 6 of its Particulars of Claim and in its literature, but also with that of the anonymous author of the entry for “SAS language” in Wikipedia (as last modified on 25 April 2010), which states:

“… SAS can be considered a general programming language, though it serves largely as a database programming language and a language with a wide variety of specialized analytic and graphic procedures.”

There is no dispute that each of the SAS Components is an original computer program in which copyright subsists. Nor is there any dispute that the creation of each of the SAS Components involved very substantial skill, judgement and labour on the part of SAS Institute’s employees. Nor is there any dispute that the skill, judgement and labour involved in creating each of the SAS Components included both (i) skill, judgement and labour in determining the requirements for the software and (ii) skill, judgement and labour in designing and writing source code to implement those requirements.

SAS Institute publishes numerous technical manuals for the SAS System which describe the operation and use of the various components. The SAS Manuals are extremely long and detailed. For example, the SAS Manual for SAS/STAT alone runs to some 7,886 pages. The SAS Manuals exhaustively document the functionality of each part of each component, the necessary inputs and, where appropriate, the expected outputs. The SAS Manuals are designed to, and do, give users of the SAS System a great deal of information about the external behaviour of the SAS System.

The SAS Manuals are designed for users who understand statistics. As a result, they frequently employ mathematical equations and formulae to describe the analyses and computations which can be carried out. For the reader seeking more information, they frequently contain references to the statistics literature. Thus the section on PROC UNIVARIATE contains 29 references to the literature.

The SAS Manuals do not contain information about the internal behaviour of the SAS System. Thus they do not give any information about the internal file formats used by the SAS System, the intermediate language used to compile DATA Steps, the SAS object code which the SAS System uses when executing SAS application programs, or the statement structures which result from parsing the PROC steps. Still less do the SAS Manuals contain any details of the source code of the SAS System.

SAS Institute relies on a long list of SAS Manuals which is set out is a Schedule to its Reply. It is not necessary for the purpose of this judgment to reproduce that list. The Manuals to which most reference was made at trial were the SAS 9.2 Language Reference: Dictionary (Second Edition), Base SAS 9.2 Procedures Guide: Statistical Procedures (Second Edition) and SAS/STAT 9.2 User’s Guide (Second Edition), although some others were referred to as well.

SAS Institute adduced evidence from Mrs Council as to the way in which the SAS Manuals were written. She explained that SAS Institute regards the writing of the Manuals as an important task on which significant resources are expended. SAS Institute’s Publications Division currently employs 54 writers and 14 editors. They are technically qualified and have also received substantial training from SAS Institute. There is close liaison between the technical writers who write the Manuals and the developers who develop the relevant components of the SAS System. In some cases, such as SAS/STAT, a lot of the Manual was written by the developers themselves. The relevant parts of each Manual are planned, and to some extent written, while the software is being developed so that the Manuals can be completed as soon as possible after the software.

The SAS Manuals are published in book form and can be purchased from outlets such as Amazon. In addition, online versions are accessible via SAS Institute’s website.

There is no dispute that each of the SAS Manuals is an original literary work in which copyright subsists. Nor is there any dispute that the creation of each of the SAS Manuals involved substantial skill, judgement and labour on the part of SAS Institute’s employees.

WPL was founded by Dr Peter Quarendon, Tom Quarendon, Martin Jupp and two other colleagues in August 2002. Dr Quarendon had worked for IBM for 29 years before leaving in 1995. He had encountered the SAS System during his time at IBM and had written application programs for various data analysis projects using it. After leaving IBM he had acquired a set of manuals for SAS version 6. As I understand it, Dr Quarendon was the only one of the founders with previous experience of the SAS System.

WPL started developing what became WPS soon afterwards, initially in parallel with other work. As Dr Quarendon explained to me, he believed as a result of his experience in the industry, and in particular his time at IBM, that there would be nothing unlawful in creating such software provided that WPL did not copy the source code of the SAS System. In about January 2003 WPL decided to concentrate its efforts on the development of WPS.

To begin with, WPL wrote WPS in the Java language. Subsequently Dr Quarendon and his colleagues decided that this was unsatisfactory, and in 2005 they translated the software into C++.

Over the years WPL has progressively increased the number of functions of the SAS Components that are supported by successive versions of WPS. Even in the current version of WPS, version 2.4, however, there are many functions that are not supported.

There is no dispute that none of the developers of WPS have ever had access to the source code of any of the SAS System. In particular, there is no suggestion that any of them have ever decompiled any SAS object code or even attempted to do so.

WPL has always openly admitted that in writing WPS its programmers studied (i) the SAS Manuals, (ii) the response of the SAS Learning Edition to a large number of SAS programs and (iii) the response of the SAS Full Edition to a large number of SAS programs. In this judgment I am not concerned with the use that WPL has made of the SAS Full Edition, but only the use it has made of the SAS Manuals and the SAS Learning Edition.

SAS Institute contends that WPL used the SAS Manuals, supplemented wherever necessary by observation of the Learning Edition, as the specification of requirements for WPS. WPL does not dispute that its programmers used the SAS Manuals and Learning Edition as sources of requirements information. As Dr Worden pointed out in his report, however, WPL’s programmers also used a variety of other sources. In paragraph 105 of his report (as amended by his supplementary report), he provided a helpful list of the sources used by WPL’s developers:

•

“The SAS User manuals

•

Observation of the behaviour of SAS Learning Edition

•

Customer feedback on the behaviour of the SAS Full Edition

•

Expert SAS users

•

Knowledge of applied statistics

•

Published papers and books on applied statistics

•

Knowledge of production platforms (particularly IBM)

•

Knowledge of products which SAS would interface with (e.g. Oracle)

•

Knowledge of WPL’s own internal requirements (e.g. maintainability, testability, extensibility)

•

Knowledge of the check-in process and bug management systems.”

Dr Quarendon’s estimate, which I accept, was that about 90% of the source code of WPS did not relate to the SAS Manuals.

WPL has also always openly admitted that, with limited exceptions, the response of WPS to SAS scripts and data is intended to be identical to the response of the SAS Components and is in fact identical. The reason why this is so is that that is what customers for WPL insist on. Dr Quarendon explained to me that, even though it was always intended that the response would generally be identical, he had been surprised at the extent to which customers perceived any deviation at all from an identical response as constituting a bug in WPS.

As noted above, WPS does not support all of the functions of the SAS Components. Where a SAS script attempts to invoke a function which WPS does not support, WPS is intended to “fail gracefully” by returning an error message stating that the function is not supported.

As indicated above, WPL produces two forms of documentation for WPS which are the subject of complaint by SAS Institute in these proceedings (there are also some others which are not complained of). The first is the WPS Manual, entitled WPS Reference for Language Elements version 2. Although this is 693 pages long, it is a rather compressed document which describes the functionality of various elements briefly by reference to syntax diagrams. The second is the WPS Guides. These are a set of four quick reference guides. Each guide simply lists elements of the SAS Language and indicates (by means of a Y or a blank in a column headed “Supported”) whether or not it is supported by the current edition of WPS.

The 31 examples of similarities between WPS and the SAS Manuals relied on by SAS Institute, some of which comprise more than one instance of the same kind of thing, were grouped by counsel for SAS Institute into nine categories, although some examples fall into more one category. In this section of the judgment I shall consider each category of similarities in turn. I do not consider that it is necessary to discuss each and every similarity individually, but I have borne in mind all of them, and the evidence concerning them. At this stage I will confine my analysis to what these examples show as a question of fact. I will analyse the legal consequences of the facts found below.

Copying of formulae and descriptions of formulae in the SAS Manuals

This category consists of examples 14-21 inclusive and 30. SAS Institute contends that these examples “demonstrate direct reproduction of the descriptions in the SAS Manuals into the WPS source code”. This is perhaps the most contentious category. It generated a great deal of evidence, but I shall state my conclusions as shortly as possible.

It is sufficient to refer by way of illustration to one of these examples, example 14. This concerns the calculation of percentiles. A percentile is the value or score below which a certain percentage of observations fall e.g. the 80^th percentile is the value below which 80% of observations fall. While this might seem a relatively simple subject, in fact there is no definitive, universally recommended, definition of a percentile. Over the past 40 years many different definitions for percentile computations have been proposed in the statistics literature for various kinds of applications and with different statistical properties. In practice, the numerical differences between calculations using these definitions may be minor (especially with large samples), but for some customers, or where sample sizes are smaller, the use of a particular definition can matter.

Pages 328-330 in Chapter 4 of the Base SAS 9.2 Procedures Guide: Statistical Procedures Guide (Second Edition) describe the calculation of percentiles using PROC UNIVARIATE. At page 328 the Manual states:

Table 4.83 on page 329 sets out five definitions of percentiles which the user may choose to use by means of the PCTLDEF option. This is reproduced highlighted in different colours in exhibit PAI11 as follows:

The routine GetQuantiles in UnivariateSummarizer.cpp in WPS contains the following C++ code, which again is reproduced in exhibit PAI11 highlighted in different colours for ease of comparison with Table 4.83 in the Manual:

Counsel for SAS Institute submitted that this example demonstrated two things. The first was the skill, judgement and labour which SAS Institute put into the development of the SAS System apart from the skill, judgement and labour expended on actually designing and writing the source code.

It appears that four of the percentile definitions in PROC UNIVARIATE came from lectures on non-parametric statistics given by Professor Dana Quade at the University of North Carolina in 1975-1977. Both Professor Harrell and Dr DeLong attended these lectures. Professor Harrell was a SAS user and an active member of the SAS Users Group International, which provided user feedback to SAS Institute on the SAS System. He also wrote two procedures for SAS Institute without compensation while employed by the University, PROC PCTL and PROC DATACHECK, which were precursors of PROC UNIVARIATE. These included percentile calculations using the definitions which Professor Harrell took from the course notes for Professor Quade’s lectures. PROC UNIVARIATE was first introduced in 1979. The development of it was led by Dr DeLong. When originally introduced, PROC UNIVARIATE offered users the option of calculating percentiles using the same four definitions as its precursors.

Since its introduction, PROC UNIVARIATE has been the subject of continuous development. It is now much more complicated and sophisticated than it was in 1979, and SAS Institute has devoted considerable effort to its improvement. Between 1979 and 1985 two changes were made to the percentile definitions available in PROC UNIVARIATE. First, definition 2 was changed by Dr DeLong to make specific provision for the case where g= ½. Secondly, Warren Searle, a colleague of Dr DeLong, suggested adding definition 5 in the light of customer feedback. Definition 5 is now the default option. (Subsequently, a sixth definition involving weighting, which is not shown in Table 4.83, was devised by Dr Yang Yuan.)

In March 1990 Rudolph S. Parrish published a paper (“Comparison of Quantile Estimators in Normal Sampling”, Biometrics, 46, 247-257) in which he compared ten different definitions of quantiles (percentiles), referred to as Y₁ to Y₁₀. Parrish noted at 249 that the first five were implemented in the UNIVARIATE procedure of SAS, referencing the SAS 5 User’s Guide: Basics manual (1985). He also provided the following information at 250:

“Quade (unpublished report, SAS Institute, 1986) provided a summary of earliest references for most of these estimators. Y₁, Y₃, and Y₈ are given in Parzen (1979); Y₄ appears in Read (1972). Hogg and Craig (1978) discuss the case of (n + l)p an integer, wherein Y₄ and Y₅ coincide. Y₇ is treated by Moore (1969) and by Cramer (1946) for np not an integer; Cramer also presents definitions Y₃ and Y₅. Y₃ is given in Ogawa (1962), and Y₅ appears in Greenberg (1978).”

As counsel for SAS Institute pointed out, no reference is given for Y₂, and the other four definitions implemented in PROC UNIVARIATE appear to have originated from disparate sources in the literature.

Parrish’s conclusion at 256 was as follows:

“Many of the nonparametric estimators considered here are frequently encountered in data analysis software, and, in some cases, data analysts are permitted to choose from among several of these. The various estimators can deliver significantly different results, especially when sample sizes are small. The performance characteristics of individual estimators depend on the manner in which the estimator is defined, the sample size, the quantile level being estimated, and the underlying parent distribution. Characteristics such as bias, variance, mean squared error, and relative efficiency facilitate comparisons among estimators. In order to make comparisons that were computationally exact, this investigation assumed a normal parent distribution. Under this condition, the estimators defined by Y₈ and Y₉ are clearly the best of the set. The worst performers are those based on a single order statistic, such as the empirical CDF estimators Y₃, Y₆, and Y₇. Y₈ proved to be the least biased, while Y₉ provided the lowest MSE values overall. Y₉ requires much greater computational effort than does Y₈, and, in practical situations, this may offset its MSE advantage. It will be of interest to consider further the performance characteristics of these estimators for nonnormal parent distributions.”

As counsel for SAS Institute pointed out, this suggests that the best definition to implement in a computer program is Y₈, and yet this is not implemented in PROC UNIVARIATE.

I accept that this evidence shows that the decision as to which percentile definitions to include in PROC UNIVARIATE, and which to leave out, required skill, judgement and labour on the part of SAS Institute’s employees, albeit that most of the definitions appear to have come from sources in the public domain. I also accept that this is representative of the substantial skill, judgement and labour expended by SAS Institute in developing the SAS System apart from the skill, judgement and labour in designing and writing the source code. As in this case, to a large extent the SAS System performs statistical computations in accordance with methods devised by others and published in the statistical literature, but in some case the methods have been devised (or at least modified) by SAS Institute’s employees. As in this case, SAS Institute has to make a choice as which methods or combinations of methods to offer the users of the SAS System.

On the other hand, I do not accept that this evidence shows that any particular skill, judgement or labour was expended by SAS Institute’s employees in devising the precise form of the formulae as they appear in the SAS Manual as distinct from other mathematically-equivalent formulae. Nor do I accept that the evidence demonstrates the expenditure of such skill, judgement or labour in devising the precise form of any of the formulae in the examples in this category. On the contrary, my impression is that the authors of the SAS Manual generally use formulae more or less in the form in which they are to be found in the statistics literature. (I say “more or less” because there is some evidence of small variations which appear to be due to factors such as groups of formulae being presented in a common format.) This is not surprising.

The second thing which counsel for SAS Institute submitted that this example demonstrated was that the formulae contained in the SAS Manual had been “transposed directly into the WPS source code”. Whether this is so depends on what one means by “transposed directly”. Two points are common ground. The first is that, in cases such as this, WPL’s programmers have taken the mathematical formulae set out in the SAS Manuals as concisely, precisely and unambiguously defining the statistical calculations which the SAS System carries out and which WPS must also perform. The second is that WPL’s programmers have chosen to write the corresponding portions of their C++ code in a style which resembles the form of the equations as set out in the SAS Manuals at least to some extent.

It is important to be clear as to the limits of these two points. So far as the first is concerned, it is not possible to take an equation and simply “translate” it into source code. An equation is a static statement of mathematical equivalence, whereas a computer program dynamically calculates values. It follows that an equation is a statement of what is to be calculated, not of how the calculation is to be performed. Furthermore, the primary object of WPL’s programmers was not to write a computer program to implement these mathematical formulae. Their primary object was to write a computer program that would respond to WPL’s customers’ SAS application programs and data by producing the same results, including, if specified by the user by means of a PROC UNIVARIATE step, percentiles calculated from the data in accordance with any of the five definitions set out in Table 4.83.

In carrying out this task, the programmer’s starting point is his knowledge of how to write C++ code, which is a so-called “object-oriented” language, and in particular his knowledge of how to represent the various entities as “objects” in the computer’s memory and how to make those objects interact with each other. Next, the programmer must already have written software to parse scripts containing PROC UNIVARIATE steps, applying the correct defaults for any parameters the user has not specified. The programmer then needs to consider how to provide the functionality of PROC UNIVARIATE. What resources does the programmer draw upon apart from his knowledge of C++ syntax and good programming practices such as how to divide the required functions among files, objects and methods on objects? First, he is guided by the fact he will be writing a PROC object i.e. an object exhibiting known characteristics enabling any PROC to interact with the rest of the code. Secondly, he reads the narrative prose of the SAS Manual which explains the operation of PROC UNIVARIATE. Thirdly, he must understand the statistics involved, which in part guides the decision as to which objects will be created and maintained in memory. For this purpose he may well need to refer to text books and academic papers, including those referenced in the SAS Manuals. Fourthly, he treats the mathematical formulae as specifying the statistical calculations to be performed.

The effort on the part of WPL required to write the necessary code is indicated by the fact that at least 5000 lines of WPS source code are required to support the PROC UNIVARIATE function. As Dr Quarendon (who was the programmer in this case) said, the code extracted by Professor Ivey in example 14 represents the tip of an iceberg.

By the time the programmer comes to percentiles, the data has been acquired from a data source into a data structure in memory which can be iterated through and the programmer has ensured that the data has been sorted. For this purpose Dr Quarendon adapted a sort routine which had already been implemented elsewhere in WPS. Next the programmer must calculate the number of values, which in WPS is held in the variable y0_. None of this code in included in example 14.

The programmer will understand that the calculation of any percentile involves an iteration process over values in the data set. What Dr Quarendon did was to define a method getQuantiles() in the UnivariateSummarizer object which invokes the iterator on the ValuesMap object. It counts the variables into the upper or lower quantile according to the PCTLDEF.

So far as the second point is concerned, the degree of resemblance of form should not be either understated or overstated. In the SAS Manual the algorithm for PCTLDEF is expressed in mathematical formulae using entities n (the number of non-missing values for a variable x), p (t/100 i.e. the t th percentile), j (the integer part of np i.e. n times p), g (the non-integer part), x_j (the jth member of the set of values of x) and x_j+1 (the j+1th member of the set of values of x). In getQuantiles() Dr Quarendon has defined variables np, xj and xjplus1. The definition of np does not multiply n by p, although the variable contains the product of n and p. Nor are xj and xjplus1 references to the elements in the data set themselves.

In each case Dr Quarendon has named the variables after the entities which appear in the SAS Manual. As he pointed out, these variable names are very common ones. Nevertheless, he was frank that this was not the reason for using them here. The reason was to ensure “traceability” of the WPS source code. What this means is that the use of names which resemble the notation used in the formulae in SAS Manual makes it easy for the same programmer when returning to the WPS source code after an interval of time or a different programmer picking up the code for the first time, for example when trying to find and fix a bug or to improve the program, to understand and follow the code, and in particular to see how the code relates to what is described in the SAS Manual. It would have been possible for Dr Quarendon to use different variable names, but the code would be harder to relate to the SAS Manual.

Despite the similarity between the names of the variables in the WPS code and the entities in the SAS Manual, they are used in a somewhat different way. Counsel for SAS Institute suggested that this is simply because it is not possible to use subscripts (or superscripts) in C++, but it goes beyond that. Overall, however, there is some resemblance of form between the formulae in the SAS Manuals and the corresponding portions of the WPS code.

I would add that, although some of the other examples in this category show a similar degree of resemblance of form (e.g. examples 15 and 16), others show a lesser, if any, degree of resemblance (e.g. examples 19, 20 and 30).

The upshot in my judgment is that this category of examples shows two things with regard to what WPL has done. First, it confirms what WPL has always admitted, namely that it has used the SAS Manuals to emulate functionality of the SAS System in WPS. Secondly, it shows that to some extent WPL has reproduced aspects of the SAS Manuals going beyond that which was strictly necessary in order for WPS to emulate the functions of the SAS System.

What it does not show is reproduction of the SAS source code by WPS going beyond the reproduction of its functionality. This is no direct evidence as to how closely or not the SAS source code resembles the SAS Manuals in this respect. My impression from such evidence as there is is that the SAS source code resembles the SAS Manuals less closely than the WPS source code.

PROC UNIVARIATE parser - reproduction of keywords

This category consists of a single example, example 1. Pages 233-240 in chapter 4 of the Base SAS 9.2 Procedures Guide: Statistical Procedures (Second Edition) Manual set out and describe the options that can be used with the PROC UNIVARIATE statement. The keywords for these options range from ALL to WINSOR. Example 2 contains an extract from \procunivariate\ProcUnivariateParser.cpp, which is the WPS parser for PROC UNIVARIATE, which reproduces most of these keywords, albeit not in exactly the same order, in the following manner (taking ALL for instance):

addUnsupportedKeywordOption (“all”);

Counsel for SAS Institute particularly relied on the fact that the WPS parser reproduces the keywords even where, as in this case, the option is not supported by WPS. In my judgment the fact that some of the options are not supported is of no significance. The keywords should still be parsed in order that, if detected by the parser, WPS can take the appropriate action. As stated above, where the options are not supported, WPS is designed to “fail gracefully” by returning an error message.

All this example shows, therefore, is that, as one would expect given WPL’s objective, the WPS parser reproduces the relevant keywords from the relevant SAS Manual and, by that route, from the SAS source code.

Reproduction of names for procedures, functions, etc

This category consists of examples 1-3, 16-17, 24 and 28. SAS Institute’s complaint here is two-fold. The first is that the source code of WPS reproduces SAS keywords such as those discussed above in places other than the parser, and therefore where it is not strictly necessary to do so in order to emulate the functionality of the SAS Components. The second is that the source code of WPS reproduces the names for procedures, functions and so on from the SAS Manuals when it is not strictly necessary to do so in order to emulate the functionality of the SAS Components.

It is sufficient to refer by way of illustration to example 17, which counsel for SAS Institute particularly relied on. Page 225 in chapter 10 of the Base SAS 9.2 Procedures Guide manual sets out in Table 10.1 a key for interpreting the SYSINFO macro return codes from PROC COMPARE. For instance, the first line in the table is as follows:

In my judgment this category of examples again shows reproduction of elements of the SAS Manuals going beyond that which was strictly necessary in order for WPS to emulate the functions of the SAS System. This category of examples also shows reproduction by that route of elements of the SAS source code.

Reproduction of default values

This category consists of examples 2-3 and 26-29. Default values are values which a program adopts if no other values are specified by the user. These examples show that, where it supports a SAS function, WPS has the same defaults.

It is sufficient to refer to the first of these examples, example 2. Page 5479 of the SAS/STAT 9.2 User’s Guide states that, in the MODEL statement of PROC REG, SLENTRY=value and SLE=value specifies the significance level for entry into the model used in the FORWARD and STEPWISE methods and the defaults are 0.50 for FORWARD and 0.15 for STEPWISE. Similarly, it states that SLSTAY=value and SLS=value specifies the significance level for staying in the model for the BACKWARD and STEPWISE methods and the defaults are 0.10 for BACKWARD and 0.15 for STEPWISE. The extract from \proc_reg\RegModelStmt.cpp set out in example 2 shows that WPS has the same defaults of 0.50 and 0.15, 0.10 and 0.15 respectively.

In my judgment counsel for WPL is right to say that none of these examples demonstrate any more than WPL had already admitted, namely that when writing WPS WPL’s developers made it function in the same way as the SAS System. As Dr Quarendon said, all the relevant defaults in WPS have to be the same as in the SAS System in order to ensure that WPS produces the same output in response to the same input as does the SAS System. It is not the case, for example, that the WPS code in example 2 reproduces the wording of the SAS Manual (other than the individual keywords and default values).

Copying of output text

This category consists of example 4. This concerns the “log” file. Like many programs, when the SAS System is run, it produces a log file which produces a record of the activities of the system. The log file contains the script written by the user, and then information provided by the SAS System as to its actions in relation to the script. Some of that information depends upon the nature of the SAS script. Other sections use pre-determined or fixed text.

In this example the SAS Learning Edition log is as follows:

The numbered lines in black or grey are the SAS script which was run to produce the log file. The log file itself contains the unnumbered lines in blue which record the SAS System’s responses to that script. These contain both fixed text (e.g. “The minimum record length was”) and variable data (e.g. “3”). The items of fixed text are contained within the relevant SAS source code.

The WPS log produced by the same SAS script is as follows:

Again, the log file itself contains the unnumbered lines in blue which record WPS’s responses to that script. It can be seen that the fixed content of the responses is very similar, although there are small differences in wording (e.g. “records were read from file” rather than “records were read from the infile” and “A new line was read” rather than “SAS went to a new line”) and presentation (e.g. the format of times and a different shade of blue). Again, the items of fixed text are contained within the WPS source code, although the way in which they are structured in the WPS code is different to that in the SAS code.

It follows that in this instance WPL has copied elements of the SAS source code into the source code of WPS. It is common ground that it did so by examining log files produced during use of the SAS Learning Edition.

In my judgment this is another instance of WPS emulating the functionality of the SAS System in the sense of producing a very similar output to the same input. It may be said to be strictly unnecessary to the actual data processing functions for WPS to do this, but Tom Quarendon explained that WPL did this because users were used to scanning SAS log files for information and therefore expected the information to be presented in a certain way. This is very similar to the reason why easyJet wanted BulletProof to copy the VT100 screens and reports in the Navitaire case.

Comments in the WPS source code

This category consists of examples 5-13 and 31. SAS Institute contends that these examples “demonstrate the slavish copying of the SAS System by WPS” in that they show that information found in the SAS Manuals or by running the SAS System “has been reproduced in the WPS code, even where the reason was not understand, or where WPL believed it could achieve a better solution”.

It is worth noting that Professor Ivey explained that he and his assistants had undertaken a complete search of the WPS code for comments of this kind. Accordingly, it is probable that these 10 examples are not merely representative, but complete.

It is sufficient to quote the first of the two instances in example 7, which was highlighted by counsel for SAS Institute in SAS Institute’s skeleton argument:

In my judgment counsel for WPL is right to say that, in general, none of these examples demonstrate any more than WPL had already admitted, namely that when writing WPS WPL’s developers made it function in the same way as the SAS System even if it either appeared to do something other than what they regarded as correct or deviated from what was described in the SAS Manuals. I say “in general” because SAS Institute advance an additional argument specifically in relation to examples 22 and 31 which I must deal with separately.

Use of SAS Manuals as the source of optimising information for WPS

This category consists of example 22 and 31. SAS Institute contends that these examples demonstrate not merely slavish copying of the SAS System by WPS, but more specifically the copying of features which optimise the performance (i.e. execution speed) of the software as distinct from its functionality.

It is sufficient to refer to example 31. Chapter 51 of the SAS/STAT 9.2 User’s Guide (Second Edition) discusses PROC LOGISTIC. Page 3328 explains that (I have added some of the surrounding text to that contained in PAI11 to provide the context):

“The LOGISTIC procedure uses a simple empirical approach to recognize the data configurations that lead to infinite parameter estimates. The basis of this approach is that any convergence method of maximising the log likelihood must yield a solution giving complete separation, if such a solution exists. In maximizing the log likelihood, there is no checking for complete or quasi-complete separation if convergence is attained in eight or fewer iterations. Subsequent to the eighth iteration, the probability of the observed response is computed for each observation. If the predicted response equals the observed response for every observation, there is a complete separation of data points and the iteration process is stopped.”

The comment in the WPS source code reads as follows

There is a considerable amount of evidence about this example, and the related example 22, but I shall state my conclusions as shortly as possible. Professor Ivey’s analysis of both these examples was marred by his lack of understanding of the underlying statistics, whereas Dr Worden understood the statistics rather better and therefore was able to present a clear and convincing analysis. To summarise briefly, PROC LOGISTIC does an iterative search to try and maximise the likelihood that the proposed rule fits the data. When it can improve the likelihood no more, the search is said to have converged. One problem which can arise is that there is separation in the data, which means that more than one rule fits the data equally well. In cases where there is separation, there is no convergence and so the search algorithm will continue its iteration for as long as it is allowed to continue. The SAS System normally applies a check for separation after eight iterations and stops if it finds separation (or quasi-separation), although checking for separation can be turned off by the parameter NOCHECK. It does not check for separation in the first eight iterations.

Counsel for SAS Institute put it to Dr Worden, on the strength of the description of the NOCHECK parameter in the Manual which is quoted in example 22, that the reason why the SAS System does not check for eight iterations was to reduce execution time. Dr Worden agreed that that was what the Manual appeared to indicate, and so I will assume that it is correct although I cannot find any direct evidence which establishes it.

It is common ground that, as the comment shows, WPS adopts the same approach of not checking for eight iterations even though WPL’s programmer did not understand the rationale for this. In my judgment it is clear from the evidence that the reason for this is to ensure that WPS gives the same result as the SAS System. As Dr Worden explained, where there is separation, the result after eight iterations will be quite different to the result after, say, five iterations. So making the first check for separation after five iterations could yield a different result to making the first check after eight iterations.

Counsel for SAS Institute submitted that WPL had used the information in the SAS Manual to optimise the performance of WPS. I do not accept this. What these examples show, once again, is that WPL has copied the functionality of the SAS System. In this case, that would appear to have the incidental result that WPS gets the same performance benefit as the SAS System; but that is not why it was done.

Use of SAS data file formats

This category consists of example 23. This concerns a SAS data file format called SAS7BDAT. I have already commented on Professor Ivey’s evidence in relation to the data file formats in paragraph 23 above. The example consists of “the SAS source code used to generate files in the SAS7BDAT formats” on the one hand and 15 identified routines in the WPS source code and their functions on the other hand. As presented in exhibit PAI11, example 23 appears to be an allegation of (presumably indirect) reproduction of unidentified SAS source code. That is also the flavour of Professor Ivey’s evidence in the body of his report. As stated above, however, not merely is the SAS source code unidentified, but also Professor Ivey had not even inspected the relevant source code. There is no evidence that the SAS source code sets out the SAS7BDAT format, as opposed to reading and writing files in that format.

Counsel for SAS Institute submitted that this example demonstrated that WPL had “copied elements of the SAS source code into the source code of WPS”. In my judgment it shows nothing of the kind. All it shows is that, by examining the SAS System in operation, WPL worked out enough of the format of SAS7BDAT data files to be able to write source code which reads and writes data files in that format. This is precisely what the extract from WPL’s internal wiki exhibited by Professor Ivey as exhibit PAI9 to his report indicated that it did, and precisely what Mr Clemow said in his witness statement that he did. The cross-examination of Mr Clemow took matters no further.

Use of SAS test scripts from the SAS Manuals

This category consists of examples 24 and 25. Both concern the use made by WPL of SAS test scripts from the SAS Manuals.

It is sufficient to refer to example 24. Page 514 of the SAS 9.2 Language Reference: Dictionary (Second Edition) gives the following example of the use of the CALL SORTN routine:

It is convenient to mention at this point that WPL admit to running SAS test scripts of this kind first through the SAS Learning Edition to observe the output it produced and then through WPS to check that WPS produced the same output (or failed gracefully).

WPL’s source code contains the following routine in call-routines/CallSortnTest.cpp:

Tom Quarendon explained that this routine, and others like it, are used by WPL for nightly in-house regression testing of WPS, and do not form part of the executable WPS code supplied to customers. The purpose of the regression tests is to ensure that the output produced by the WPS software remains consistent.

The 51 examples of similarities between the respective manuals are contained in trial bundle F prepared by SAS Institute’s solicitors. This sets out on the left hand and right hand sides of each page the corresponding parts of the manuals with identifying letters added in red for ease of reference and yellow highlighting added to identify the similarities relied on. This is a very useful forensic tool. It should nevertheless be borne in mind that, as counsel for WPL pointed out, the true comparison is between the actual manuals. In certain, mostly minor, respects the comparisons in bundle F make the respective manual entries seem more similar than they really are.

Not all of the examples were examined at trial, although quite a number were. Again, I do not consider that it is necessary to discuss all of the examples, but I have borne in mind all of them and all the evidence concerning them. It is sufficient by way of illustration to refer to the first example, which concerns the CALL SCAN routine. I reproduce the relevant pages from bundle F in the Appendix to this judgment. I take this example not merely because it is first, but also because it is neither SAS Institute’s best one nor its worst. It is therefore reasonably representative.

There is no dispute that the text of the WPS Manual is very similar to the text of the corresponding parts of the SAS Manuals. The disputes are as to the reasons for the similarities and whether those matter.

SAS Institute contends that the reason for the similarities is that WPL copied the SAS Manuals, and that it is immaterial whether WPL did this directly from the SAS Manuals or indirectly via the WPS source code. Furthermore, SAS Institute contends that in at least some cases it is more likely that the WPL programmer copied directly from the SAS Manuals than indirectly via the SAS source code. In support of this contention SAS Institute points to the general similarities in language between the corresponding passages in the respective manuals. SAS Institute also points to apparent mistakes in the WPS Manual which it suggests are indicia of copying. For instance, in the case of the CALL SCAN example, in section E on page 4 the SAS Manual reads “o or O processes the charlist and modifier arguments only once” while the WPS Manual reads “o or O processes the and only once”. Similarly, in section D on page 5 the SAS Manual reads “t or T trims trailing blanks from the string and charlist arguments” while the WPS Manual reads “t or T trims trailing blanks from the and arguments”.

WPL’s evidence is that it was its policy that the WPS Manual should not be copied directly from the SAS Manuals because it was conscious of the danger of infringing the copyright in the SAS Manuals, and that the WPL programmers did not do so. WPL say that the reasons for the similarities are that: (i) the SAS Manuals describe in detail the functionality, and in particular the syntax and semantics, of the relevant elements of the SAS System; (ii) WPL used the SAS Manuals to reproduce the functionality, and in particular the syntax and semantics, of those elements of the SAS System in the WPS source code; (iii) the WPS Manual was written from the WPS source code; and accordingly (iv) the similarities are attributable to the fact both the SAS Manuals and the WPS Manual are describing the same functionality, and in particular the same syntax and semantics. WPL accepts this process amounts to it having indirectly copied the SAS Manuals, but contends that the distinction is an important one.

Once again, there is a lot of evidence on this subject, but I shall state my conclusions as briefly as possible.

In general, each WPL programmer was responsible for writing the parts of the WPS Manual which document the code written by him. Although there were slight differences between the working methods of the different programmers, in general the way in which they proceeded was to read the relevant part of the relevant SAS Manual, then write the WPS code and then produce the relevant part of the WPS Manual. By way of exception to this general rule, parts of the WPS Manual were written by Mr Brent and Mr Jupp although they had not written the underlying code. They used the SAS Manuals to understand the operation of the function in question, and in particular to ascertain what options and sub-options were listed. They then ascertained either directly from the programmer or by running WPS which options and sub-options WPS supported, and then wrote the relevant section of the WPS Manual.

It is important to appreciate that WPL partly automated the production of the WPS Manual. The first version of the WPS Manual was partly generated from the information in the Java classes that drove the parser in the first, Java version of WPS. In October 2004 Tom Quarendon wrote a tool to interrogate the Java classes to extract the language elements supported, that is which PROCs, formats, informats, functions, global statements and DATA step statements were supported. For the PROCs it also extracted what options were valid for the PROC and the types of those options, and what statements were valid within the PROC. For the functions it extracted the number and type of the arguments required for the function, and whether those arguments were optional. For the formats and informats it extracted the range of valid widths for the format or informat. It then produced skeleton documentation in an XML format which included a syntax diagram and left blanks for the descriptions of the elements which the programmers filled in manually.

Subsequently WPL produced a set of XML tags which were used to produce XML documents which were then processed into formatted HTML documents. The XML scripts were structured so that each language element was contained in its own file. Where possible, a file would be based on a pre-existing one, but otherwise a new one would be created. In the XML scripts a <name> tag contains the name of the language element (e.g. CALL SCAN), an <overview> tag contains a short description of the basic function of the language element (e.g. “The CALL SCAN returns the position and length of a specific word from its first argument”), a <syntax> tag defines the syntax, a <syntax-description> tag links elements in the syntax with associated textual descriptions and so on. When adding the names of the arguments to the <syntax> tags, the argument names are given IDs (e.g. “argument1”, “argument2” etc). When the programmer comes to write the descriptions, he can write tags which refer to the IDs at the relevant places and those tags should be automatically replaced by the name of the arguments.

The significance of this is illustrated by the CALL SCAN example. Tom Quarendon explained that the missing words in the description of CALL SCAN in the WPL Manual referred to in paragraph 138 above were due to an error he made in writing the XML script for CALL SCAN. As can be seen from the script which he exhibited, in the <syntax> tag he defined “argument1” as “source” (the expression used in the SAS Manual is “string”), “argument5” as “delimiters” (the expression used in the SAS Manual is “charlist”) and “argument6” as “modifiers” (the expression used in the SAS Manual is “modifier”). When it came to <syntax description> tags, in the case of “o or O”, he should have written >processes the <grammar-ref refid=”argument5”/> and <grammar-ref refid=”argument6”/> only once…. In fact he wrote >processes the <grammar-ref name=”argument5”/> and <grammar-ref name=”argument6”/> only once… This mistake led to the tags being resolved to an empty string. He made a similar error in the case of “t or T”.

Although this explanation was not challenged at a technical level, counsel for SAS Institute nevertheless put it to Mr Quarendon in cross-examination and submitted in closing submissions that the mistake showed that Mr Quarendon had been copying the SAS Manual. In my judgment, the reverse is true. What the mistake shows is that, at least to this extent, he was not copying the SAS Manual.

The fact that the argument that the missing words are an indication of copying turns out to be a bad point is by no means the end of the matter, however. SAS Institute is still left with the multiple similarities in language between the SAS MANUAL and the WPS Manual, for example the phraseology “trims trailing blanks from the … arguments” which is used to describe “o or O” in CALL SCAN in both cases. (As indicated above, in some of the other examples there are even more similarities in language.) What is the explanation for this?

Tom Quarendon’s evidence was that, as described above, his general approach was to read the SAS Manual entry, then write the WPS code and then write the XML script for the WPS Manual entry from the WPS code. He accepted that he referred to the SAS Manual when writing the WPS code. The time gap between him studying the SAS Manual and writing the WPS Manual entry varied, but sometimes it could be on the same day. Furthermore, if there was a long gap he would sometimes need to refresh his memory as to what some of the options did. If he knew that he had looked at the SAS Manual recently, he would sometimes go back to it after writing his WPS Manual entry to check that the wording was different and, if he thought it too similar, change it. While he did not specifically recall writing CALL SCAN, he thought it likely that he followed his usual approach in this case.

I am prepared to accept that WPL’s manual writers did not copy directly from the SAS Manuals in the sense of having one of the SAS Manuals open in front of them when writing the WPS Manual and intentionally either transcribing or paraphrasing the wording. I also accept that a considerable degree of similarity in both content and language between the SAS Manual entries and the WPS Manual entries is to be expected given that they are describing identical functionality. Nevertheless, in my judgment the degree of resemblance in the language goes beyond that which is attributable to describing identical functionality. Moreover, it is plain that the reason for this is that WPL’s writers had the language of the SAS Manuals firmly in their minds when writing the WPS Manual entries. Although they tried to avoid describing the functions in the same language, they did not succeed in avoiding the use of very similar language.

In this section of the judgment I shall set out the relevant legislation and some of the existing case law relevant to the issues in this case.

Berne Convention. The International Convention for the Protection of Literary and Artistic works signed at Berne on 9 September 1886 (Paris Act of 1971 as amended in 1979) (“the Berne Convention”), to which all Member States of the European Union are parties, includes the following provisions:

“Article 2

(1)

The expression ‘literary and artistic works’ shall include every production in the literary, scientific and artistic domain, whatever may be the mode or form of its expression, such as books, pamphlets and other writings; lectures, addresses, sermons and other works of the same nature; dramatic or dramatico-musical works; choreographic works and entertainments in dumb show; musical compositions with or without words; cinematographic works to which are assimilated works expressed by a process analogous to cinematography; works of drawing, painting, architecture, sculpture, engraving and lithography; photographic works to which are assimilated works expressed by a process analogous to photography; works of applied art; illustrations, maps, plans, sketches and three-dimensional works relative to geography, topography, architecture or science.

…

(5)

Collections of literary or artistic works such as enyclopaedias and anthologies which, by reason of the selection and arrangement of their contents, constitute intellectual creations, shall be protected as such, without prejudice to the copyright in each of the works forming part of such collections.

(6)

The works mentioned in this Article shall enjoy protection in all countries of the Union. This protection shall operate for the benefit of the author and his successors in title.

…

Article 9

(1)

Authors of literary and artistic works protected by this Convention shall have the exclusive right of authorising the reproduction of these works, in any manner or form.

(2)

It shall be a matter for legislation in the countries of the Union to permit the reproduction of such works in certain special cases, provided that such reproduction does not conflict with a normal exploitation of the work and does not unreasonably prejudice the legitimate interests of the author.

…

Article 20

The Governments of the countries of the Union reserve the right to enter into special agreements among themselves, in so far as such agreements grant to authors more extensive rights than those granted by the Convention, or contain other provisions not contrary to this Convention. The provisions of existing agreements which satisfy these conditions shall remain applicable.”

TRIPS. The Agreement on Trade-related Aspects of Intellectual Property Rights (commonly known as “TRIPS”) which forms Annex 1C to the Agreement establishing the World Trade Organisation signed in Morocco on 15 April 1994, to whom the European Union and all its Member States are parties, includes the following provisions:

“Article 9

Relation to the Berne Convention

1.

Members shall comply with Articles 1 through 21 of the Berne Convention (1971) and the Appendix thereto. However, Members shall not have rights or obligations under this Agreement in respect of the rights conferred under Article 6bis of that Convention or of the rights derived therefrom.

2.

Copyright protection shall extend to expressions and not to ideas, procedures, methods of operation or mathematical concepts as such.

Article 10

Computer Programs and Compilations of Data

1.

Computer programs, whether in source or object code, shall be protected as literary works under the Berne Convention (1971).

2.

Compilations of data or other material, whether in machine readable or other form, which by reason of the selection or arrangement of their contents constitute intellectual creations, shall be protected as such. Such protection, which shall not extend to the data or material itself, shall be without prejudice to any copyright subsisting in the data or material itself.

…

Article 13

Limitation and Exceptions

Members shall confine limitations or exceptions to exclusive rights to certain special cases which do not conflict with a normal exploitation of the work and do not unreasonably prejudice the legitimate interests of the author.”

WIPO Copyright Treaty. The World Intellectual Property Organisation Copyright Treaty agreed in Geneva on 20 December 1996 (“the WIPO Copyright Treaty”), to which the European Union and all its Member States are parties, includes the following provisions:

“Article 1

Relation to the Berne Convention

At the Diplomatic Conference at which the WIPO Copyright Treaty was adopted, a number of statements concerning its interpretation were agreed. These include the following statements:

“Concerning Article 4

The scope of protection for computer programs under Article 4 of this Treaty, read with Article 2, is consistent with Article 2 of the Berne Convention and on a par with the relevant provisions of the TRIPS Agreement.

Concerning Article 5

The scope of protection for compilations of data (databases) under Article 5 of this Treaty, read with Article 2, is consistent with Article 2 of the Berne Convention and on a par with the relevant provisions of the TRIPS Agreement.”

Although the WIPO Copyright Treaty was approved by the Council of the European Community on behalf of the European Community by Council Decision 2000/278/EC of 16 March 2000, the European Union and many of its Member States, including the United Kingdom, did not ratify the Treaty until 14 December 2009, although some Member States ratified it at an earlier date. The Treaty entered into force with respect to the European Union and the United Kingdom on 14 March 2010.

Software Directive. Council Directive 91/250/EEC of 14 May 1991 on the legal protection of computer programs (“the Software Directive”) was recently replaced by a codified (i.e. consolidated) version, European Parliament and Council Directive 2009/24/EC of 23 April 2009 in which both the sequence of the recitals and the numbering of some of the Articles has been changed. Since the original version was in force at the time of most of the alleged infringing activity by WPL and since the case law discussed below refers to the original version, however, it is convenient to continue to refer to the original version. There is no relevant difference of substance between the two versions.

The Software Directive includes the following recitals (numbering added for ease of reference):

“[1] Whereas computer programs are at present not clearly protected in all Member States by existing legislation and such protection, where it exists, has different attributes;

[2] The development of computer programs requires the investment of considerable human, technical and financial resources while computer programs can be copied at a fraction of the costs needed to develop them independently;

…

[4] Whereas certain differences in the legal protection of computer programs offered by the laws of the Member States have direct and negative effects on the functioning of the common market as regards computer programs and such differences could well become greater as Member States introduce new legislation on this subject;

[5] Whereas existing differences having such effects need to be removed and new ones prevented from arising, while differences not adversely affecting the functioning of the common market to a substantial degree need not be removed or prevented from arising;

[6] Whereas the Community's legal framework on the protection of computer programs can accordingly in the first instance be limited to establishing that Member States should accord protection to computer programs under copyright law as literary works and, further, to establishing who and what should be protected, the exclusive rights on which protected persons should be able to rely in order to authorize or prohibit certain acts and for how long the protection should apply;

[7] Whereas, for the purpose of this Directive, the term 'computer program' shall include programs in any form, including those which are incorporated into hardware; whereas this term also includes preparatory design work leading to the development of a computer program provided that the nature of the preparatory work is such that a computer program can result from it at a later stage;

…

[10] Whereas the function of a computer program is to communicate and work together with other components of a computer system and with users and, for this purpose, a logical and, where appropriate, physical interconnection and interaction is required to permit all elements of software and hardware to work with other software and hardware and with users in all the ways in which they are intended to function;

[11] Whereas the parts of the program which provide for such interconnection and interaction between elements of software and hardware are generally known as ‘interfaces’;

[12] Whereas this functional interconnection and interaction is generally known as ‘interoperability’; whereas such interoperability can be defined as the ability to exchange information and mutually to use the information which has been exchanged;

[13] Whereas, for the avoidance of doubt, it has to be made clear that only the expression of a computer program is protected and that ideas and principles which underlie any element of a program, including those which underlie its interfaces, are not protected by copyright under this Directive;

[14] Whereas, in accordance with this principle of copyright, to the extent that logic, algorithms and programming languages comprise ideas and principles, those ideas and principles are not protected under this Directive;

[15] Whereas, in accordance with the legislation and jurisprudence of the Member States and the international copyright conventions, the expression of those ideas and principles is to be protected by copyright;

…

[19] Whereas a person having a right to use a computer program should not be prevented from performing acts necessary to observe, study or test the functioning of the program, provided that these acts do not infringe the copyright in the program;

[20] Whereas the unauthorized reproduction, translation, adaptation or transformation of the form of the code in which a copy of a computer program has been made available constitutes an infringement of the exclusive rights of the author;

[21] Whereas, nevertheless, circumstances may exist when such a reproduction of the code and translation of its form within the meaning of Article 4 (a) and (b) are indispensable to obtain the necessary information to achieve the interoperability of an independently created program with other programs;

[22] Whereas it has therefore to be considered that in these limited circumstances only, performance of the acts of reproduction and translation by or on behalf of a person having a right to use a copy of the program is legitimate and compatible with fair practice and must therefore be deemed not to require the authorization of the rightholder;

[23] Whereas an objective of this exception is to make it possible to connect all components of a computer system, including those of different manufacturers, so that they can work together;

[24] Whereas such an exception to the author's exclusive rights may not be used in a way which prejudices the legitimate interests of the rightholder or which conflicts with a normal exploitation of the program;

…”

The Software Directive includes the following operative provisions:

“Article 1

Object of protection

1.

In accordance with the provisions of this Directive, Member States shall protect computer programs, by copyright, as literary works within the meaning of the Berne Convention for the Protection of Literary and Artistic Works. For the purposes of this Directive, the term 'computer programs' shall include their preparatory design material.

2.

Protection in accordance with this Directive shall apply to the expression in any form of a computer program. Ideas and principles which underlie any element of a computer program, including those which underlie its interfaces, are not protected by copyright under this Directive.

3.

A computer program shall be protected if it is original in the sense that it is the author's own intellectual creation. No other criteria shall be applied to determine its eligibility for protection.

…

Article 4

Restricted Acts

Subject to the provisions of Articles 5 and 6, the exclusive rights of the rightholder within the meaning of Article 2, shall include the right to do or to authorize:

(a)

the permanent or temporary reproduction of a computer program by any means and in any form, in part or in whole. Insofar as loading, displaying, running, transmission or storage of the computer program necessitate such reproduction, such acts shall be subject to authorization by the rightholder;

(b)

the translation, adaptation, arrangement and any other alteration of a computer program and the reproduction of the results thereof, without prejudice to the rights of the person who alters the program;

…

Article 5

Exceptions to the restricted acts

1.

In the absence of specific contractual provisions, the acts referred to in Article 4 (a) and (b) shall not require authorization by the rightholder where they are necessary for the use of the computer program by the lawful acquirer in accordance with its intended purpose, including for error correction.

2.

The making of a back-up copy by a person having a right to use the computer program may not be prevented by contract insofar as it is necessary for that use.

3.

The person having a right to use a copy of a computer program shall be entitled, without the authorization of the rightholder, to observe, study or test the functioning of the program in order to determine the ideas and principles which underlie any element of the program if he does so while performing any of the acts of loading, displaying, running, transmitting or storing the program which he is entitled to do.

Article 6

Decompilation

1.

The authorization of the rightholder shall not be required where reproduction of the code and translation of its form within the meaning of Article 4 (a) and (b) are indispensable to obtain the information necessary to achieve the interoperability of an independently created computer program with other programs, provided that the following conditions are met:

(a)

these acts are performed by the licensee or by another person having a right to use a copy of a program, or on their behalf by a person authorized to do so;

(b)

the information necessary to achieve interoperability has not previously been readily available to the persons referred to in subparagraph (a); and

(c)

these acts are confined to the parts of the original program which are necessary to achieve interoperability.

2.

The provisions of paragraph 1 shall not permit the information obtained through its application:

(a)

to be used for goals other than to achieve the interoperability of the independently created computer program;

(b)

to be given to others, except when necessary for the interoperability of the independently created computer program; or

(c)

to be used for the development, production or marketing of a computer program substantially similar in its expression, or for any other act which infringes copyright.

3.

In accordance with the provisions of the Berne Convention for the protection of Literary and Artistic Works, the provisions of this Article may not be interpreted in such a way as to allow its application to be used in a manner which unreasonably prejudices the right holder's legitimate interests or conflicts with a normal exploitation of the computer program.

…

Article 9

Continued application of other legal provisions

1.

The provisions of this Directive shall be without prejudice to any other legal provisions such as those concerning patent rights, trademarks, unfair competition, trade secrets, protection of semiconductor products or the law of contract. Any contractual provisions contrary to Article 6 or to the exceptions provided for in Article 5(2) and (3) shall be null and void.

…”

Information Society Directive. European Parliament and Council Directive 2001/29/EC of 22 May 2001 on the harmonisation of certain aspects of copyright and related rights in the information society (“the Information Society Directive”) includes the following recitals:

“(15)

The Diplomatic Conference held under the auspices of the World Intellectual Property Organisation (WIPO) in December 1996 led to the adoption of two new Treaties, the ‘WIPO Copyright Treaty’ and the ‘WIPO Performances and Phonograms Treaty’, dealing respectively with the protection of authors and the protection of performers and phonogram producers. Those Treaties update the international protection for copyright and related rights significantly, not least with regard to the so-called ‘digital agenda’, and improve the means to fight piracy world-wide. The Community and a majority of Member States have already signed the Treaties and the process of making arrangements for the ratification of the Treaties by the Community and the Member States is under way. This Directive also serves to implement a number of the new international obligations.

…

(20)

This Directive is based on principles and rules already laid down in the Directives currently in force in this area, in particular Directives 91/250/EEC, 92/100/EEC, 93/83/EEC, 93/98/EEC and 96/9/EC, and it develops those principles and rules and places them in the context of the information society. The provisions of this Directive should be without prejudice to the provisions of those Directives, unless otherwise provided in this Directive.

…

(44)

When applying the exceptions and limitations provided for in this Directive, they should be exercised in accordance with international obligations. Such exceptions and limitations may not be applied in a way which prejudices the legitimate interests of the rightholder or which conflicts with the normal exploitation of his work or other subject-matter. The provision of such exceptions or limitations by Member States should, in particular, duly reflect the increased economic impact that such exceptions or limitations may have in the context of the new electronic environment. Therefore, the scope of certain exceptions or limitations may have to be even more limited when it comes to certain new uses of copyright works and other subject-matter.

…”

The Information Society Directive includes the following operative provisions:

“Article 1

Scope

…

2.

Except in the cases referred to in Article 11, this Directive shall leave intact and shall in no way affect existing Community provisions relating to:

(a)

the legal protection of computer programs;

…

Article 2

Reproduction right

Member States shall provide for the exclusive right to authorise or prohibit direct or indirect, temporary or permanent reproduction by any means and in any form, in whole or in part:

(a)

for authors, of their works;

…

Article 5

Exceptions and limitations

...

2.

Member States may provide for exceptions or limitations to the rights provided for in Articles 2 and 3 in the following cases:

…

(d)

quotations for purposes such as criticism or review, provided that they relate to a work or other subject-matter which has already been lawfully made available to the public, that, unless this turns out to be impossible, the source, including the author's name, is indicated, and that their use is in accordance with fair practice, and to the extent required by the specific purpose;

…”

The European Communities Act 1972 (as amended) includes the following provisions:

“Short title and interpretation

1.

…

(3)

If Her Majesty by Order in Council declares that a treaty specified in the Order is to be regarded as one of the EU Treaties as herein defined, the Order shall be conclusive that it is to be so regarded…

General implementation of Treaties

2.(1) All such rights, powers, liabilities, obligations and restrictions from time to time created or arising by or under the Treaties, and all such remedies and procedures from time to time provided for by or under the Treaties, as in accordance with the Treaties are without further enactment to be given legal effect or used in the United Kingdom shall be recognised and available in law, and be enforced, allowed and followed accordingly; and the expression ‘enforceable EU right’ and similar expressions shall be read as referring to one to which this subsection applies.

(2)

Subject to Schedule 2 to this Act, at any time after its passing Her Majesty may by Order in Council, and any designated Minister or department may by order, rules, regulations or scheme, make provision-

(a)

for the purpose of implementing any EU obligation of the United Kingdom, or enabling any such obligation to be implemented, or of enabling any rights enjoyed or to be enjoyed by the United Kingdom under or by virtue of the Treaties to be exercised; or

(b)

for the purpose of dealing with matters arising out of or related to any such obligation or rights or the coming into force, or the operation from time to time, of subsection (1) above…

(4)

The provision that may be made under subsection (2) above includes, subject to Schedule 2 to this Act, any such provision (of any such extent) as might be made by Act of Parliament, any enactment passed or to be passed, other than one contained in this Part of this Act, shall be construed and have effect subject to the foregoing provisions of this section…”

The Copyright, Designs and Patents Act 1988 was amended pursuant to section 2(2) of the 1972 Act by the Copyright (Computer Programs) Regulations 1992, SI 1992/3233, to implement the Software Directive. The 1988 Act was amended pursuant to section 2(2) of the 1972 Act by the Copyright and Related Rights Regulations 2003, SI 2003/2498, to implement the Information Society Directive. As amended by the 1992 Regulations and the 2003 Regulations, and other legislation, the 1988 Act includes the following provisions:

“Copyright and copyright works

1.(1) Copyright is a property right which subsists in accordance with this Part in the following descriptions of work-

(a)

original literary, dramatic, musical or artistic works,

….

Literary, dramatic and musical works

3.(1) In this Part:

‘literary work’ means any work, other than a dramatic or musical work, which is written, spoken or sung, and accordingly includes:

(a)

a table or compilation other than a database,

(b)

a computer program,

(c)

preparatory design material for a computer program, and

(d)

a database.

…

The acts restricted by copyright in a work

16.(1) The owner of the copyright in a work has, in accordance with the following provisions of this Chapter, the exclusive right to do the following acts in the United Kingdom –

(a)

to copy the work (see section 17);

…

(3)

References in this Part to the doing of an act restricted by the copyright in a work are to the doing of it –

(a)

in relation to the work as a whole or any substantial part of it

(b)

either directly or indirectly;

and it is immaterial whether any intervening acts themselves infringe copyright.

(4)

This Chapter has effect subject to-

(a)

the provisions of Chapter III (acts permitted in relation to copyright works)

…

Infringement of copyright by copying

17.(1) The copying of the work is an act restricted by the copyright in every description of copyright work; and references in this Part to copying and copies shall be construed as follows.

(2)

Copying in relation to a literary, dramatic, musical or artistic work means reproducing the work in any material form. This includes storing the work in any medium by electronic means.

…

Criticism, review and news reporting

30.(1) Fair dealing with a work for the purpose of criticism or review, of that or another work or of a performance of a work, does not infringe any copyright in the work provided that it is accompanied by a sufficient acknowledgement and provided that the work has been made available to the public

…

Observing, studying and testing of computer programs

50BA.(1) It is not an infringement of copyright for a lawful user of a copy of a computer program to observe, study or test the functioning of the program in order to determine the ideas and principles which underlie any element of the program if he does so while performing any of the acts of loading, displaying, running, transmitting or storing the program which he is entitled to do.

(2)

Where an act is permitted under this section, it is irrelevant whether or not there exists any term or condition in an agreement which purports to prohibit or restrict the act (such terms being, by virtue of section 296A, void)

…

Minor definitions

178.

In this Part-

…

‘sufficient acknowledgement’ means an acknowledgement identifying the work in question by its title or other description, and identifying the author unless –

(a)

in the case of a published work, it is published anonymously;

…

Avoidance of certain terms

296A.(1) Where a person has the use of a computer program under an agreement, any term or condition in the agreement shall be void in so far as it purports to prohibit or restrict –

...

(c)

the observing, studying or testing of the functioning of the program in accordance with section 50BA.”

The WTO Agreement, which includes TRIPS, was declared to be an EU Treaty pursuant to section 1(3) of the 1972 Act by the European Communities (Definition of Treaties) (The Agreement Establishing the World Trade Organisation) Order 1995, SI 1995/265. The WIPO Copyright Treaty was declared to be an EU Treaty by the European Communities (Definition of Treaties) (WIPO Copyright Treaty and WIPO Performances and Phonograms Treaty) Order 2005, SI 2005/3431. It does not appear, however, that the 1988 Act has been amended in any respect relevant to the present case as a consequence of either of these Orders.

It is well established that domestic legislation, and in particular legislation specifically enacted or amended to implement a European directive, must be construed so far as is possible in conformity with, and to achieve the result intended by, the directive: Case C-106/89 Marleasing SA v La Comercial Internacional de Alimentación SA [1990] ECR I-4135 at [8]; Litster v Forth Dry Dock and Engineering Co Ltd [1990] 1 AC 546, HL at 558C-H (Lord Templeman) and 576E-577D (Lord Oliver of Aylmerton); Cases C-397/01 to C-403/01 Pfeiffer v Deutsches Rotes Kreuz, Kreisverband Waldshut eV [2004] ECR I-8835 at [113]-[117]; and R (IDT Card Services Ireland Ltd) v Commissioners for Her Majesty’s Revenue and Customs [2006] EWCA Civ 29, [2006] STC 1252 at [73]-[92] (Arden LJ). This is a strong duty of interpretation. Indeed, at the end of the passage in her judgment in IDT that I have referred to, Arden LJ suggests that there is little difference between the Court’s duty under the Marleasing principle and its duty under section 3 of the Human Rights Act 1998 to interpret domestic legislation as far as possible in a manner which is compatible with Convention rights under the European Convention on Human Rights.

The ordinary rule of statutory construction is that domestic legislation which has been enacted in order to give effect to the UK’s obligations under an international convention or treaty should be construed in the same sense as the convention or treaty if the words of the statute are reasonably capable of bearing that meaning: The Jade [1976] 1 WLR 430 at 436 (Lord Diplock). It is generally thought that this rule imposes a weaker duty of interpretation on the courts than the Marleasing principle.

A European directive falls to be interpreted according to principles of interpretation of European Union legislation developed by the Court of Justice of the European Union. The basic rule of interpretation, which has been frequently reiterated by the European Court of Justice, is that stated in Case C-306/05 Sociedad General de Autores y Editores de España v Rafael Hoteles SA [2006] ECR I-11519 at [34]:

“According to settled case-law, in interpreting a provision of Community law it is necessary to consider not only its wording, but also the context in which it occurs and the objectives pursued by the rules of which it is part (see, in particular, Case C-156/98 Germany v Commission [2000] ECR I-6857, paragraph 50, and Case C-53/05 Commission v Portugal [2006] ECR I-6215, paragraph 20)”.

As is well known, in applying this rule the ECJ routinely refers to the recitals of the measure as well as its operative provisions and frequently refers to pre-legislative materials such as the Explanatory Memoranda which accompany the Commission’s legislative proposals.

Another rule of interpretation applied by the ECJ is that stated in SGAE at [35]:

“Moreover, Community legislation must, so far as possible, be interpreted in a manner that is consistent with international law, in particular where its provisions are intended specifically to give effect to an international agreement concluded by the Community (see, in particular, Case C-341/95 Bettati [1998] ECR I-4355, paragraph 20 and the case-law cited).”

In that case the Court interpreted Article 3(1) of the Information Society Directive in accordance with Article 8 of the WIPO Copyright Treaty.

It follows that, in a field of intellectual property law where the European Union has legislated, national courts must interpret both European and domestic legislation as far as possible in the light of the wording and purpose of relevant international agreements to which the EU is a party, such as TRIPS and the WIPO Copyright Treaty: see Case C-53/96 Hermès International v FHT Marketing Choice BV [1998] ECR I-3603 at [28]; Joined Cases C-300/98 and C-392/98 Parfums Christian Dior SA v Tuk Consultancy BV [2000] ECR I-11307 at [47]; Case C-89/99 Schieving-Nijstad VOF v Groeneveld [2001] ECR I-5851 at [35]; Case C-245/02 Anheuser-Busch Inc v Budejovicky Budvar NP [2004] ECR I-10989 at [55]-[57]; and Case C-431/05 Merck Genéricos – Produtos Farmacêuticos Lda v Merck & Co Inc [2007] ECR I-7001 at [35].

It can be seen from sections 16(3) and 17(2) of the 1988 Act that it is an infringement of the copyright in a literary work to reproduce any substantial part of it. The general principles to be applied in determining whether a substantial part has been reproduced are summarised in Copinger and Skone James (15^th ed, Sweet & Maxwell, 2005) at §7-27. They include the following:

“(a)

As already stated, the overriding question is whether, in creating the defendant's work, substantial use has been made of the skill and labour which went into the creation of the claimant’s work and thus those features which made it an original work.

(b)

The quality or importance [Catnic Components Ltd v Hill & Smith Ltd [1982] R.P.C. 183 at 223] of what has been taken is much more important than the quantity. [There are statements in many authorities to this effect. See, e.g. Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 W.L.R. 273 at 276; Designers Guild Ltd v Russell Williams (Textiles) Ltd [2001] 1 W.L.R. 2416, at 2422, 2425, 2431…..] The issue thus depends therefore not just on the physical amount taken but on its substantial significance [Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 W.L.R. 273, at 283.] or importance to the copyright work, [Designers Guild Ltd v Russell Williams (Textiles) Ltd [2001] 1 W.L.R. 2416, per Lord Millett, at 2426.] so that the quality, or importance, of the part is frequently more significant than the proportion which the borrowed part bears to the whole. In this context, expressions such as quality or importance need to be properly understood. A literary work may convey a very important idea but, as already discussed, the law of copyright is concerned with the protection of the expression of such ideas, not the idea itself. … Quality and importance must therefore be understood in terms of the features of the work which made it an original work in the first place. It follows that the quality relevant for the purposes of substantiality in the case of a literary work refers to the literary originality of that which has been copied. …

(c)

Depending on the circumstances, the question may depend on whether what has been taken is novel or striking, or is merely a commonplace arrangement of words or well known material. [Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 W.L.R. 273, at 276.] In this respect it may be a helpful shortcut to ask whether the part taken could itself be the subject of copyright, [….] although this should not be used as a substitute for the proper and full test of substantial part. [Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 W.L.R. 273, at 277.] In any event it is in reality merely a restatement of the basic question, which is whether use has been made of those features of the claimants work which, by reason of the skill and labour employed in their production, constitute it an original copyright work.

(d)

As a corollary of the last point, the more simple or lacking in substantial originality the copyright work, the greater the degree of taking will be needed before the substantial part test is satisfied. In the case of works of little originality, almost exact copying will normally be required to amount to infringement. […]

(e)

The more abstract and simple a copied idea, the less likely it is to constitute a substantial part. Originality, in the sense of the contribution of the authors skill and labour, tends to lie in the detail with which the basic idea is presented. [Designers Guild Ltd v Russell Williams (Textiles) Ltd [2001] 1 W.L.R. 2416 , at 2423, per Lord Hoffmann: ….]

(f)

In general it is wrong to dissect the claimant’s work, taking each part which has been copied and asking whether each part could be the subject of copyright if it had stood alone. It is the work as a whole which must be considered, [ibid., at 2421.] particularly where the originality of the work lies in the creation of the work as a whole. …

…

(j)

If what has been taken from the claimants work is material which was not original to the maker of the work, because for example it was copied from another work, then it should be disregarded in deciding this issue. [Warwick Film Productions v Eisinger [1969] Ch 1969] If, however, part of the originality of the claimant’s work consisted of the choosing and collecting together of unoriginal material, and the defendant has made unfair use of this labour and skill, then the test will be satisfied. This can often be seen in compilation cases … On the other hand, that which would not attract copyright except by reason of its collocation will, when robbed of that collocation, not be a substantial part of the copyright. [Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 W.L.R. 273 at 293.] It is, therefore, often important to isolate what it is the real worth of a work from the copyright point of view, in particular what parts of the work constitute the original contribution of the author and thus represent the skill or labour of his creation.

…”

The leading authority on what amounts to reproduction of a substantial part in cases of what is sometimes referred to as “non-textual” copying is the decision of the House of Lords in Designers Guild Ltd v Russell Williams (Textiles) Ltd [2001] 1 WLR 2416, [2001] FSR 11 (the latter report has the advantage of added paragraph numbers, and hence has frequently been quoted from in subsequent cases) which is repeatedly cited in the above passage. Although this was a case about the copying of an artistic work, Lord Hoffmann also considered literary works in a passage in his speech which has subsequently been cited in several cases concerning literary works:

“[23] It is often said, as Morritt L.J. said in this case, that copyright subsists not in ideas but in the form in which the ideas are expressed. The distinction between expression and ideas finds a place in the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) (O.J. 1994 L. 336 , p. 213), to which the United Kingdom is a party (see article 9.2: ‘Copyright protection shall extend to expressions and not to ideas…’). Nevertheless, it needs to be handled with care. What does it mean? As Lord Hailsham of St. Marylebone said in L.B. (Plastics) Ltd v. Swish Products Ltd. [1979] R.P.C. 551, 629, ‘it all depends on what you mean by “ideas”’.

[24] Plainly there can be no copyright in an idea which is merely in the head, which has not been expressed in copyrightable form, as a literary, dramatic, musical or artistic work. But the distinction between ideas and expression cannot mean anything so trivial as that. On the other hand, every element in the expression of an artistic work (unless it got there by accident or compulsion) is the expression of an idea on the part of the author. It represents her choice to paint stripes rather than polka dots, flowers rather than tadpoles, use one colour and brush technique rather than another, and so on. The expression of these ideas is protected, both as a cumulative whole and also to the extent to which they form a ‘substantial part’ of the work. Although the term ‘substantial part’ might suggest a quantitative test, or at least the ability to identify some discrete part which, on quantitative or qualitative grounds, can be regarded as substantial, it is clear upon the authorities that neither is the correct test. Ladbroke (Football) Ltd. v. William Hill (Football) Ltd. [1964] 1 W.L.R. 273 establishes that substantiality depends upon quality rather than quantity (Lord Reid, at p. 276, Lord Evershed, at p. 283, Lord Hodson, at p. 288, Lord Pearce, at p. 293). And there are numerous authorities which show that the ‘part’ which is regarded as substantial can be a feature or combination of features of the work, abstracted from it rather than forming a discrete part. That is what the judge found to have been copied in this case. Or to take another example, the original elements in the plot of a play or novel may be a substantial part, so that copyright may be infringed by a work which does not reproduce a single sentence of the original. If one asks what is being protected in such a case, it is difficult to give any answer except that it is an idea expressed in the copyright work.

[25] My Lords, if one examines the cases in which the distinction between ideas and the expression of ideas has been given effect, I think it will be found that they support two quite distinct propositions. The first is that a copyright work may express certain ideas which are not protected because they have no connection with the literary, dramatic, musical or artistic nature of the work. It is on this ground that, for example, a literary work which describes a system or invention does not entitle the author to claim protection for his system or invention as such. The same is true of an inventive concept expressed in an artistic work. However striking or original it may be, others are (in the absence of patent protection) free to express it in works of their own: see Kleeneze Ltd. v. D.R.G. (U.K.) Ltd. [1984] F.S.R. 399. The other proposition is that certain ideas expressed by a copyright work may not be protected because, although they are ideas of a literary, dramatic or artistic nature, they are not original, or so commonplace as not to form a substantial part of the work. Kenrick & Co. v. Lawrence & Co. (1890) 25 Q.B.D. 99 is a well known example. It is on this ground that the mere notion of combining stripes and flowers would not have amounted to a substantial part of the plaintiff's work. At that level of abstraction, the idea, though expressed in the design, would not have represented sufficient of the author's skill and labour as to attract copyright protection.

[26] Generally speaking, in cases of artistic copyright, the more abstract and simple the copied idea, the less likely it is to constitute a substantial part. Originality, in the sense of the contribution of the author's skill and labour, tends to lie in the detail with which the basic idea is presented. Copyright law protects foxes better than hedgehogs. In this case, however, the elements which the judge found to have been copied went well beyond the banal and I think that the judge was amply justified in deciding that they formed a substantial part of the originality of the work.”

The penultimate sentence of this passage was explained by Arden LJ in L. Woolley Jewellers Ltd v A & A Jewellery Ltd [2002] EWCA Civ 1119, [2003] FSR 15:

“[9] The judge did not explain the last sentence of his citation from Lord Hoffmann's speech in the Designers Guild case, in which the judge had appeared as leading counsel. Indeed Lord Hoffmann did not elaborate on his reference to hedgehogs and foxes. However, it appears that it is a reference to a fragment of Greek poetry of the seventh century BC, with which the late Sir Isaiah Berlin begins his famous essay on Tolstoy:

‘There is a line among the fragments of a Greek poet Archilochus which says “The fox knows many things, but the hedgehog knows one big thing”’.’ (The Hedgehog and the Fox: An Essay on Tolstoy's View of History by Isaiah Berlin. (1953, as revised in 1978) (Phoenix) (1999) p.3).

[10] Sir Isaiah points out that scholars have differed about the correct interpretation of these ‘dark’ words. They may, on the one hand, mean no more than that the fox, for all his cunning, is defeated by the hedgehog's one defence. But the fragment may also be taken figuratively as contrasting those with a single central vision and organising principle as against those who pursue many ends, often unrelated or contradictory. It was, I think, in the figurative sense that Lord Hoffmann was using his metaphor.”

In the subsequent case of Baigent v Random House Group Ltd [2007] EWCA Civ 247, [2007] FSR 24 Lloyd LJ said at [5]:

“Copyright does not subsist in ideas; it protects the expression of ideas, not the ideas themselves. No clear principle is or could be laid down in the cases in order to tell whether what is sought to be protected is on the ideas side of the dividing line, or on the expression side.”

He went on to quote parts of Lord Hoffmann’s speech in Designers’ Guild at [24] to [25].

Mummery LJ said:

“145.

… The decided cases help in identifying the relevant necessary and sufficient conditions for substantiality. Thus, it is not necessary for the actual language of the copyright work to be copied or even for similar words to be used tracking, like a translation, the language of the copyright work. It is sufficient to establish that there has been substantial copying of the original collection, selection, arrangement, and structure of literary material, even of material that is not in itself the subject of copyright.

146.

It is not, however, sufficient for the alleged infringing work simply to replicate or use items of information, facts, ideas, theories, arguments, themes and so on derived from the original copyright work.

147.

I agree with Lloyd LJ that no clear principle can be laid down on how or where to draw the line between the legitimate use of the ideas expressed and the unlawful copying of their expression.”

At present, the leading case on the copyright protection of computer programs in this country is the judgment of Pumfrey J in Navitaire v easyJet (supra). This was the first substantial decision on the copyright protection of computer software in this country on facts occurring after the implementation of the Software Directive. Before turning to the detail of the case itself, it is worth briefly recording the qualifications of Pumfrey J to decide it. In addition to 22 years’ experience as a barrister, and 6 years’ experience as a judge, specialising in intellectual property law, he had a degree in physics and considerable experience as an amateur computer programmer. He was therefore able to approach the case with both expertise in copyright law and a good understanding of the technology.

In Navitaire the claimant, Navitaire, was the developer of an airline reservation system called OpenRes. The first defendant, easyJet, was an airline. easyJet was a licensee of OpenRes and its staff were accustomed to using it. Much of OpenRes had what would now be regarded as a rather old-fashioned “green screen” user interface. This involved users typing in commands using a set of command codes such as A for an availability query, in many cases together with appropriate data (such as departure date, city pair, return date, fare class, etc) in a defined syntax. The results of executing the commands would then be somewhat crudely displayed on VT100 screens, which were character-based terminals, either as individual screen displays or as longer reports. In addition, part of OpenRes had a more modern graphical user interface (or GUI) enabling users to control the system by clicking on icons. easyJet commissioned the second defendant, BulletProof, to develop a replacement system called eRes that had the same user interfaces as OpenRes. In short, what easyJet wanted, and BulletProof supplied, was a “‘drop-in’ replacement” for OpenRes (see the judgment of Pumfrey J at [109](iv)).

Navitaire was a complicated case involving a range of factual and legal issues, but the most important issue was concisely identified by Pumfrey J at the outset of his judgment at [2] and [5]:

“2.

The case … raises starkly an issue of considerable importance in the law of copyright. While there are comparatively minor allegations of infringement by copying of certain code and an allegation in relation to the databases, that I summarise below, the striking feature of this action is that Navitaire does not suggest that easyJet or BulletProof ever had access to the source code of the OpenRes system. What is alleged, and not disputed, is that easyJet wanted a new system that was substantially indistinguishable from the OpenRes system, as easyJet used it, in respect of its ‘user interface’. This term is used to denote the appearance that the running software presents to the user, who may be an agent in a call centre or a private individual seeking to make a booking by use of the World Wide Web. It substantially achieved this far-from-simple goal. It is not in dispute that none of the underlying software in any way resembles that of OpenRes, save that it acts upon identical or very similar inputs and produces very similar results, but it is said that the copyright in OpenRes is infringed by what was called ‘non-textual copying.’

…

5.

There is here an issue of general importance. To emulate the action of a piece of software by the writing of other software that has no internal similarity to the first but is deliberately designed to ‘look’ the same and achieve the same results is far from uncommon. If Navitaire are right in their most far-reaching submission, much of such work may amount to the infringement of copyright in the original computer program, even if the alleged infringer had no access to the source code for it and did not investigate or decompile the executable program.”

So far as it is relevant for present purposes, Navitaire advanced three main claims (there were various other claims which can be ignored). First, Navitaire contended that the defendants had infringed the copyright which subsisted in the commands as copyright works distinct from the source code of OpenRes by copying a substantial proportion of the commands. This claim was put in three different ways: copyright subsisted in each command (meaning the code letter or letters, any associated syntax and any associated symbols recognised by that syntax such as brackets, slashes, etc) as a copyright work in its own right; in the alternative, copyright subsisted in so-called “complex” commands (essentially, those which did have an associated syntax); and in the further alternative copyright subsisted in the set of commands as a compilation. Secondly, Navitaire contended that the defendants had infringed the copyright which subsisted in a number of individual screen displays and reports as copyright works distinct from the source code of OpenRes by copying those displays and reports. As indicated above, this claim concerned two different types of screen displays, namely VT100 screen displays and reports and GUI screen displays. Thirdly, Navitaire contended that the defendants had infringed the copyright in the source code of OpenRes by copying what prior to trial Navitaire and its expert witness described as the “business logic” of OpenRes, although at trial counsel for Navitaire preferred to describe this claim as one of “non-textual copying” analogous to copying the plot of a novel. Save for the claim in relation to the GUI screens, Pumfrey J dismissed each of these claims.

For present purposes, three sets of passages in Pumfrey J’s reasoning are of particular importance. The first concerns the claim in respect of the commands. At [79]-[80] he rejected Navitaire’s contention that the individual command codes were copyright works on the ground that they were not literary works. At [81]-[84] he rejected Navitaire’s contention that the complex commands were copyright works on the ground that the syntaxes of the complex commands were only parsed by (that is to say, recognised by, but not stated in) the software, and therefore again the complex commands were not literary works. He went on:

“85.

However, I am acutely conscious that this may not be a satisfactory answer to the problem. It depends too much upon the way in which OpenRes was written. In answer to a question from me, it was made clear by Dr Hunt that it would be possible to record the command names and their syntax expressly and use a program (a ‘parser generator’) automatically to construct a parser that recognised such commands accompanied by arguments according to such a syntax. The commands and their syntax would, in such a case, be recognisable as such in the source code for the parser generator. In such a case the copyright owner could point to a written work describing exactly how the alleged infringer's program parsed the code and the consequences would be very different. I am most reluctant to come to a conclusion that depends upon the happenstance of the manner in which the programmer decided to set about constructing his parser. In either case, the ultimate result is a computer program which recognises the input according to the prescribed rules.

86.

I consider that the better approach is to take the view that it is not possible to infringe the copyright that subsists either in the source code for a parser or in the source code for a parser generator by observing the behaviour of the final program and constructing another program to do the same thing. In expressing this view, I am verging on drawing a distinction between the ‘idea’ of the program and its ‘expression’, and, Navitaire contends, that is not a distinction known to English law and is entirely contrary to the observations of Jacob J. (as he then was) in the leading case of Ibcos Computers Ltd v Barclays Mercantile Highland Finance Ltd [1994] F.S.R. 275.

87.

I think the problem should be approached in the following way. To define a series of commands and their syntax to be recognised by the computer is to define a computer language. It is exactly the same as defining a language such as BASIC or a simple language to control a calculator program. A program consists of a statement or series of statements in that ‘language’. Thus, to take the availability command as an example, one would say that the language includes an availability statement that starts with the letter ‘A’ and one of the permissible forms of which is A[date][City-Pair]┘. An example of a statement that will be parsed as an allowable statement to control the computer in accordance with this language is A13JUNLTNAMS┘. Recitals 13, 15 and 14 (fn They are not numbered in the original, but this is a convenient way to refer to them) of the Software Directive are as follows: ….

88.

The Software Directive is a harmonising measure. I must construe any implementing provision in accordance with it: if the implementing provision means what it should, the Directive alone need be consulted: if it departs from the Directive, then the latter has been incorrectly transposed into UK law. (fn See Case C-106/89: Marleasing [1990] E.C.R. I-4135, and in Customs and Excise Commissioners v Century Life Plc [2001] S.T.C. 38.) The recitals quoted are said by Laddie, Prescott & Vitoria, Modern Law of Copyright and Designs, (3rd edn), para.34.19 to make it clear that ‘computer languages are not included in the protection afforded to computer programs’. With this conclusion I agree, although the point cannot be said to be entirely clear and will require to be referred to the Court of Justice. In my view, the principle extends to ad hoc languages of the kind with which I am here concerned, that is, a defined user command interface. It does not matter how the ‘language’ of the interface is defined. It may be defined formally or it may be defined only by the code that recognises it. Either way, copyright does not subsist in it. This is of course not to suggest that the expression of a program in a particular language is not entitled to copyright. Quite the reverse. What this recital [sic], and the associated dispositive provision of Art.1(2), appear to be intended to do, is to keep the language free for use, but not the ideas expressed in it: …

89.

There is here more than an echo of a conceptual distinction between idea and expression, but it is unprofitable to pursue this approach in the light of the express reference to computer languages and interfaces in the recital and to the interfaces in Art.1(2).”

He then considered Navitaire’s claim to copyright in the set of commands as a compilation. He rejected this claim at [90]-[92] on the grounds both that it was not a compilation and that the commands were a computer language and thus not entitled to copyright. He added at [94]:

“Copyright protection for computer software is given, but I do not feel that the courts should be astute to extend that protection into a region where only the functional effects of a program are in issue. There is a respectable case for saying that copyright is not, in general, concerned with functional effects, and there is some advantage in a bright line rule protecting only the claimant's embodiment of the function in software and not some superset of that software. The case is not truly analogous with the plot of a novel, because the plot is part of the work itself. The user interface is not part of the work itself. One could permute all the letters and other codes in the command names, and it would still work in the same way, and all that would be lost is a modest mnemonic advantage. To approach the problem in this way may at least be consistent with the distinction between idea and expression that finds its way into the Software Directive, but, of course, it draws the line between idea and expression in a particular place which some would say lies too far on the side of expression. I think, however, that such is the independence of the particular form of the actual codes used from the overall functioning of the software that it is legitimate to separate them in this way, and not to afford them separate protection when the underlying software is not even arguably copied.”

The second set of passages concerns Navitaire’s claims in respect of the screen displays. At [96] Pumfrey J concluded that the individual VT100 screen displays were properly to be viewed as tables, and hence as literary in character for the purposes of copyright. He then said:

“They are, in my view, ‘ideas which underlie its interfaces’ in the sense used in Art.1(2) of the Directive: they provide the static framework for the display of the dynamic data which it is the task of the software to produce.”

It appears from this that he considered that the VT100 screen displays were not protected by copyright. He went on in Annexe 5 of the judgment at [308] to consider the individual displays, from which it appears that he considered that the defendants had not reproduced a substantial part of anything protected by literary copyright in any event. His analysis in respect of the reports in Annexe 6 of the judgment is to the same effect.

By contrast, he held at [93]-[94] that the GUI screens stood in a different position, because they qualified as artistic works:

“The Directive is concerned only with the protection of computer programs as literary works, and I do not read it as having any impact on relevant artistic copyrights.”

Thus the claim in respect of the GUI screens succeeded for the reasons given in more detail in Annexe 5 at [309]-[317].

The third set of passages concerns Navitaire’s claim in respect of the “business logic” or non-textual copying. Pumfrey J first considered in more detail the nature of Navitaire’s case in respect of this claim, beginning with the following observation at [107]:

“As I have indicated above, the case advanced by Navitaire is based on the fact that the functions of OpenRes and eRes are identical to the user so far as the aspects of the system of interest to easyJet are concerned. The case had its origin in the suggestion that what was called the ‘business logic’ of OpenRes had been appropriated.”

Having referred to some evidence on the subject given by Navitaire’s expert witness, he went on:

“112.

I shall return to first principles. For present purposes, a computer running a particular program is a deterministic machine. A particular input to the machine will produce a predictable result derived from all previous inputs to the machine. If therefore one studies a machine in operation, it should be possible to identify the machine's response to all possible sequences of inputs, and so construct a new machine that operates to give the same outputs for the same sequences of inputs by writing an appropriate program. Navitaire contend that if this is done, it follows axiomatically that any copyright in the source code for the first machine must be infringed in writing the second program. Indeed, it was urged on me at an earlier hearing that it was unnecessary to consider any of the source code for the OpenRes system in determining whether there had been copying of a substantial part of the copyright(s) subsisting in the source code for it.

113.

There is no doubt that easyJet and BulletProof had no access to the source code of OpenRes, and it is not in dispute that in languages used, actual code and architecture (subject to the claim in respect of the database) the systems are quite different. There is no suggestion that the eRes code represents a translation or adaptation of the OpenRes code. The term ‘non-textual copying’ might be replaced by the more accurate ‘copying without access to the thing copied, directly or indirectly’.

114.

The claim depends first upon the contention that the manner in which a machine behaves under the control of a program represents part of the skill and labour that went into the program. This is not an unreasonable observation. On the contrary, it is the whole object of the programmer to get the computer to behave in the required manner.

115.

To copy an operating machine in this manner avoids the need to conduct any systems analysis or the production of functional specifications. Thus, it may be observed that although the copyist has not avoided the need to write software to achieve the desired result, he has avoided the need to identify the result by any of the normal methods of analysis that either precede or accompany the writing of a substantial piece of business software….”

At [119]-[121] Pumfrey J considered earlier case law on copying the plot of a novel or play (Holland v Vivian van Damm Productions Ltd [1936–45] MacG CC 69, Harman Pictures NV v Osborne [1967] 1 WLR 723 and the speech of Lord Hoffmann in Designers’ Guild at [23]-[24]) and on copying compilations (Jarrold v Houlston (1857) 3 K&J 708 and Macmillan & Co Ltd v Cooper (1923) 93 LJPC 113). At [122]-[124] he considered two earlier cases on copying computer programs which pre-dated the implementation of the Software Directive in the United Kingdom, John Richardson Computers Ltd v Flanders [1993] FSR 497 and Ibcos Computers Ltd v Barclays Mercantile Highland Finance Ltd [1994] FSR 275.

He then said:

“125.

This does not answer the question with which I am confronted, which is peculiar, I believe, to computer programs. The reason it is a new problem is that two completely different computer programs can produce an identical result: not a result identical at some level of abstraction but identical at any level of abstraction. This is so even if the author of one has no access at all to the other but only to its results. The analogy with a plot is for this reason a poor one. It is a poor one for other reasons as well. To say these programs possess a plot is precisely like saying that the book of instructions for a booking clerk acting manually has a plot: but a book of instructions has no theme, no events, and does not have a narrative flow. Nor does a computer program, particularly one whose behaviour depends upon the history of its inputs in any given transaction. It does not have a plot, merely a series of pre-defined operations intended to achieve the desired result in response to the requests of the customer.

126.

The view in favour of the Navitaire case is expressed concisely by the authors of The Modern Law of Copyright and Designs , 3rd ed (2000) (I have assumed that when they speak of ‘obtains … from the original program’ they do not mean obtain directly, but indirectly from watching the program work), at para 34.64:

‘For instance, the writing of a financing program may require as part of the task a careful elucidation of the relevant tax regulations—so that they may be reduced to a series of unambiguous statements—and it will be evident to any lawyer that this alone will probably involve a very large amount of work. A competitor might write a program of his own in a different computer language and arranged in a different way and with many improvements of his own but if he obtains the rules for calculating the tax from the original program instead of working these out for himself it is hard to see why he should not be considered a plagiarist.’

127.

There is a counter-example that throws some light on the nature of the problem. Take the example of a chef who invents a new pudding. After a lot of work he gets a satisfactory result, and thereafter his puddings are always made using his written recipe, undoubtedly a literary work. Along comes a competitor who likes the pudding and resolves to make it himself. Ultimately, after much culinary labour, he succeeds in emulating the earlier result, and he records his recipe. Is the later recipe an infringement of the earlier, as the end result, the plot and purpose of both (the pudding) is the same? I believe the answer is no.

128.

I think that the answer to the problem is to be gathered from the passage in Lord Hoffmann's speech immediately following that quoted above …. from the Designers' Guild case: [he proceeded to cite [25]-[26]].

129.

The questions in the present case are both a lack of substantiality and the nature of the skill and labour to be protected. Navitaire's computer program invites input in a manner excluded from copyright protection, outputs its results in a form excluded from copyright protection and creates a record of a reservation in the name of a particular passenger on a particular flight. What is left when the interface aspects of the case are disregarded is the business function of carrying out the transaction and creating the record, because none of the code was read or copied by the defendants. It is right that those responsible for devising OpenRes envisaged this as the end result for their program: but that is not relevant skill and labour. In my judgment, this claim for non-textual copying should fail.

130.

I do not come to this conclusion with any regret. If it is the policy of the Software Directive to exclude both computer languages and the underlying ideas of the interfaces from protection, then it should not be possible to circumvent these exclusions by seeking to identify some overall function or functions that it is the sole purpose of the interface to invoke and relying on those instead. As a matter of policy also, it seems to me that to permit the ‘business logic’ of a program to attract protection through the literary copyright afforded to the program itself is an unjustifiable extension of copyright protection into a field where I am far from satisfied that it is appropriate.”

In the subsequent case of Nova Productions Ltd v Mazooma Games Ltd [2007] EWCA Civ 219, [2007] Bus LR 1032, the claimant was the developer of a video game called Pocket Money which simulated a game of pool. The defendants developed two video games called Jackpot Pool and Trick Shot which also simulated pool games and which the claimant alleged infringed its copyrights. A remarkable feature of the case was that, although there were certain similarities between the defendants’ games and the claimant’s at a high level of generality, at a more detailed level they were very different.

In the Court of Appeal the claimant advanced two main claims. The first was that the sequence of the frames (i.e. the images displayed to the user) of its game was an artistic work. It was common ground that the individual frames were artistic works and that there had been no frame-for-frame reproduction by the defendants. The Court of Appeal held that there was no separate copyright in the sequence of frames, and that in any event Kitchin J had rightly held that the defendants had not reproduced a substantial part of the claimant’s work.

The second claim was that the defendants had infringed the copyright in two literary works, namely the claimant’s designer’s design notes and the source code of the claimant’s game itself. The Court of Appeal held that Kitchin J had rightly rejected this claim on two distinct, but related, bases. The first was that the similarities between the defendants’ programs and the claimant’s works were so general as to amount to ideas, rather than part of the expression of the claimant’s works, and thus the defendants had not reproduced a substantial part of those works. The second was that, applying the principles set out by Pumfrey J in Navitaire, nothing which was protectable by copyright in computer software had been copied.

So far as the first ground is concerned, Jacob LJ (with whom Sir Andrew Morritt Ch and Lloyd LJ agreed) said:

“31.

Mr Howe had to face the formidable objection created by Art. 1.2 of the [Software] Directive and recitals 13 and 15. To my mind these provisions are abundantly clear. The well-known dichotomy between an idea and its individual expression is intended to apply and does to copyright in computer software. When I say ‘well-known’ I mean not just known to copyright lawyers of one country but well-known all over the world. Recital 15 refers to the protection of the expression of ideas as being ‘in accordance with the legislation and jurisprudence of the Member States and the international copyright conventions’ and is clearly a reference to this dichotomy. The TRIPS agreement of 1994 likewise recognises this dichotomy, see particularly Art, 9.2.

32.

Mr Howe suggested that the dichotomy was intended to apply only to ideas which underlie an element of a program – what he called a ‘building block’. He cited (as did the Judge) what Lord Hoffmann said about the dichotomy outside the context of computer programs in Designers Guild: [quoting [25]].

33.

As regards [Lord Hoffmann’s] first proposition I said much the same thing in Ibcos Computers v. Barclays Mercantile [1994] FSR 275 at p.291

‘The true position is that where an ‘idea’ is sufficiently general, then even if an original work embodies it, the mere taking of that idea will not infringe. But if the ‘idea’ is detailed, then there may be infringement. It is a question of degree. The same applies whether the work is functional or not, and whether visual or literary. In the latter field the taking of a plot (i.e. the “idea”) of a novel or play can certainly infringe—if that plot is a substantial part of the copyright work. As Judge Learned Hand said (speaking of the distinction between “idea” and “expression”): “Nobody has ever been able to fix that boundary and nobody ever can”, Nichols v Universal Pictures (1930) 45 F. (2d) 119.’

34.

Mr Howe then submitted that the ‘idea’ of the cue pulsing with the power-meter could not be discounted within Lord Hoffmann's first category because here we are concerned with copyright in a computer program. You cannot say the "idea" has no connection with the nature of the work. Nor did it fall within the second category because it was not held ‘commonplace,’ merely ‘obvious.’ He sought to bolster the argument by reference to the travaux préparatoires to the Directive. At the very least, he submitted, the position was unclear and that we should refer some questions to the European Court of Justice pursuant to Art. 234 of the Treaty.

35.

I reject all of that. First I think the fact that we are considering a computer program copyright does not in any way preclude a mere ‘idea’ as to what the program should do from being excluded as having nothing to do with the nature of the work. The nature of the work is a computer program having all the necessary coding to function. The general idea is only faintly related to that – no different from the relationship of the general idea of a plastic letter-box draught excluder to the artistic works consisting of the drawings for a particular excluder in the Kleeneze case. Indeed I have to say that, as Mr Howe waxed lyrical about the combination of features in the animation, he sounded more like counsel in a patent case than one in a copyright case. Not all of the skill which goes into a copyright work is protected – the obvious example being the skill involved in creating an invention which is then described in a literary work. An idea consisting of a combination of ideas is still just an idea. That is as true for ideas in a computer program as for any other copyright work.

36.

Nor am I impressed by Mr Howe's attempt to limit the dichotomy to ‘building blocks’. He sought to do this by reference to recital 14 which refers to ‘logic, algorithms and programming languages’ as comprising ‘ideas and principles.’ I see no reason to suppose that recital 13 is thereby limited. Recital 14 is clearly drawn on the basis that the basic position of recital 13 - no protection for ideas and principles - applies also to those specified matters.

37.

The same conclusion is reached if one considers TRIPS. Although normally a UK Act is not to be construed by reference to a later international Treaty, I note that Lord Hoffmann considered TRIPS to be of relevance to our domestic copyright law in Designers Guild (…). Here the position is much clearer because we are dealing with EU law. The ECJ has held that TRIPS (to which the EU as well as its Member States is a party) is relevant to the construction of earlier EU legislation concerned with intellectual property [citing Case C-89/99 Schieving-Nijstad v Groeneveld].

38.

Mr Howe sought to escape from the fact that we should construe the UK Act in accordance with the Directive which should be construed so far as possible as to conform to TRIPS by a further submission: that TRIPS was only concerned with minimum standards for intellectual property rights and that its signatories were free to provide for greater rights. So, he said, the EU could provide that copyright protection extended to ideas if it so wanted. Now it is in general true that a party to TRIPS can provide more extensive protection than called for TRIPS, see Art.1. But the concluding words of the first sentence of Art.1 add ‘provided that such protection does not contravene the provisions of this Agreement’. Art. 9.2 positively provides that ‘copyright protection shall extend to expressions and not to ideas etc. as such’. So in this instance TRIPS lays down a positive rule as to the point beyond which copyright protection may not go. To protect by copyright mere ideas as such would contravene TRIPS. The Software Directive must be construed so as to conform to TRIPS and so must be construed as not to protect ideas as such.

39.

As to the travaux Mr Howe took us through the following:

i)

the original Commission "Proposal for a Council Directive on the legal protection of computer programs" COM(88) 816 final –SYN 183, submitted on 5 January 1989 (89/C91/05);

ii)

the amended proposal COM (90) 509 final –SYN 183 (1990/C 320/11), submitted on 18 October 1990; and

iii)

the Opinion of the Economic and Social Committee.

40.

Mr Howe was unable to point to any clear unequivocal statement anywhere suggesting that copyright in computer programs should extend to ideas. Given that state of affairs it would be a waste of time to set out all the material in detail. It is sufficient to record that the submission is based on an argument by implication from unexplained alterations from the original proposal. In particular the original proposal did not use the expression ‘element of a computer program’ in proposed Art.1.2 but merely to ‘the ideas, principles, logic, algorithms or programming languages.’ I cannot extract from this any intention to extend protection to ideas provided they are not ‘elements.’

41.

I am reinforced in my view by the fact that the Economic and Social Committee (whose report is specifically recited in the Directive as being a document to which the Council had regard in making the Directive) clearly did not think the proposal was limited as suggested by Mr Howe. It said tersely:

‘There is no dispute that 'ideas and principles' are outside the protection of the law of copyright.’”

So far as the second reason is concerned, Jacob LJ quoted the passage from Pumfrey J’s judgment in Navitaire at [125]-[130] that I have set out above and then said:

“48.

Mr Howe attacked that. I quote his skeleton argument:

‘this analogy is a poor one. The reason is that the first chef has deployed two quite distinct types of skill and labour. The first is the skill of devising a recipe, a skill which on no view forms part of the skill and labour protected by copyright in literary works. The second is skill and labour in reducing the recipe he has devised to written form. A copyist who copies from his pudding rather than from his recipe book may appropriate the former skill and labour but none of the latter. By contrast, a copyist who copies the function of a computer program to write his own program to achieve the same results is clearly appropriating part of the skill and labour expended in designing the program.’

49.

He further developed the argument basing himself on the recital 7 of the Directive. … He asked us to suppose a case where there are two clear stages in the making of a program—a first stage where the designer sets out all the things he wants the program to be able do and a second stage (which may be by a different person) where the actual program code is written. Mr Howe contended that the first stage was intended to be protected as such, even if it consisted only of ideas as to what the program should do. Going back to the analogy, the ‘preparatory work; for the program is like the skill of devising the recipe and the actual program writing like the reduction of the recipe to written form. The difference, he submitted, is that for computer programs, unlike the recipe, the preparatory work is to be protected.

50.

I reject the argument. The reason is simple. The Directive does not say that mere ideas by way of preparatory design work are to be protected. As I have said it makes it clear that for computer programs as a whole (which includes their preparatory design work) ideas are not to be protected. What is protected by way of preparatory design work is that work as a literary work—the expression of the designs which are to go into the ultimate program, not the ideas themselves.

51.

So, for example, if Mr Jones had actually written a description of the pulsing, rotating cue, and synchronised power meter his description would (if not too trivial at least) be protected as a literary work. People could not copy that. But they could use the same idea. Similarly and more generally, a written work consisting of a specification of the functions of an intended computer program will attract protection as a literary work. But the functions themselves do not. Of course to someone familiar with the prior English law it is self-evident that copyright could subsist in such a description. The fact that a work can get copyright even if mundane, is old and familiar to an English lawyer. But the Directive needed to say that protection as a literary work should be provided for preparatory design work because not all member states under their existing laws necessarily provided that. That is the whole point of the Directive—and the clear reason for it is recited in article 1.

52.

So I think Mr Howe's attack on the Navitaire case [2006] RPC 111 fails. The reasoning in the Navitaire case provides a second reason for dismissing this appeal. Pumfrey J was quite right to say that merely making a program which will emulate another but which in no way involves copying the program code or any of the program's graphics is legitimate.

53.

Finally Mr Howe suggested the law was sufficiently uncertain as to warrant a reference being made to the European Court of Justice. No less than six elaborate draft questions were supplied. I do not think it is necessary to make any reference to resolve this case. It is wholly unrealistic to suppose that the European Court of Justice would hold that copyright protection was to be given to ideas at such a high level of abstraction as those in this case.”

The claimant subsequently petitioned the House of Lords for leave to appeal and suggested that it would be appropriate for the House to refer questions to the ECJ. The House refused leave to appeal.

The current domestic case law on section 30(1) of the 1988 Act is summarised on Copinger at §§9-35 to 9-43. These paragraphs refer both backward and forwards to certain other paragraphs which are relevant. For present purposes the following passages are particularly pertinent:

“9-23 Approach to construction. … As to the various expressions, ‘criticism or review’ and ‘reporting current events’, these are expressions which should be interpreted liberally and their precise boundaries cannot be plotted: the nearer any use comes to the boundaries, unplotted as they are, the less likely is the use to be fair. [Pro Sieben Media AG v Carlton UK Television Ltd [1999] 1 W.L.R. 605; [1999] F.S.R. 610 The same is no doubt true of the expressions ‘non-commercial research’ and ‘private study’.] All the fair dealing provisions use the words ‘for the purpose of’. As to whether this expression imports an objective or subjective test, it is important to construe the composite phrases ‘for the purpose of non-commercial research’ and ‘for the purpose of private study’, etc. rather than each single word. When this is done, the precise mental element on the part of the user ceases to be of great importance. The words ‘in the context of’ or ‘as part of an exercise in’ could be substituted for ‘for the purpose of’ without any significant change of meaning. The task of the court is to consider the use made of the work and then ascertain what was the perceived purpose of that use….

9-30 Form of acknowledgment. The acknowledgment of the work may take the form of identification of its title or some other description of it. [CDPA 1988, s.178 .] Identification of the author need not by reference to his proper name, particularly if this would be unlikely to have any particular significance or would be meaningless to the bulk of the intended audience. [Pro Sieben Media AG v Carlton UK Television Ltd [1999] 1 W.L.R. 605 ; [1999] F.S.R. 610; Newspaper Licensing Agency Ltd v Marks and Spencer plc [1999] R.P.C. 536 (CA). …] Use of a pseudonym by which an author is known to the public will therefore be sufficient.]….

9-38 For the purposes of criticism or review. The proper approach to the construction of these expressions has already been considered. […] In Australia, the words ‘criticism’ and ‘review’ have both been given their dictionary definitions. Thus ‘criticism’ has been defined as: (1) The act or art of analysing and judging the quality of a literary or artistic work, etc. (as in literary criticism). (2) The act of passing judgment as to the merits of something. (3) A critical comment, article or essay. [De Garis v Neville Jeffress Pidler Pty Ltd (1990) 18 I.P.R. 292 at 299.] ‘Review’ has been defined as: ‘a critical article or report, as in a periodical, on some literary work, commonly some work of recent appearance; a critique’. [De Garis v Neville Jeffress Pidler Pty Ltd (1990) 18 I.P.R. 292.] ‘Criticism’ can thus be seen as describing the ‘critical application of the mental faculties’, whilst ‘review’ describes the results of such a process. [De Garis v Neville Jeffress Pidler Pty Ltd (1990) 18 I.P.R. 292.] Criticism may be strongly expressed and unbalanced, the issue then usually being whether the use is fair dealing. [Pro Sieben Media AG v Carlton UK Television Ltd [1999] 1 W.L.R. 605, [1999] F.S.R. 610, CA. An author's remedy for unbalanced criticism which still amounts to fair dealing will lie, if anywhere, in the law of defamation: Pro Sieben.]

…

9-55 Relevant considerations. Relevant factors to be taken into account in judging whether the dealing was fair have been identified in various cases. None is determinative and the weight to be attached to them will vary from case to case. In particular, the various factors will carry different weight according to the type of dealing. Cases of fair dealing for purposes of criticism, review and the reporting of current events usually raise more difficult problems than cases of non-commercial research and private study. The three most important factors have been identified to be: [Ashdown v Telegraph Group Ltd [2001] EWCA Civ 1142; [2002] Ch.149…]

(1)

The degree to which the alleged infringing use competes with exploitation of the copyright work by the owner. This is likely to be a most important factor. …

(2)

Whether the work has been published or not….

(3)

The extent of the use and the importance of what has been taken. In many cases this will be a highly important factor…”

In Case C-5/08 Infopaq International A/S v Danske Dagblades Forening [2009] ECR I-0000 the Court of Justice had to consider for the first time the reproduction right under Article 2(a) of the Information Society Directive and the exception to that right contained in Article 5(1) of the same Directive in the context of an allegation that copyrights in literary works, namely newspaper articles, had been infringed by various acts including the reproduction of 11 word extracts.

With regard to the question of whether the reproduction right extended to the reproduction of 11 word extracts, the Court held:

“31.

It is clear that Directive 2001/29 does not define the concept of either ‘reproduction’ or ‘reproduction in part’.

32.

In those circumstances, those concepts must be defined having regard to the wording and context of Article 2 of Directive 2001/29, where the reference to them is to be found and in the light of both the overall objectives of that directive and international law (see, to that effect, SGAE, paragraphs 34 and 35 and case-law cited).

33.

Article 2(a) of Directive 2001/29 provides that authors have the exclusive right to authorise or prohibit reproduction, in whole or in part, of their works. It follows that protection of the author’s right to authorise or prohibit reproduction is intended to cover ‘work’.

34.

It is, moreover, apparent from the general scheme of the Berne Convention, in particular Article 2(5) and (8), that the protection of certain subject-matters as artistic or literary works presupposes that they are intellectual creations.

35.

Similarly, under Articles 1(3) of Directive 91/250, 3(1) of Directive 96/9 and 6 of Directive 2006/116, works such as computer programs, databases or photographs are protected by copyright only if they are original in the sense that they are their author’s own intellectual creation.

36.

In establishing a harmonised legal framework for copyright, Directive 2001/29 is based on the same principle, as evidenced by recitals 4, 9 to 11 and 20 in the preamble thereto.

37.

In those circumstances, copyright within the meaning of Article 2(a) of Directive 2001/29 is liable to apply only in relation to a subject-matter which is original in the sense that it is its author’s own intellectual creation.

38.

As regards the parts of a work, it should be borne in mind that there is nothing in Directive 2001/29 or any other relevant directive indicating that those parts are to be treated any differently from the work as a whole. It follows that they are protected by copyright since, as such, they share the originality of the whole work.

39.

In the light of the considerations referred to in paragraph 37 of this judgment, the various parts of a work thus enjoy protection under Article 2(a) of Directive 2001/29, provided that they contain elements which are the expression of the intellectual creation of the author of the work.

…

45.

Regarding the elements of such works covered by the protection, it should be observed that they consist of words which, considered in isolation, are not as such an intellectual creation of the author who employs them. It is only through the choice, sequence and combination of those words that the author may express his creativity in an original manner and achieve a result which is an intellectual creation.

46.

Words as such do not, therefore, constitute elements covered by the protection.

47.

That being so, given the requirement of a broad interpretation of the scope of the protection conferred by Article 2 of Directive 2001/29, the possibility may not be ruled out that certain isolated sentences, or even certain parts of sentences in the text in question, may be suitable for conveying to the reader the originality of a publication such as a newspaper article, by communicating to that reader an element which is, in itself, the expression of the intellectual creation of the author of that article. Such sentences or parts of sentences are, therefore, liable to come within the scope of the protection provided for in Article 2(a) of that directive.

48.

In the light of those considerations, the reproduction of an extract of a protected work which, like those at issue in the main proceedings, comprises 11 consecutive words thereof, is such as to constitute reproduction in part within the meaning of Article 2 of Directive 2001/29, if that extract contains an element of the work which, as such, expresses the author’s own intellectual creation; it is for the national court to make this determination.

49.

It must be remembered also that the data capture process used by Infopaq allows for the reproduction of multiple extracts of protected works. That process reproduces an extract of 11 words each time a search word appears in the relevant work and, moreover, often operates using a number of search words because some clients ask Infopaq to draw up summaries based on a number of criteria.

50.

In so doing, that process increases the likelihood that Infopaq will make reproductions in part within the meaning of Article 2(a) of Directive 2001/29 because the cumulative effect of those extracts may lead to the reconstitution of lengthy fragments which are liable to reflect the originality of the work in question, with the result that they contain a number of elements which are such as to express the intellectual creation of the author of that work.”

With regard to the exception contained in Article 5(1), the Court held:

“56.

For the interpretation of each of those conditions in turn, it should be borne in mind that, according to settled case-law, the provisions of a directive which derogate from a general principle established by that directive must be interpreted strictly (Case C-476/01 Kapper [2004] ECR I-5205, paragraph 72, and Case C-36/05 Commission v Spain [2006] ECR I-10313, paragraph 31).

57.

This holds true for the exemption provided for in Article 5(1) of Directive 2001/29, which is a derogation from the general principle established by that directive, namely the requirement of authorisation from the rightholder for any reproduction of a protected work.

58.

This is all the more so given that the exemption must be interpreted in the light of Article 5(5) of Directive 2001/29, under which that exemption is to be applied only in certain special cases which do not conflict with a normal exploitation of the work or other subject-matter and do not unreasonably prejudice the legitimate interests of the rightholder.”

SAS Institute’s claim that WPS infringes the copyrights in the SAS Components raises some fundamental issues of copyright law. These include three distinct, but related questions: (1) the extent to which copyright protects ideas, procedures, methods of operation and mathematical concepts as distinct from expressions of those ideas etc; (2) the extent to which copyright protects the functionality and interfaces of computer programs and the programming languages in which they are expressed; and (3) the test to be applied to determine what amounts to reproduction of a substantial part in cases such as this. The current domestic law on these matters is set out above, in particular in the speech of Lord Hoffmann in Designers’ Guild, in the judgments of the Court of Appeal concerning ordinary literary works in Baigent and in the judgments of Pumfrey J in Navitaire and Jacob LJ in Nova concerning computer programs.

As counsel for SAS Institute submitted, the starting point for consideration of these questions is that Article 10(1) of TRIPS, Article 4 of the WIPO Copyright Treaty and Article 1(1) of the Software Directive all provide that computer programs are to be protected as literary works within the Berne Convention. Thus, subject to specific exceptions such as those contained in Articles 5 and 6 of the Software Directive, computer programs are protected in the same manner and to the same extent as any other literary work.

It is nevertheless inescapable, for the reasons so eloquently explained by Pumfrey J in Navitaire, in particular at [112] and [125], that computer programs are different from other kinds of literary works and present peculiar problems for copyright law. As in that case, so too in this one, counsel for SAS Institute sought to rely upon doctrines established in relation to other kinds of literary work. While I agree that the general principles of copyright law must be applied to computer programs, it does not necessarily follow that particular doctrines which were established in relation to other types of literary works can be applied to computer programs, or applied in the same way.

Both Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty distinguish between “expressions”, which are protected by copyright, and “ideas, procedures, methods of operation and mathematical concepts as such”, which are not. Although Article 9(2) of TRIPS was referred to by Lord Hoffmann in Designers’ Guild and by Jacob LJ in Nova, none of the judgments in the cases mentioned in paragraph 196 above refer to Article 2 of the WIPO Copyright Treaty. When the present case was opened, I invited the parties to consider the relevance of Article 2 of the WIPO Copyright Treaty, as well as Article 9(2) of TRIPS. By the end of the trial there was a considerable measure of common ground on these questions.

Relevance of Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty to domestic law. It is clear that the United Kingdom courts must interpret both the Software Directive and domestic law in conformity with both the relevant provisions of TRIPS and the relevant provisions of the WIPO Copyright Treaty. This duty arises both as a matter of EU law and as matter of domestic law. It arises as a matter of EU law as a result of the principles of interpretation laid down by the ECJ which are set out in paragraphs 167 and 168 above. In addition, it is apparent from recitals (15) and (20) and Article 1(2)(a) of the Information Society Directive that the Community legislature considered that the Software Directive was consistent with the WIPO Copyright Treaty (by contrast recital (61) of the Information Society Directive says that Directives 92/100/EEC and 93/98/EEC are being amended “in order to comply with the WIPO Performances and Phonograms Treaty”). The duty arises as a matter of domestic law under section 2(1) of the European Communities Act 1972 by virtue of the fact that, as noted in paragraph 162 above, both TRIPS and the WIPO Copyright Treaty have been designated as EU Treaties.

Counsel for SAS Institute accepted that domestic law with regard to literary works other than computer programs should be construed in conformity with TRIPS and the WIPO Copyright Treaty for the second reason, but not the first. He submitted that the first reason did not apply because the EU had not legislated in relation to ordinary literary works. Although it does not matter given the concession in relation to the second reason, I do not accept this. The EU has legislated in relation to ordinary literary works, notably in the Information Society Directive. Whilst that is a measure of harmonisation and approximation of only some aspects of national copyright law as it affects literary works, it plainly brings ordinary literary works within the scope of the acquis communitaire. The ramifications of this can be seen from the ECJ’s decision in Infopaq.

Counsel for SAS Institute took no point on the fact that the EU and its Member States only ratified the WIPO Copyright Treaty this year, after the commission of many of the infringing acts alleged. In my view he was right not to do so, since both sources of the interpretative duty arose well before then.

Nor did counsel for SAS Institute dispute that the duty of interpretation is a strong one, in accordance with the Marleasing principle explained above.

The correct approach to interpretation of Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty. There is a broad consensus among scholars that both Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty are intended to be confirmatory of the position under Article 2 of the Berne Convention. Articles 9(2) and 10(1) of TRIPS, and Articles 2 and 4 of the WIPO Copyright Treaty, are historically linked. The inclusion of Article 10(1) in the drafts of TRIPS, requiring that computer programs be protected by copyright as literary works, led to concerns that such copyrights might be used to monopolise ideas, procedures, methods of operation or mathematical concepts. It was initially proposed that what became Article 9(2) be an exception to Article 10(1), but its application was widened to all forms of copyright protection. When the WIPO Copyright Treaty came to be drafted, it was desired to preserve the same balance between the provisions, and hence the replication of Articles 9(2) and 10(1) of TRIPS in Articles 2 and 4 of the WIPO Copyright Treaty. If Article 2 of the WIPO Copyright Treaty is compared to Article 9(2) TRIPS, it can be seen that the wording is identical, save that “shall extend to” has been replaced by “extends to.” This subtle change of words is deliberate, indicating that Article 2 of the WIPO Copyright Treaty is intended to be confirmatory of an existing recognised principle, consistently with the WIPO Copyright Treaty’s status as a special agreement under the Berne Convention. The linkage between Articles 2 and 4 of the WIPO Copyright Treaty, and the intention that they should re-state the effects of Articles 9(2) and 10(1) of TRIPS, is made explicit in the Agreed Statement concerning Article 4 of the WIPO Copyright Treaty. The Agreed Statement also makes it clear that Article 2 of the WIPO Copyright Treaty is intended to be consistent with Article 2 of the Berne Convention. The same point is implicit in the relationship between Article 9(1) and 9(2) in TRIPS. On these points see: Ficsor, The Law of Copyright and the Internet: The 1996 WIPO Copyright Treaties, Their Interpretation and Implementation (OUP, 2002), pp. 457-460; Gervais, The TRIPS Agreement: Drafting History and Analysis (3rd ed, Sweet & Maxwell, 2008), pp. 220-223; Reinbothe and von Lewinski, The WIPO Treaties 1996: Commentary and Legal Analysis (Butterworths, 2002) pp. 27, 45-47; Ricketson and Ginsburg, International Copyright and Neighbouring Rights: The Berne Convention and Beyond (2^nd ed, OUP, 2007), pp. 150-151, 514-519.

Effect on domestic law. Counsel for SAS Institute submitted that it followed from the points discussed above that Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty did not change the law. I do not agree with this. In Ibcos (supra) Jacob J (as he then was) pointed out at 292 that the United Kingdom Copyright Act differed from the United States Copyright Act in that it contained no counterpart to section 102(b) of Title 17 of the United States Code, which provides:

“In no case does copyright protection for an original work of authorship extend to any idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied in such work.”

At that time there was nothing to require the United Kingdom courts to distinguish between “expressions” and “ideas, procedures, methods of operation and mathematical concepts as such”, since, even if they had recognised that Article 2 of the Berne Convention should be interpreted in that manner, the principle of statutory construction articulated by Lord Diplock in The Jade did not require them to interpret the 1988 Act in the same way. For the reasons given above, the 1988 Act must now be interpreted in conformity with Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty. As a result, the United Kingdom courts must now interpret the 1988 Act so as to protect “expressions” and not “ideas, procedures, methods of operation and mathematical concepts as such”.

The distinction. For the reasons given above, it is necessary to distinguish between “expressions” on the one hand and “ideas, procedures, methods of operation and mathematical concepts as such” on the other. What is protected by copyright in a literary work is the form of expression of the literary work itself. Other things which are conveyed by or described in the literary work, of which “ideas, procedures, methods of operation and mathematical concepts” is evidently a non-exhaustive list, are not protected. Thus these provisions draw a line between copyright protection and the public domain: see Reinbothe and von Lewinski, pp. 46-47.

Counsel for SAS Institute emphasised the words “as such”, and submitted that the effect of the exclusion was a question of the level of generality at which an idea etc. was expressed: at a high level of generality the idea etc. could not be protected, whereas the detailed expression of the idea etc. at a lower level of generality was protected. Counsel for WPL submitted that the distinction between expression and ideas etc. was not, or not merely, a matter of level of generality. In my judgment counsel for WPL is correct: the distinction is one between different kinds of skill, judgement and labour. Skill, judgement and labour in devising ideas, procedures, methods of operation and mathematical concepts is not protected by the copyright in a literary work. What is protected by copyright in a literary work is the skill, judgement and labour in devising the form of expression of the literary work.

Article 1(2) of the Software Directive. Recital [15] of the Software Directive suggests that Article 1(2) of the Software Directive is intended to be interpreted in accordance with Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty, particularly now that the original version has been replaced by the codified version (in which the relevant wording is now to be found in recital (11)) which post-dates those treaties. In any event, Article 1(2) of the Software Directive must be interpreted in conformity with Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty for the reasons explained above. In my judgment there is no conflict between them. Although Article 1(2) of the Software Directive is differently worded, and in particular refers to “ideas and principles” rather than “ideas, procedures, methods of operation and mathematical concepts”, the distinction it draws is essentially the same as that drawn by Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty. If and in so far as there is any difference, however, Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty make it clear that Article 1(2) of the Software Directive must be broadly interpreted.

It can be seen from the paragraphs 178-185 above that in Navitaire Pumfrey J interpreted Article 1(2) of the Software Directive as meaning that copyright in computer programs did not protect (i) programming languages, (ii) interfaces or (iii) the functionality of a computer program. It can be seen from paragraph 190 above that his conclusion on point (iii) was upheld by the Court of Appeal in Nova.

Counsel for SAS Institute submitted that Pumfrey J was wrong on each of these questions, although he also sought to distinguish the present case from Navitaire and Nova on the facts. I shall consider these questions in turn. Before doing so, however, it is convenient to mention three general points on which counsel for SAS Institute relied. The first is the point, which I have already commented on, that the Software Directive requires computer programs to be protected as literary works. Secondly, he relied on the fact that recital [2] of the Software Directive recognises that the development of computer programs requires substantial investment which it is the purpose of copyright to protect. Thirdly, he argued that the Software Directive treats the dichotomy between ideas and expression in accordance with the same general principles which apply to other types of literary work. I accept each of these points, but I do not consider that either individually or cumulatively they support the contention that programming languages, interfaces and functionality are protected by copyright in computer programs.

Programming languages. Counsel for SAS Institute pointed out that recital [14] of the Software Directive says that “to the extent that … programming languages comprise ideas and principles, those ideas and principles are not protected under this Directive [emphasis added]”. He submitted that this was not the same thing as saying that programming languages themselves were not protected, and that there was no reason why the expression of ideas and principles in the form of a programming language should not be protected. Furthermore, he argued that the legislative history of the Software Directive pointed away from the conclusion that programming languages were not protected.

So far as the legislative history is concerned, counsel for WPL relied on passages from the European Commission Green Paper On Copyright and the Challenge of Technology – Copyright Issues Requiring Immediate Action dated 7 June 1988 (COM(88) 172 final). I am doubtful whether this is a legitimate aid to interpretation, however, given that it preceded the proposal for the Software Directive. In any event, I think it is of less assistance than later documents.

In the Commission’s original Proposal of 5 January 1989 (Com (88) 816/final), the draft text of the proposed Directive included the following provision as Article 1(3):

“Protection in accordance with this Directive shall apply to the expression in any form of a computer program but shall not extend to the ideas, principles, logic, algorithms or programming languages underlying the program. Where the specification of interfaces constitutes ideas and principles which underlie the program, those ideas and principles are not copyrightable subject matter.”

On 11 July 1990 the European Parliament approved the proposal subject to amendments, including the following amendment to Article 1(3):

“Protection in accordance with this Directive shall apply to the expression in any form of a computer program. Ideas and principles which underlie any aspect of a program, including its interfaces, shall not be protected by copyright under this Directive.”

The Commission’s introduction to its Amended Proposal of 18 October 1990 (Com (90) 509) stated that the Amended Proposal contained three major modifications to the original Proposal, the first of which was described as follows:

“As regard the scope of protection given by the application of copyright to computer programs, the original proposal has been abridged and simplified as proposed by the European Parliament to make clear the basic doctrine of copyright law which the Directive seeks to apply. That doctrine, widely applied in the jurisprudence of the Member States even if not always articulated explicitly in each national legislation, is that copyright protection only applies to the expression of an idea or principle, and not to the idea or principle itself.”

In the Amended Proposal, the European Parliament’s text was adopted as Article 1(2) subject to the replacement of the word “aspect” by “element” and recitals [13], [14] and [15] were introduced for the first time. The Commission’s commentary on Article 1(2) was as follows:

“The Commission has accepted the amendment of the Parliament to the text of article 1.3 of the original proposal which re-states, in a simplified form, the general principle of copyright law on which the scope of protection of this proposal is based. That principle is the following: copyright protects the way in which an idea is expressed, but does not give a monopoly in the idea itself. A second author may take up an idea which he has found in an existing work and re-formulate it, using a different expression, and acquiring his own copyright in that new expression. The text proposed by the Parliament makes clear that this copyright principle is to be applied to every part of a program: it is therefore a formulation capable of being interpreted by national courts across a broad range of situations… ”

I am not persuaded that Pumfrey J was wrong to conclude that programming languages are not protected. While I acknowledge that recital [14] can be read in the manner contended for by counsel for SAS Institute, it should not be construed as if it were an operative provision in an English statute. It is there to guide courts as to the purpose of Article 1(2). It must be read in its context between recitals [13] and [15] (a point which is reinforced in the codified version by the fact that all three have been combined in one recital, recital (11)). Read in that context, the words “to the extent that” can be understood as meaning “in as much as”. As for the legislative history, I do not agree that this demonstrates, as counsel for SAS Institute argued, that an exclusion of programming languages was deliberately not included in the Software Directive. To the contrary, I consider that it indicates that Article 1(2) is to be broadly interpreted. Furthermore, I think that the distinction which Pumfrey J drew between a computer program and the language it is written in is, despite his hesitancy on the point, perfectly consistent with the distinction between expressions and ideas, procedures, methods of operation and mathematical formulae.

I also agree with Pumfrey J, however, that the correct interpretation of Article 1(2) of the Software Directive on this point is not acte clair and that a reference to the ECJ is required in order to determine it. Furthermore, as the evidence in this case demonstrates, even if Pumfrey J is right that programming languages are not protectable, there is room for debate as to how broadly the concept of a programming language should be interpreted in this context.

Interfaces. The specific issue on interfaces which arose in Navitaire, which concerned the “user interface”, and in particular the VT100 screen displays and reports, does not arise in the present case. In this case there is no real counterpart to the VT100 screen displays and reports. The nearest equivalent is the log file, but counsel for WPL did not argue that this was an interface. Counsel for WPL did contend, however, that the syntax and data file formats of the SAS Language were both interfaces. I will consider whether they are interfaces below. At this stage, I will confine my analysis to the legal question as to whether interfaces are protected by the copyright in a computer program.

Counsel for SAS Institute pointed out that recital [13] and Article 1(2) of the Software Directive exclude protection for “ideas and principles which underlie any aspect of a computer program, including those which underlie its interfaces [emphases added]”. He submitted that this was not the same thing as saying that interfaces were not protected, and that there was no reason why the expression of ideas and principles in the form of an interface should not in general be protected.

Furthermore, he pointed out that recital [11] of the Software Directive defined “interfaces” as “the parts of the program which provide for such interconnection and interaction between elements of software and hardware” and that recital [12] referred to such “interconnection and interaction” as “interoperability”. He submitted that recitals [21]-[24] and Article 6 provided for a specific, narrow exception for interoperability of computer programs, and that that was inconsistent with Article 1(2) being interpreted as a broad exclusion for interfaces.

Counsel for WPL submitted that the legislative history showed that interfaces were not protected. He relied in particular on the following passage in Part One of the original Proposal:

“3.11.

In order to produce interoperative systems it is necessary to replicate the ideas, rules or principles by which interfaces between systems are specified, but not necessarily to reproduce the code which implements them. Ideas, rules or principles are not copyrightable subject matter. Such ideas, rules or principles may be used by any programmer in the creation of an independent implementation of them in an interoperative program.

3.12

Competitors are therefore free, once they establish through independent analysis which ideas, rules or principles are being used, to create their own implementation of the ideas, rules or principles in order to make compatible products. They may build on the identical idea, but may not use the same expression as that of other protected programs. There is thus no monopoly on the information itself, but only a protection of the form of expression of that information.

3.13

If similarities in the code which implements the ideas, rules or principles occur as between interoperative programs, due to the inevitability of certain forms of expression, where the constraints of the interface are such that in the circumstances no different implementation is possible, then no copyright infringement will normally occur, because in these circumstances it is generally said that idea and expression have merged.

3.14

Although it is technically possible to decompile a program in order to find out information concerning access protocols and interfaces this is a lengthy, costly and inefficient procedure. It is usually more efficient for the parties concerned to agree on the terms under which the information will made available. Problems of access to information may have to be addressed by other means which are outside the scope of this Directive.”

As counsel for WPL pointed out, at that stage the proposed Directive did not contain a provision corresponding to Article 6 of the Directive as enacted. Thus these comments related to what was then Article 1(3) of the draft Directive. Counsel for WPL argued that, for the reasons explored above in relation in programming languages, it had not been intended to narrow the exclusion of interfaces from protection in the final Directive.

As for Article 6, this had its origin in Article 5a of the European Parliament’s amended text. This was substantially accepted by the Commission in its Amended Proposal. In the introduction to the Amended Proposal, the third major modification to the original Proposal was described as follows:

“In response to concerns expressed by the European Parliament and by part of the industry, a further exception to the author’s exclusive rights for the purpose of creating an interoperable program has been accepted.

Computer programs have to interoperate with hardware and other software in order to perform their functions and in order to form systems and networks. If a manufacturer wishes to interconnect his products with others supplied by a different manufacturer he may need information from the manufacturer about how his products are designed to interconnect.

Such information may be at the present time usually available through materials supplied by manufacturers or by the growing move towards the use of publicly available open standards where the means to interconnect have been standardised and described and documented by international standards bodies.

However if information is not forthcoming or if the design of the means of interconnection is a non-standard proprietary one, manufacturers could find themselves unable to derive sufficient detailed information without committing acts which technically violate the author's exclusive rights to prevent the reproduction and translation of his program. These acts of reproducing and translating the object code version of the program, which is the version normally supplied the public, back into a language representing something more like the original source code in which the programmer devised a programme are often referred to loosely as ‘reverse engineering’ the program.

Although a dominant supplier who refused to make information available to provide for interoperability between programs or between programs and hardware could be subject to the application of the competition rules under Articles 85 and 86 of the EEC treaty, the Commission has been persuaded that the original proposal, which left the matter of ‘reverse engineering’ not explicitly regulated, lacked sufficient clarity. It is therefore proposed that an additional Article 5bis dealing with a derogation allowing ‘reverse engineering’ of programs for the purposes of interoperability of the programme should be added. Nothing in this Directive should prevent however the ‘reverse engineering’ of a program, whether incorporated into hardware or not, under the conditions of Article 5bis for the purpose of independently creating an interoperable program, wherever it may be incorporated.”

Similarly, the commentary to Article 5bis stated:

“The problem which this article addresses stems from the nature of the computer program highlighted in the context of Article 5 paragraph 5 above, that is to say, the fact that it cannot be easily ‘read’ by a human user.

However, a computer program may be required to interconnect and interact with other computer programs, for example, an applications program with an operating system. In order for the creators of computer programs to understand how their creations can interconnect and interact with those of others, they must be able to perceive in detail how the first manufacturer has provided for the exchange of data between his program and other programs.

In many instances the creator of the original program will have made available through published manuals or on request, sufficient information about the parts of his program whose function is to provide for its interoperation with other programs. In other instances, the design of those parts of the program will be standardized and publicly documented for all creators of programs to work to a common agreed interface specification. In a certain number of cases it may be that information is not forthcoming by either of these means. In such cases the creator of the original program, by withholding information from competitors, can ensure that only he can supply the range of other programs which will interoperate with his original program.

Article 5 paragraph 5 is intended to permit much of the information required for the purpose of interoperability to be derived from observation, study or testing of the program without committing infringements of the author’s exclusive rights. These techniques will be adequate in many circumstances. However, where their use does not produce sufficient information and where other non-infringing means such as the use of publicly available material or published documentation is also inadequate, the Commission’s amended proposal ensures, that as a last resort, a person having a right to use a copy of the program may commit acts of reproduction and translation of the machine-readable form of the code in which the copy has been supplied without the authorization of the right holder, subject to certain limitations.

In this way, the amended proposal provides a safety mechanism by which an independently created program can be made to be interoperable with an existing program, even when the creator of the existing program has chosen not to reveal to third parties the specification of the interfaces whose function is to provide a means of interconnection with other elements of a computer system.”

Again, I am not persuaded that Pumfrey J was wrong to conclude that interfaces as described in recital [15] of the Software Directive are not protected by the copyright in a computer program. In my judgment the legislative history supports this interpretation. The inclusion of Article 6 in the Software Directive does not support the opposite interpretation. It can be seen from the passages quoted above that the purpose of Article 6 is to entitle third parties to obtain information about interfaces by decompiling the object code of a program where the necessary information is not available from either (i) published sources such as manuals, (ii) common standards or (iii) observation, study or testing of the program. Where information about an interface can be obtained in one or more of those ways, it is evident that the Commission considered that competitors would be free to copy the interface anyway.

Again, however, I do not consider that the point can be said to be acte clair and a reference to the ECJ is required to determine it. This is not a matter which Pumfrey J addressed, perhaps because he felt that it was unnecessary to repeat what he had said in relation to programming languages.

Functionality. Counsel for SAS Institute submitted that Pumfrey J was wrong to conclude that it was not an infringement of the copyright in a computer program to reproduce the functionality of the program. He contended that the correct view of the law was that expressed in the passage from Laddie, Prescott & Vitoria, The Modern Law of Copyright (3^rd ed, Butterworths, 2000) at §36.64 quoted by Pumfrey J at [126]. In support of this contention he relied, in addition to the general arguments I have already considered, on the legislative history of the Software Directive, and in particular on the inclusion of “preparatory design material” in Article 1(1) during the course of the adoption of the Directive.

The Commission’s original Proposal described the nature of the intellectual property which it was intended to protect in Part One as follows:

“2.2

As far as the property right is concerned, a computer program, in common with other works protected by intellectual property legislation, is the result of a creative intellectual human activity. While its mode of expression or fixation may still be unfamiliar to many, the degree of creativity, skill and inventiveness required to devise a program make it no less deserving of protection than other works protected by copyright. The fact that computer programs have utilitarian function does not change this.

2.3

These elements of creativity, skill and inventiveness manifest themselves in the way in which the program is elaborated. The tasks to be performed by a computer program need to be defined and an analysis of the possible ways to achieve these results must be carried out. A selection has been made of the various solutions and the steps to achieve the end result must be listed. The way in which these steps are expressed gives the program its particular characteristics of speed, efficiency and even style. A program has a structure, with sections and subsections, through which information flows. In common with other literary works, the computer program also has an underlying logic in the presentation of the various steps.

2.4

These steps, the algorithms, from which the program is built up, should not be protected as such against unauthorized reproduction. They are the equivalent of the words by which the poet or the novelist creates a work of literature, or the brushstrokes of the artist or the musical scales of the composer.

2.5

As with literary works in general, protection can only be envisaged for a computer program from the point at which the selection and compilation of these elements indicate the creativity and skill of the author, and set his work apart from that of other authors.

…

2.7

In practice, computer programs are rarely of such simplicity that authors will arrive at totally identical programs, independently of each other. On the other hand many sub-routines which programmers habitually use in order to build up programs are in themselves commonplace in the industry and the originality of the program may lie in the selection and compilation of these otherwise commonplace programs.

2.8

The success of the program in terms of its ability to perform the task which is required will to a large extent be conditioned by these choices made by the author of the program at every step along the way. This success will manifest itself in a program which is quicker, easier, more reliable, more comprehensive, more productive to use than its predecessors or its competitors. ”

Similarly, in the commentary on Article 1 of the proposed Directive in Part Two, the original Proposal stated:

“1.3

Copyright protects the expression of ideas but not the ideas themselves. Therefore the protection given to computer programs will extend to the program as a whole, and to its constituent parts, in so far as they represent a sufficient degree of creativity to qualify as ‘works’ in themselves. The only criterion which should be applied to determine eligibility for protection is that of originality, that is, that the work has not been copied. No other aesthetic or qualitative test should be applied. Sub-routines and routines which go together to form modules which in turn form programs may all qualify for protection independently of the protection given to the program as a whole, that is, as a compilation of such elements. The algorithms which go to make up the sub-routines are not normally in themselves capable of receiving protection under copyright insofar as they are similar in nature to mathematical formulae. They may in exceptional circumstances attract patent protection. Similarly, the ideas, principles, or logic which underlie the program will not be copyrightable.

1.4a Many algorithms and many sub-routines are commonplace in the industry. They may have been placed or have fallen into the public domain or they may be de facto standard routines or algorithms. Where a program is composed wholly or in part of such commonplace or unprotected algorithms and routines, it should nevertheless be protected as a compilation, provided that it is original in the above mentioned sense and that the creator demonstrated skill and labour in the creation of the compilation.”

At that stage, Article 1 of the draft Directive did not contain any reference to “preparatory design material”. The inclusion of “preparatory design material” was first proposed by the European Parliament in Article 1(1) of its amended text. The Commission essentially accepted this proposal. The commentary to Article 1(1) in the Commission’s Amended Proposal states:

“… The proposal also takes up the clarification proposed by Parliament that preparatory design work leading to the development of a computer program is protected as a computer program provided the nature of the preparatory work is such that a computer program can result from it at a later stage.”

This commentary formed the basis for recital [7] of the Software Directive.

Even leaving aside the decision of the Court of Appeal in Nova, I am not persuaded that Pumfrey J was wrong to hold that it is not without more an infringement of the copyright in a computer program to create another computer program which has the same functionality. I accept that copyright protection is not limited to the text of the source code of the program, but extends to protecting the design of the program, that is, what has been referred to in some cases as its “structure, sequence and organisation”. If there were any doubt about this, then the conferring of protection on “preparatory design material” confirms it. But there is a distinction between protecting the design of the program and protecting its functionality. It is perfectly possible to create a computer program which replicates the functionality of an existing program, yet whose design is quite different.

In my judgment Pumfrey J was right to say that at [129] the key question is “the nature of the skill and labour”. If one takes the example of the tax calculation program postulated in The Modern Law of Copyright, the skill, judgement and labour that goes into understanding and elucidating the tax regulations is the wrong kind of skill, judgement and labour to be protected by copyright in the resulting computer program (including any preparatory design material). Copyright in the computer program (including any preparatory design material) protects the skill, judgement and labour in devising the form of expression of the program (including any preparatory design material), that is to say, its design and source code.

Counsel for SAS Institute relied before me, as he had before Pumfrey J in Navitaire, on the line of cases which establishes that it may be an infringement of the copyright in a novel or play to reproduce the plot even though no single sentence is reproduced. I agree with Pumfrey J, however, that the functions of a computer program are not analogous to the plot of a novel or play. The reason why the plot of a novel or play is protected in an appropriate case is that the plot forms part of the expression of the literary work. The correct analogy in the case of computer programs is with the design of the computer program. The functions of a computer program have no counterpart in the case of novel or play because a novel or play has no function in that sense.

Counsel for SAS Institute also relied on the fact that copyright may be relied on to protect literary works which are wholly utilitarian in nature. Thus in Anderson v Lieber Code Co [1917] 2 KB 469 it was held that copyright subsisted in a telegraphic code of invented five letter words. But what was protected in that case was the skill, judgment and labour in devising the form of expression of the code.

Accordingly, I consider that the functionality of a computer program falls on the wrong side of the line drawn by Article 1(2) of the Software Directive, Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty.

In any event, Pumfrey J’s judgment on this point was upheld by the Court of Appeal in Nova, and that decision is binding on me unless and until overruled by either the Supreme Court or the Court of Justice.

On the other hand, I do not interpret the Court of Appeal in Nova as having held that the matter was acte clair. What Jacob LJ said at [53] was that he did not think that it was “necessary to make any reference to resolve this case [emphasis added]”. In my view the question is not acte clair, and a reference to the ECJ is required in order to determine it.

Counsel for SAS Institute submitted that it was wrong to dissect a copyright work into protectable and non-protectable parts when considering whether there had been a reproduction of a substantial part of it. In support of this submission he relied on the decisions of the House of Lords in Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 273 and Designers’ Guild.

Although counsel referred to all of the speeches in Ladbroke, it is sufficient to cite what Lord Reid said at 276-277:

“The first question to be determined is whether or to what extent copyright attaches to these coupons. The respondents say that a coupon must be regarded as a single work and that as such it is protected by copyright. The appellants seek to dissect the coupon. They would not only dissect it into the 16 lists, but they would further dissect each list into heading, selection of matches, and statement of odds offered for the various kinds of bets. They admit that there is copyright in the selection and in the statements of odds offered: they can safely do that because there they did not copy. But they deny any copyright as regards the rest of the coupon.

The Copyright Act, 1956, provides, by section 2, that copyright shall subsist in every original literary work and, by section 48, that literary work includes any written table or compilation. I have no doubt that the coupon must be treated as a single compilation. The appellants' dissection theory is derived from some statements in infringement cases and I must, therefore, examine at this point the law regarding infringement. Copyright gives the exclusive right to do certain things including ‘reproducing the work in any material form’ (section 2 (5)), and reproduction includes reproduction of a substantial part of the work (section 49 (1)). Broadly, reproduction means copying, and does not include cases where an author or compiler produces a substantially similar result by independent work without copying. And, if he does copy, the question whether he has copied a substantial part depends much more on the quality than on the quantity of what he has taken. One test may be whether the part which he has taken is novel or striking, or is merely a commonplace arrangement of ordinary words or well-known data. So it may sometimes be a convenient short cut to ask whether the part taken could by itself be the subject of copyright. But, in my view, that is only a short cut, and the more correct approach is first to determine whether the plaintiffs’ work as a whole is ‘original’ and protected by copyright, and then to inquire whether the part taken by the defendant is substantial.

A wrong result can easily be reached if one begins by dissecting the plaintiffs’ work and asking, could section A be the subject of copyright if it stood by itself, could section B be protected if it stood by itself, and so on. To my mind, it does not follow that, because the fragments taken separately would not be copyright, therefore the whole cannot be. Indeed, it has often been recognised that if sufficient skill and judgment have been exercised in devising the arrangements of the whole work, that can be an important or even decisive element in deciding whether the work as a whole is protected by copyright.”

In Designers’ Guild Lord Hoffmann criticised the Court of Appeal for having engaged in dissection of the respective designs, saying at [19]:

“The exercise in dissection also, as it seems to me, involved two errors. First, it ignored substantial parts of the judge’s findings on what had been copied and, secondly, it dealt with the copied features piece-meal instead of considering, as the judge had done, their cumulative effect.”

In my judgment neither of these passages bears the weight that counsel for SAS Institute placed on them. Lord Reid’s speech was primarily directed to the fallacy of considering the constituent parts of a work when deciding whether the whole work is original, particularly in a case where the work is a compilation. Lord Hoffmann’s speech was directed to the fallacy of considering the parts reproduced in isolation and not as a whole. It is trite law that, in determining whether the defendant has reproduced a substantial part of the copyright, the court must focus on the quality of the parts that have been reproduced. As Lord Hoffmann (with whom the other members of the House of Lords agreed) explained in Newspaper Licensing Agency Ltd v Marks & Spencer plc [2001] UKHL 38, [2003] 1 AC 551 at [19] (emphasis added):

“The House of Lords decided in Ladbroke (Football) Ltd v William Hill (Football) Ltd [1964] 1 WLR 273 that the question of substantiality is a matter of quality rather than quantity. The relevant passages are too well known to require citation: see Lord Reid, at p 276, Lord Evershed, at p 283, Lord Hodson, at p 288 and Lord Pearce, at p 293. But what quality is one looking for? That question, as it seems to me, must be answered by reference to the reason why the work is given copyright protection. In literary copyright, for example, copyright is conferred (irrespective of literary merit) upon an original literary work. It follows that the quality relevant for the purposes of substantiality is the literary originality of that which has been copied. In the case of an artistic work, it is the artistic originality of that which has been copied. So, in the recent case of Designers Guild Ltd v Russell Williams (Textiles) Ltd (trading as Washington DC) [2000] 1 WLR 2416, the House decided that although not the smallest part of a fabric design had been reproduced with anything approaching photographic fidelity, the copying of certain of the ideas expressed in that design which, in their conjoined expression, had involved original artistic skill and labour, constituted the copying of a substantial part of the artistic work.”

This passage was strictly obiter, but it plainly represents the considered view of the House of Lords. In any event, the Court of Justice has now adopted the same approach in Infopaq in holding at [48] that an extract of 11 words from a literary work constituted a substantial part of the work if the extract contained an element of the work which expressed the author’s own intellectual creation. It follows from this that, when considering whether a substantial part has been reproduced, it is necessary to focus upon what has been reproduced and to consider whether it expresses the author’s own intellectual creation. To that extent, some dissection is not merely permissible, but required. On the other hand, the Court of Justice also held in Infopaq at [49] that it is necessary to consider the cumulative effect of what has been reproduced.

Counsel for SAS Institute also argued that, even though none of the copyright works relied on by SAS Institute were compilations, they contained compilations, such as compilations of keywords and compilations of mathematical formulae, and that it was appropriate to approach the question of reproduction of a substantial part on this basis. So far as this argument is concerned, it is important to be clear as to the basis upon which a compilation of unprotectable subject-matter can be protected as a copyright work. As can be seen from Article 2(5) of the Berne Convention, Article 10(2) of TRIPS and Article 5 of the WIPO Copyright Treaty, compilations are protected (whether or not the items compiled are protected) which “by reason of the selection or arrangement of their contents constitute intellectual creations”. Similarly, EU legislation adopts the test of “intellectual creation” for the originality of computer programs (Article 1(3) of the Software Directive), databases for the purposes of copyright (Article 3(1) of European Parliament and Council Directive 96/9/EC of 11 March 1996 on the legal protection of databases) and photographs (Article 6 of Council Directive 93/98/EEC of 29 October 1993 harmonizing the terms of protection of copyright and certain related rights, now codified as Directive 2006/116/EC). It is now clear from Infopaq at [31]-[48] that there will only be reproduction of a substantial part of a literary work, including a compilation, where what has been reproduced represents the expression of the intellectual creation of the author of that literary work.

Assuming that the law is as stated above, I consider that the application of these principles to the present case leads to the following conclusions.

First, as recorded above, it is common ground that WPS replicates a large part of the functionality of the SAS Components (categories (1), (2), (4), (5), (6) and (7) of the similarities relied on by SAS Institute). For the reasons given above, I conclude that this in itself does not constitute an infringement of the copyrights in the SAS Components. WPL has not reproduced any of the design of the SAS Components, let alone the actual source code.

Secondly, I have already found that the SAS Language is a programming language. It is common ground that WPS reproduces elements of the SAS Language. In particular, the WPS parser reproduces the SAS keywords, including keywords for functions that are not supported (category (2) of the similarities relied on by SAS Institute). For the reasons given above, I conclude that this in itself does not constitute an infringement of the copyrights in the SAS Components.

Thirdly, counsel for WPL argued that the syntax of the SAS Language and the SAS data file formats were both interfaces. In my view the syntax of the SAS Language is better regarded as part of the programming language than as an interface of the kind referred to in recital [15] of the Software Directive, but the conclusion that it is unprotected by the copyrights in the SAS Components is the same. As for the SAS data file formats, I agree that these are interfaces. These are precisely the kind of information which is required by third parties in order to access data stored in those formats for the purposes of interoperability. It is common ground that WPS is able to read and write files in SAS7BDAT format (category (8) of the similarities relied on by SAS Institute). For the reasons given above, I conclude that this in itself does not constitute an infringement of the copyrights in the SAS Components. In any event, in my judgment it is not an infringement on the quite separate ground that there is no evidence that in this respect WPS reproduces a substantial, or indeed any, part of the SAS source code.

Fourthly, categories (3) and (5) of the similarities relied on by SAS Institute establish that in certain respects WPL has reproduced elements of the SAS Components which go beyond what is strictly necessary in order to reproduce the functionality of the SAS Components. In my judgment this does not amount to an infringement for two alternative reasons. The first is that it remains the case that what WPL has done is to replicate the functionality of the SAS Components. As I read Pumfrey J’s judgment in Navitaire, the principle he established that functionality is not protected by copyright is not confined to that which it is strictly necessary to reproduce. The second is that, even if the principle is so confined, in the present case I do not consider that WPL has reproduced a substantial part of the SAS source code once the replication of the functionality of the SAS Components is discounted. In the case of the names of procedures etc, these are very descriptive and there is no evidence that significant skill, judgement or labour was expended by SAS Institute is devising these. Likewise, the fixed text elements of the SAS log file are short and banal pieces of text. Again, there is no evidence that significant skill, judgement or labour was expended by SAS Institute in devising these. Furthermore, the fixed text elements in the WPS log file are not identical.

Finally before leaving this subject, I wish to comment on three forensic points. First, counsel for SAS Institute emphasised that WPS was what he described as a “drop-in replacement” for the SAS Components. As I have pointed out in paragraph 175 above, the same was true of eRes in Navitaire. Secondly and more importantly, counsel for SAS Institute described the SAS Manuals as providing “a window into” the source code of the SAS Components. Counsel for WPL criticised this metaphor as inaccurate and misleading. I agree. The SAS Manuals are nothing more (and nothing less) than a very detailed description of what the software does. They give the reader very little insight into how the source code achieves that. As noted above, Professor Ivey did not compare WPS to the SAS source code. When asked, he said that he would expect the two source codes to look different. Dr Worden, who had compared certain aspects of the WPS code with the corresponding SAS code, said that they were indeed very different. Thirdly, it is remarkable that a claim for infringement of copyright can be advanced in which, not only does the claimant not suggest that the defendant ever saw the copyright work, but also the claimant does not consider it necessary for either its own expert witness or the court to see the copyright work. In this respect SAS Institute’s claim goes beyond even that presented by Navitaire: see the judgment of Pumfrey J at [29].

Counsel for SAS Institute argued strenuously that, even if Navitaire and Nova were correctly decided in all material respects, there was a fundamental factual distinction between the present case and those cases. This is that, as is common ground, WPL made extensive use of the SAS Manuals when creating WPS. In short, he submitted that (i) the SAS Manuals are ordinary literary works to which the Software Directive does not apply, and (ii) WPL has reproduced substantial parts of the SAS Manuals in the source code of WPS.

The first point to observe about this argument is that I do not accept that there is quite as great a factual distinction between the present case and Navitaire as counsel for SAS Institute suggested. Pumfrey J found as a fact in Navitaire at [177] that BulletProof had used easyJet’s OpenRes manual, together with screen shots and reports from easyJet’s OpenRes system, in order to ascertain the requirements for eRes. On the other hand, it is fair to say that Navitaire did not advance a case of substantial reproduction of, and hence infringement of copyright in, the OpenRes manual by eRes of the kind advanced by SAS Institute here.

I also note that it is evident from the passage from the Commission’s Amended Proposal for the Software Directive quoted in paragraph 225 above that the Commission considered that it was permissible for a third party to use manuals published by the creator of the original program as a source of information about interfaces in order to create an interoperable program and, to that extent, to reproduce the original program.

I accept that the SAS Manuals are ordinary literary works which are not governed by the Software Directive. Nevertheless, the fact remains that they are detailed descriptions of the functionality of the SAS System. Furthermore, as counsel for SAS Institute himself argued, Article 1(2) of the Software Directive draws the same distinction between expressions and ideas etc. as Article 9(2) of TRIPS and Article 2 of the WIPO Copyright Treaty. It is common ground that the latter two provisions do apply to the SAS Manuals. As discussed above, the reproduction right under Article 2(a) of the Information Society Directive and under sections 16(1)(a) and 17 of the 1988 Act must be interpreted as far as possible in conformity with those provisions.

Assuming that the reasoning which led Pumfrey J in Navitaire to conclude that it is not an infringement of the copyright in a computer program to replicate its functions without copying its source code or design is correct, I consider that by parity of reasoning it is not an infringement of the copyright in a manual describing those functions to use the manual as a specification of the functions that are to be replicated and, to that extent, to reproduce the manual in the source code of the new program. Once again, it is a question of the kind of skill, judgement and labour involved. Copyright in a literary work, be it a computer program or a manual, does not protect skill, judgment and labour in creating ideas, procedures, methods of operation or mathematical formulae. It follows that it is not an infringement to reproduce such things either from a computer program or a manual.

Once again, however, I do not consider that this interpretation of Article 2(a) of the Information Society Directive is acte clair. It is another question on which a reference to the ECJ is required.

Counsel for SAS Institute advanced three main arguments against this conclusion in addition to attacking its premise as discussed above. First, he submitted that WPL had reproduced the SAS Manuals in considerable detail, rather than merely at a general level. Secondly, he submitted that WPL had reproduced the form of expression of the manuals, and not merely the ideas etc. expressed in them. Thirdly, he submitted that, even if individual elements were unprotectable, WPL had reproduced compilations of those elements. In support of each of these arguments he relied in particular (but without limitation) on category (1) of the similarities relied on by SAS Institute.

So far as the first argument is concerned, Counsel for SAS Institute drew a contrast with Nova, where the only similarities between the defendants’ programs and the claimant’s were at a high level of generality. I accept that in that respect there is a factual distinction between the present case and Nova. I do not accept, however, that it is a material distinction so far as the present issue is concerned. For the reasons I have endeavoured to explain, I consider that what matters in this context is not the degree of detail with which the defendant has reproduced the claimant’s work, but the kind of skill, judgement and labour that is embodied in what has been reproduced.

Turning to the second argument, counsel for SAS Institute submitted that WPL had reproduced the form of expression of the mathematical formulae set out in the SAS Manuals in the WPS source code. In principle, this argument seems to me to address the issue from the correct perspective. As I have said, however, the evidence does not establish that SAS Institute’s employees expended any particular skill, judgement or labour in devising the form of expression of those formulae as distinct from other mathematically equivalent formulae. Furthermore, it is not possible to translate such formulae directly into source code and WPL has not done so. As I have said, there is some resemblance of form between the WPS source code and the formulae in the SAS Manuals, but it should not be overstated. Having regard to the quality as well as the quantity of what has been taken, and provided that it was permissible for WPL to reproduce the ideas etc. described in the SAS Manuals, I am not persuaded that WPL has reproduced a substantial part of the expression of the SAS Manuals.

As to the third argument, counsel for SAS Institute did not dispute that a single mathematical formula was unprotectable by copyright. He submitted, however, that the SAS Manuals contained a compilation of statistical methods and that WPL had reproduced that compilation. Thus in the case of example 14, he argued there were at least 10 possible methods of calculating percentiles in the statistics literature, SAS had exercised skill, judgement and labour in selecting five of those 10 while omitting the remainder and WPL had reproduced that selection in WPS. In my judgment, however, this argument involves an intellectual sleight of hand. If there is a compilation of statistical methods, the compilation is to be found in the SAS System. The authors of the SAS Manuals did not create that compilation, they simply reproduced it. Although the authors of the SAS Manuals undoubtedly exercised skill, judgement and labour in devising the form of expression of the Manuals, and in particular their wording, they did not exercise any skill, judgement or labour in making the compilation of statistical methods. It follows that, to use the language of Infopaq, what WPL reproduced from the SAS Manuals in this respect does not “express the intellectual creation” of the authors of the SAS Manuals.

I would add that I do not accept that the SAS System does contain a compilation of statistical methods. It is clear from the evidence that the collection of statistical methods contained in the current version of the SAS System has grown steadily over a period of around 35 years. The collection has grown by accretion as a result of both customer feedback and competitive pressure as well as SAS Institute’s own initiatives. While there may be individual procedures or functions in respect of which it could be said that an author or group of authors consciously planned and implemented a particular selection of methods, in general the collection was not planned and there was no overall design. It was therefore not the intellectual creation of an author or group of authors. In this respect it stands in the same position as the collection of commands in Navitaire: see the judgment of Pumfrey J at [92].

As noted above, WPL does not dispute having reproduced certain SAS test scripts from SAS Manuals albeit in source code which is used for testing rather than in WPS itself (category (9) of the similarities relied on by SAS Institute). This complaint raises two issues which require separate consideration.

The first issue is whether this amounts to the reproduction of a substantial part of the relevant SAS Manuals. This issue is very similar to the question of whether the 11 word extracts in Infopaq amounted to the reproduction of a substantial part of the newspaper articles. As I have already said, it is clear from Infopaq that the test to be applied in these circumstances to determine whether what has been reproduced in itself expresses the author’s own intellectual creation. There is no evidence as to the skill, judgement and labour expended in the creation of the two scripts which are the subject of examples 24 and 25. I do not even know who the authors were. I have been left to make my own assessment. Doing the best I can, I note that each of these two scripts is both very short and very basic. In the absence of evidence to the contrary, the skill, judgement and labour required to create them appears trivial. On the other hand, WPL has copied them precisely. With some hesitation, I have come to the conclusion that they do not express an intellectual creation and so WPL has not reproduced a substantial part of the SAS Manuals. The point of the requirement for a substantial part to be reproduced is that the copying of insubstantial parts of a copyright work is lawful.

The second issue is whether, if I am wrong and WPL has reproduced a substantial part of the SAS Manuals, it had an implied licence to do so. It is important to be clear as to the scope of the issue here. WPL does not suggest that it was in any better position than any other reader of the SAS Manuals. Nor does WPL suggest that it had an implied licence to reproduce other parts of the SAS Manuals. WPL’s case is based squarely on the fact that both of the scripts which are the subject of examples 24 and 25 were explicitly included in the SAS Manuals as “examples”.

Counsel for WPL relies in this regard on the evidence of Mrs Council:

“Q. I am not asking you technical questions about what this does, Mrs Council, I just wanted to ask you about the purpose of including an example like this in the manual. Why is it put there?

A. It is an example to show how the SORTN CALL would work and the results that it would produce.

Q. Would you expect someone who read this manual maybe to type that into their computer, run it and see how it worked and that it produced the result stated?

A. I'm not aware that people do that but it is possible. In this particular case, it could be done because all the data is there.

Q. Isn't that one of the reasons why you provide these sort of examples so people can try out the way these routines work with a little bit of sample code?

A.

Actually the examples are there along with the result and we assume that people believe the result that we give them but perhaps they check it out for themselves.”

As I think Mrs Council accepted, a reasonable reader of the SAS Manuals would presume that it was acceptable for him to type these scripts into his SAS System to verify the functioning which the examples illustrate. There is nothing in the SAS Manuals to suggest that he cannot do so 10 times, 100 times or 1000 times if so minded.

Counsel for SAS Institute submitted that it did not follow that the reader could use the scripts to develop a competitive system. That is not an accurate description of how WPL use them in these examples, however. Instead, WPL uses them to check the consistency of WPS’s output. On the assumption that WPS is not itself an infringing copy of the SAS source code or Manuals, I cannot see that there is any difference between running the scripts repeatedly for verification purposes through an interpreter supplied by SAS Institute (the SAS System) and an interpreter lawfully produced by WPL (WPS). Accordingly, I conclude that the defence of implied licence succeeds.

It is common ground that WPL purchased copies of four versions of the SAS Learning Edition:

WPL purchased two copies of version 1.0 in October 2003, one copy of version 2.0 in March 2005, two copies of version 4.1(a) in July 2007 and seven copies of version 4.1(b) in March 2009.

Each version of the Leaning Edition was supplied with a “click-through” licence which required the user to accept the terms of the licence before the user was permitted to access the software. The terms of the licence for use of versions 1.0 and versions 2.0 are slightly different to those for use of versions 4.1(a) and (b).

The terms of the licence for versions 1.0 and 2.0 include the following clauses:

“Important: Please carefully read the terms and conditions of this License Agreement (‘Agreement’) before clicking on the ‘Yes’ button. By clicking on the ‘Yes’ button, the individual licensing the Software (‘Customer’) agrees to these terms, and SAS Institute Inc (‘SAS’) will authorize Customer to use the SAS Learning Edition Software (‘Software’) in accordance with the terms and conditions of this Agreement. If Customer does not agree to all of the terms of this Agreement, click on the ‘No’ button and return the Software to the supplier from whom Customer obtained it. Customer will receive a refund of licence fees applicable to and paid by Customer for this Software, if any, provided Customer returns the full Software package with proof of purchase within thirty (30) days of date of purchase or the supplier’s return period, whichever is the longer.

1.

Licence Grant

1.1

In exchange for Customer’s payment of all applicable fees and compliance with all of the terms and conditions of this Agreement, SAS hereby grants Customer a[n individual] nonassignable, nontransferable and nonexclusive license to use the Software on one (1) workstation at a time, for Customer’s [own self-training] non-production purposes only. Concurrent usage or use on a network is not authorized.

1.2

Customer may not use the Software in any outsourcing, facilities management or service bureau arrangement or any data or information technology management operation by or for third parties. … Customer may not distribute the Software to any third party in modified or unmodified form.

…

2.

Intellectual Property Rights

The Software is copyrighted. ... Customer may not reverse assemble, reverse engineer, or decompile the Software or otherwise attempt to recreate the Source Code, except to the extent applicable laws specifically prohibit such restriction….

…

9.

Complete Agreement

This Agreement and any invoices relating to the Software set forth the entire agreement between Customer and SAS related to the Software and supercede any purchase order, communications or representations regarding the Software…. ”

The terms of the licence for versions 4.1(a) and (b) are the same, except that they omit the italicised words in square brackets.

It is clear from the dates set out in paragraphs 267 and 268 above that any use by WPL of the Learning Edition prior to July 2007 was of versions 1.0 or 2.0 and hence the relevant licence terms were those applicable to those versions. Equally, any use of the Learning Edition after December 2008 was of version 4.1(a) or (b) and hence the relevant licence terms were those applicable to those versions. Between July 2007 and December 2008 the use may have been of either version 2.0 or version 4.1(a).

All the versions of the licence terms are governed by the law of the State of North Carolina. There is no evidence of North Carolina law, and it is common ground that I should assume that it is no different from English law so far as is relevant to this case.

SAS Institute contends that WPL has (i) acted in breach of the licence terms in developing WPS, and hence (ii) infringed the copyright in the Learning Edition. The importance of this contention to the parties is demonstrated by the fact that, not only was this sole claim made by SAS Institute when the proceedings were first launched, but also it was the claim which counsel for SAS Institute took first in both his opening and closing submissions.

It is common ground that the licences should be construed in accordance with the principles of contractual interpretation stated by Lord Hoffmann in Investors Compensation Scheme Ltd v West Bromwich Building Society [1998] 1 WLR 896 at 912-913. As Lord Hoffmann summarised it in the later cases of Bank of Credit and Commerce International SA v Ali [2001] UKHL 8, [2002] 1 AC 251 at [37] and Chartbrook Ltd v Persimmon Homes Ltd [2009] UKHL 38, [2009] 1 AC 1011 at [14], the correct approach is to ask what a reasonable person having all the background knowledge which would have been available to the parties at the time would have understood them to be using the language of the contract to mean.

Counsel for SAS Institute submitted, and I accept, that the relevant background knowledge included the following matters:

Counsel for WPL submitted that the licence terms should be construed contra proferentem. Counsel for SAS Institute submitted that this doctrine did not assist WPL for the reasons discussed in Lewison, The Interpretation of Contracts (4^th ed, Sweet & Maxwell, 2007) at pages 265-267. I accept that submission.

There are two issues of construction of the terms of the licences. The first is as to the interpretation of the term “Customer”. The second concerns the expression “non-production purposes”.

Customer. SAS Institute contends that the licence is personal to the individual human being who clicks the “yes” button and thereby agrees to the terms and becomes the Customer. Counsel for SAS Institute accepted that a company could acquire a copy of the Learning Edition for an employee to enable that employee to learn the SAS Language, and that the company could pay for that copy. He also accepted that in such a case there would be a contract between SAS Institute and the company. He nevertheless submitted that the introductory section clearly defined “Customer” as the individual who accepted the licence terms and that it was clear from clause 1.1 that the licence was personal to the Customer so licensed.

WPL contends that, where the Learning Edition is purchased by a company, the Customer is the company. Counsel for WPL submitted that this interpretation was supported by the fact that both the introductory section and clause 1.1 referred to the Customer paying the licence fees for the software and by the fact that clause 1.1 limited the licence to a single workstation. He also submitted in such circumstances the individual who clicked the “yes” button to accept the terms did so as agent for the company.

I prefer SAS Institute’s construction. The introductory section explicitly defines the Customer as being the individual who clicks the “yes” button and thereby accepts the terms. Consistently with this, the licence granted by clause 1.1 is personal to the Customer since it is non-transferable, a point which is reinforced by the last sentence of clause 2. The fact that the licence is personal to the individual is clearer in the terms applicable to versions 1.0 and 2.0, which includes the words “individual” and “own self-training”, than it is in the terms applicable to versions 4.1(a) and (b), but I reach the same conclusion in respect of both. The limitation to use on a single workstation is an additional restriction on the scope of the licence, and it does not support the conclusion that the licence extends to the company. WPL’s construction would have the consequence that a large company could purchase a single copy of the Learning Edition, install it on a single workstation and then train a large number of employees on it. That does not appear to me to be how a reasonable person having the background knowledge set out above would understand the Agreement.

I do not see that it makes any difference to this if the individual in question is acting in the course of his or her employment by a company and the company pays for his or her copy of the Learning Edition. In the present case it is common ground that (i) WPL’s employees were acting in the course of their employment, (ii) WPL paid for the copies of the Learning Edition and (iii) WPL contracted with SAS Institute. I agree with counsel for WPL that in these circumstances the employees acted as agents for WPL. It does not follow that the employees cannot be contracting parties. As Lord Scarman said in Yeung Kai Yung v Hong Kong and Shanghai Banking Corp [1981] AC 787 at 795:

“It is not the case that, if a principal is liable, his agent cannot be. The true principle of law is that a person is liable for his engagements (as for his torts) even though he is acting for another, unless he can show that by the law of agency he is to be held to have expressly or impliedly negatived his personal liability.”

There are numerous cases in which even disclosed agents have been held personally liable on contracts made on behalf of principals, a recent example being Foxtons Ltd v Thesleff [2005] EWCA Civ 524, [2005] 2 EGLR 29. That being so, the fact that the employees entered into the licences as agents for WPL does not lead to the conclusion that “Customer” must be interpreted as meaning their principal.

Non-production purposes. SAS Institute contends that the restriction of the licence to “non-production purposes” excludes any commercial use of the Learning Edition. Counsel for SAS Institute submitted that the licence was limited to use of the Learning Edition by an individual for their own self-education. He accepted that this was clearer in the terms applicable to versions 1.0 and 2.0, but submitted that it was clear enough from the terms applicable to versions 4.1(a) and (b), particularly when read against the background set out above.

WPL contends that the restriction of the licence to “non-production purposes” excludes use of the Learning Edition by the customer to process data for production purposes, that is to say, to produce output of value to the customer in its business, trade or profession. Counsel for WPL submitted that the purpose of the limitation was to prevent customers from using the Leaning Edition as a cheap substitute for the Full Edition.

In my view the meaning of the expression “non-production purposes” in this context is not clear. On balance, however, I prefer SAS Institute’s construction. It seems to me that the essence of the restriction is that the Customer must not use the Leaning Edition to produce anything, but only to learn about the SAS Language, how to write SAS scripts and how to use the SAS System. The terms applicable to versions 1.0 and 2.0 are more explicit and thus clearer in this regard, but with some hesitation I reach the same conclusion in respect of the terms applicable to versions 4.1(a) and (b).

SAS Institute alleges that WPL has acted in breach of the licence terms in two distinct ways. First, SAS Institute alleges that more than one employee of WPL has used copies of the Learning Edition purchased by WPL. Secondly, SAS Institute alleges that WPL has used the Learning Edition in ways which extend beyond the scope of the licence.

There is little dispute as to the facts pertaining to these allegations. So far as the first allegation is concerned, it is clear from the evidence of WPL’s witnesses that copies of the Learning Edition purchased by WPL were used by more than one employee, at least prior to WPL’s acquisition of multiple copies of version 4.1(b) in March 2009. It follows that such use was outside the scope of the licence as I have construed it. For the avoidance of doubt, use of each copy of the Learning Edition by the individual who clicked the “yes” button was licensed, subject to the second allegation.

So far as the second allegation is concerned, it is clear from the evidence of WPL’s witnesses that WPL has used the Learning Edition in the following ways:

In my judgment these uses were outside the scope of the licence as I have construed it. Counsel for WPL argued that, even if the terms of licence were to be interpreted as contended for by SAS Institute, it was necessary to distinguish between (a) use of the Learning Edition to learn how the SAS System functioned and (b) the subsequent creation and testing of WPS. I do not accept this argument. In my view it is clear that, if licence terms are to be construed as I have construed them, WPL’s use of the Learning Edition went beyond the scope of the licence.

It is common ground that sections 50BA and 296A of the 1988 Act must be interpreted in the same way as Articles 5(3) and 9(1) of the Software Directive for the reasons set out in paragraph 163 above. As is now customary in intellectual property litigation in this country, counsel argued the case directly by reference to the provisions of the Software Directive.

It is also common ground that the drafting of Article 5(3) of the Software Directive is less than pellucid in its clarity. As Professor Blocher observes in Walter and von Lewinski (eds), European Copyright Law: A Commentary (OUP, 2010) at §5.5.43:

“Like the previous one, this paragraph is also not drafted satisfactorily. Taken literally, it essentially states a pleonasm, namely, that the person having a right to use a copy of a computer program shall be entitled to do all the acts he is entitled to do.”

Professor Blocher goes on to suggest that recital [19] gives little assistance in clarifying the provision, but that the legislative history provides more help.

It is also common ground that, if and to the extent that the licence terms for the Learning Edition are contrary to Article 5(3), then by virtue of the last sentence of Article 9(1) they are null and void.

SAS Institute contends that Article 5(3) is a “for the avoidance of doubt” provision, which simply confirms that acts of observing, studying and testing a computer program are not infringements provided that the user is licensed to use the program in the manner in question. Counsel for SAS Institute advanced a number of arguments in support of this contention.

First, he submitted that the words “which he is entitled to do” at the end Article 5(3) could not be ignored and had to be given effect to. If the user was not entitled to do certain acts because his licence did not extend to those acts, then Article 5(3) did not apply.

Secondly, he argued that, as a derogation from the restricted acts in Article 4 of the Software Directive, Article 5(3) should be construed narrowly, in accordance with the principle stated by the ECJ in Infopaq at [57].

Thirdly, he pointed out that Article 6 of the Software Directive was expressed in terms which were plainly intended to conform with the three-step test set out in Article 9(2) of the Berne Convention and Article 13 of TRIPS (see Article 6(3) and recital [24]), whereas Article 5(3) was not. He did not contend that Article 5(3) should nevertheless be interpreted and applied by national courts in accordance with the three-step test (cf. the effect of Article 5(5) of the Information Society Directive on Article 5(1) as interpreted in Infopaq at [58]). Instead, he argued that, if Article 5(3) had been intended to operate as a substantive defence to acts which would otherwise infringe copyright, then it would have been expressed in a way which conformed with the three-step test.

Fourthly, he relied on the fact that recital [19] states that the user should not be prevented from observing, studying or testing the program “provided that these acts do not infringe the copyright in the program”.

Fifthly, he submitted that this interpretation was supported by the legislative history of the Software Directive. Article 5(3) has its origin in new Article 5(2a) of the European Parliament’s amended text, which provided:

“Notwithstanding the provisions of Article 4(a), the legitimate holder of a copy of a program may, without having to request the authorisation from the right-holder, observe, study or test the working program in order to determine its underlying ideas, principles and other characteristics where these are not protected by copyright, in the course of loading, viewing, running, transmission or storage.”

The Commission adopted a modified form of this provision in Article 5(5) of its Amended Proposal. In the commentary to this the Commission stated:

“A further provision has been incorporated into Article 5 of the amended proposal in line with the opinion of Parliament. Paragraph 5 is intended to make clear that a person who has a right to use a copy of a computer program is not prevented, as some commentators have suggested, from studying how the program functions. The form in which a computer program is usually supplied is not readily accessible to the human user. It is in a form known as ‘object’ or ‘machine-readable’ code, which, even when rendered visible to the human observer, is difficult to decipher in large quantities. Nevertheless, as the program is run, a person skilled in the art may observe and test the functioning of the program by a variety of means, including the use of electronic testing and monitoring techniques. The use of such techniques does not involve reproducing, translating or adapting the program. Such techniques do not therefore infringe the author’s rights in his program.

The amended proposal makes it clear that if a person has a right to use a program, that right must include at least the ability to load and run the program. During such running of the program any non-infringing act necessary to observe, study or test the functioning of the program may be carried out.

If in addition to loading and running the program, the user is also entitled to display, transmit or store the program, he may observe, study or test the functioning of the program during these operations also. He may not, however, claim rights to perform acts beyond these necessary for use or permitted under the licence merely in order to carry out additional study of the functioning of the program.”

Sixthly, counsel for SAS Institute argued that the history of the UK’s implementation of Article 5(3) demonstrated the importance of the words “which he is entitled to do”. When the UK first implemented the Software Directive, there was no section 50BA: the whole exception was covered by section 296A(1)(c) as it then stood. At that time section 296A(1)(c) did not contain the words “which he is entitled to do”. The Report from the Commission to the Council, the European Parliament and the Economic and Social Committee concerning the implementation of the Software Directive dated 10 April 2000 (COM(2000) 199 final) noted that the UK (and Finland) had omitted these words and stated:

“This omission is significant in that the phrase ‘which he is entitled to do’ ensures appropriate limitation of the permitted acts which can be performed by the user to observe, study or test the functioning of the program”.

The 1988 Act was subsequently amended by introducing section 50BA, and amending section 296A(1)(c) to refer back to section 50BA.

WPL contends that the words “the acts … which he is entitled to do” in Article 5(3) refer to the kind of acts which the user is entitled to do, not to their purpose. Thus WPL contends that the kinds of acts in question are those referred to immediately before the words “which he is entitled to do”, namely “loading, displaying, running, transmitting or storing the program”. Thus if the licence does not entitle the user to transmit the program, Article 5(3) does not permit transmission; but if the licence entitles the user to load and run the program, then the user can observe, study and test while loading and running. Counsel for WPL advanced a number of arguments in support of this contention.

First, he submitted that the words “without the authorisation of the rightholder” were crucial, and made it clear that the purpose of Article 5(3) was to prevent the performance of restricted acts even by a lawful acquirer of a copy of the program being dependent on the authorisation of the rightholder. Otherwise, the rightholder could achieve indirect protection for ideas and principles even though they could not be directly protected by the copyright in the program.

Secondly, he pointed out that the acts of “loading, displaying, running, transmitting or storing the program” referred to in Article 5(3) corresponded to the acts which were restricted by the copyright in the program by virtue of Article 4(a), but not those restricted by virtue of Article 4(b) (translation, adaptation etc). He submitted that this confirmed that the words “which he is entitled to do” referred to the kind of restricted act.

Thirdly, he submitted that it could not be right to interpret Article 5(3) as depending on the purpose for which the user was licensed to use the program. Otherwise, a licence for, say, academic use would prevent the program being observed, studied and tested with a view to creating an interoperable program.

Fourthly, he argued that recital [19] was consistent with this interpretation.

Fifthly, he argued that the contrast between Article 6(3) and Article 5(3) showed that the legislature considered that the latter inherently complied with the three-step test.

Sixthly, he submitted that the legislative history, and in particular the passage from the Amended Proposal quoted in paragraph 298 above, supported WPL’s interpretation rather than SAS Institute’s.

Seventhly, he argued that the history of the UK implementation of Article 5(3) was entirely neutral on the point.

I have come to the conclusion that this is an issue upon which it is necessary to seek the guidance of the ECJ. It is plain that the interpretation of Article 5(3) is not acte clair. On the contrary, it is attended by considerable difficulty. Moreover, it is an issue upon which an authoritative ruling can only be given by the ECJ, which will have the advantages of being able to compare the different language versions of the Software Directive, to consider all the travaux préparatoires and to receive the submissions of the Commission. Nevertheless, in case it is of assistance to the ECJ, I shall express my own provisional view.

My view is that WPL’s interpretation is to be preferred for the reasons given by counsel for WPL. I would add two points. First, the starting point is that Article 5(3) is expressed to be an exception to the restricted acts referred to in Article 4. In my opinion it follows that it should be interpreted as a positive defence to a claim of copyright infringement and not merely a “for the avoidance of doubt” provision with no substantive effect. This is a pointer against SAS Institute’s interpretation.

Secondly, Article 5(3) must be read together with the last sentence of Article 9(1). That makes it clear that the copyright proprietor cannot override Article 5(3) by contract. To allow copyright proprietors to override Article 5(3) by the use of standard form licence terms of the kind relied on by SAS Institute in the present case would make it very easy for the proprietors to circumvent Article 5(3). That would be contrary to the important public interest which underlies Article 5(3).

On the assumption that Article 5(3) is to be construed as WPL contends, it follows the licence terms for the Learning Edition are null and void to the extent that they make it an infringement for the user to observe, study and test the Learning Edition in order to determine the ideas and principles which underlie any element of the program. Counsel for SAS Institute submitted that, even on that assumption, Article 5(3) did not entitle WPL to commit all the acts listed in paragraph 289 above.

Again, I regard this as a matter on which the guidance of the ECJ is required. My provisional view, however, is that the words “ideas and principles which underlie any element of the program” are to be construed in the same way in Article 5(3) as in Article 1(2) of the Software Directive, that is to say, broadly, for the reasons explained in paragraph 208 above.

On that basis, I would hold that none of the acts listed in paragraph 289 was an infringement, except perhaps the last.

SAS Institute’s claims in respect of the WPS Manual and the WPS Guides raise distinct issues and must be considered separately.

The only issue here is whether WPL has reproduced a substantial part of the relevant SAS Manuals. WPL contends that it has gone no further than indirectly copying the content of the SAS Manuals via the WPS source code, and that this is not an infringement because all it has done is to copy the ideas, procedures, methods of operation and mathematical formulae described in the SAS Manuals rather than the expression of the SAS Manuals.

On the assumption that the law is correctly stated in paragraphs 199-207 above, I accept that, in so far as WPL merely copied the ideas etc. described in the SAS Manuals, then that was not an infringement. Thus I do not consider that WPL infringed the copyright in the SAS Manuals by including syntax diagrams in the WPL Manual which concisely describe the syntax of the SAS Language elements supported by WPS. In this regard it should be noted that the form of expression used by WPL in the WPL Manual is not only conventional, but different from that used by SAS Institute in the SAS Manuals. As illustrated by the CALL SCAN example reproduced in the Appendix, whereas the SAS Manuals contain syntax statements, the WPL Manual contains syntax diagrams. Similarly, I do not consider that WPL infringed the copyright in the SAS Manuals by reproducing the relevant keywords for reasons I shall explain in more detail in relation to the WPS Guides.

Even on that basis, however, I have concluded that WPL went further than merely copying the ideas etc. described in the SAS Manuals. For the reasons given in paragraph 148 above, I consider that WPL substantially reproduced the language of the SAS Manuals even though its policy was not to do so. Applying the test laid down in Infopaq, namely whether the parts which have been reproduced express the intellectual creation of the authors of the SAS Manuals, I conclude they do.

I have briefly described the WPS Guides in paragraph 75 above. Mr Jupp confirmed in evidence that he had obtained the lists of keywords in the WPS Guides from the SAS Manuals. In simple terms, he made a long list of the headings, sub-headings and sub-sub-headings in the SAS Manuals. Subsequently WPL has indicated which keywords are supported.

There are two issues in respect of the WPS Guides. The first is whether they reproduce a substantial part of the SAS Manuals. The second is whether, if so, WPL has a defence of fair dealing for the purposes of criticism or review under section 30(1) of the 1988 Act.

Reproduction of a substantial part. SAS Institute contends that the list of keywords represents a substantial part of the SAS Manuals. In support of this contention counsel for SAS Institute argued that the SAS Manuals contained a compilation of keywords, and WPL had reproduced the entirety of that compilation. In my judgment this argument suffers from the same flaws as the argument discussed in paragraphs 259 and 260 above. First, if there is a compilation of keywords, the compilation is to be found in the SAS System. The authors of the SAS Manuals did not create that compilation, they simply reproduced it. Secondly, I do not accept that the SAS System contains a compilation of keywords. Once again, the collection of keywords in the SAS System grew by accretion. It is not an intellectual creation of an author or group of authors. Accordingly, applying the test in Infopaq, WPL has not reproduced the intellectual creation of the authors of the SAS Manuals.

Fair dealing for the purpose of criticism or review. In case I am wrong on the first issue, I shall briefly address the second issue. In this regard, I should record that neither counsel addressed me on Article 5(3)(d) of the Information Society Directive.

Counsel for SAS Institute submitted that this defence failed for three reasons: (i) the keywords had not been reproduced for the purpose of “criticism or review”, (ii) the dealing was not fair and (iii) it was not accompanied by a sufficient acknowledgement. I shall consider these points in turn.

So far as the expression “criticism or review” is concerned, counsel for WPL relied on the statement by Robert Walker LJ (as he then was) in Pro Sieben Media AG v Carlton UK Television Ltd [1999] 1 WLR 605 at 614 that the expression “should be interpreted liberally”. He also submitted this approach was reinforced by the requirement in section 3 of the Human Rights Act 1998 to construe legislation compatibly with the right of freedom of speech under Article 10 of the European Convention on Human Rights. I accept that submission, which is supported by the judgment of Lord Phillips of Worth Matravers MR (as he then was) in Ashdown v Telegraph Group Ltd [2001] EWCA Civ 1142, [2002] Ch 149 at [30]-[33] and [45], the judgment of Mann J in Fraser-Woodward Ltd v British Broadcasting Corp [2005] EWHC 472 (Ch), [2005] FSR 36 at [36] and (by analogy) the judgment of Lord Phillips of Worth Matravers LCJ (as he then was) in HRH Prince of Wales v Associated Newspapers Ltd [2006] EWCA Civ 1776, [2008] Ch 57 at [25].

Counsel for WPL went on to argue that the WPS Guides did reproduce the keywords for the purposes of “review”, because the WPS Guides compared the functions supported by WPS with those available in the SAS System. He said that it did not matter for this purpose that the WPS Guides contained no comment on that comparison. Although this involves taking a broad view of what constitutes review, I have come to the conclusion that in the particular circumstances of the present case I accept this argument. A third party reviewer of two rival computer programs might well wish to inform potential customers for the two programs which functions they respectively supported. In the case of rival interpreters, this could well extend to informing potential customers which functions of one interpreter the other supported. I do not see that it ceases to be a review if one of the suppliers of the programs makes the same comparison.

Turning to the question of fairness, counsel for SAS Institute relied on three main points as supporting the conclusion that the reproduction of the keywords was not fair. First, he submitted that the infringement competed with the copyright work. I do not accept this. It is, of course, true that WPS competes with the SAS Components. But that does not mean that the WPS Guides complete with the SAS Manuals. In my judgment they do not compete to any material extent. Secondly, he relied on the fact that the WPS Guides reproduced all of the keywords. I am unimpressed by this. The extent of reproduction is still very small compared to the SAS Manuals. Furthermore, the extent of the reproduction was proportionate to the purpose of the review. Thirdly, he submitted that it was not necessary to reproduce the keywords. I disagree. In my view it was necessary for WPL to reproduce the keywords in order systematically to compare the functionality of WPS with that of the SAS System.

As to whether there was a sufficient acknowledgement, there is no dispute that the SAS Manuals were published anonymously and thus no identification of the authors was required. Counsel for SAS Institute submitted that WPL had failed to identify the SAS Manuals by their title or other description. Counsel for WPL submitted that the identification need not be formal, relying by analogy on the decision of the Court of Appeal in Pro Sieben that the corporate author of a television programme could be identified by its logo. As a matter of principle, I accept that submission. The difficulty on the facts of the present case, however, is that I cannot see that the WPS Guides contain any acknowledgement of the SAS Manuals as a source whatsoever. The WPL website from which the WPS Guides may be downloaded clearly acknowledges the SAS Components by name, but in my view that is not the same thing.

Accordingly, if it were necessary, I would hold that WPL’s reproduction of the keywords in the WPS Manual was fair dealing for the purpose of review, but that the defence failed for want of a sufficient acknowledgement of the SAS Manuals.

For the reasons given above, I have concluded that the resolution of this case depends on a number of important issues of interpretation of Articles 1(2) and 5(3) of the Software Directive and Article 2(a) of the Information Society Directive. Since this court is not a court of last resort, I have a discretion as to whether or not to refer questions to the ECJ. Counsel for SAS Institute submitted that, unless I was able to decide the case in favour of SAS Institute, I should exercise my discretion to make a reference. Counsel for WPL opposed this course. He submitted that, even if I concluded that the resolution of the case depended on issues of European law which were not acte clair, I should proceed to determine those issues myself and leave it to the appellate courts to decide whether or not to refer questions to the ECJ. When I suggested to him that it would not be in WPL’s interests for me to decline to refer questions now, only for the Court of Appeal to make a reference, since the result would simply be further delay and expense, he disputed this. He pointed out that there had been an order for a speedy trial and submitted that the purpose of that order would be frustrated if I were to make a reference. Furthermore, he submitted that his client would rather have a decision in its favour now, even if there was a likelihood that the Court of Appeal would subsequently refer questions to the ECJ.

In my view it is almost inevitable that, if I were to decline to refer questions to the ECJ, the Court of Appeal would do so. In those circumstances, it seems to me that the effect of my declining to refer would simply be to cause the parties further delay and expense before the dispute can finally be resolved. I cannot see how this can be in the interests of either party. Accordingly, I have decided to refer questions to the ECJ myself.

For the reasons given above, I conclude as follows:

I shall hear further argument on the precise formulation of the questions to be referred to the ECJ. The parties should exchange proposed drafts of the questions in advance of that hearing. The parties should also consider the guidance given by Arden LJ in R (Horvath) v Secretary of State for the Environment, Food and Rural Affairs [2007] EWCA Civ 620, [2007] EuLR 770 at [80].