Copyright Act, Division 4A
44 In view of his conclusion that there had been no breach of the licence agreement, it was unnecessary for the primary judge to determine whether the making of the DR Copy was also authorised by Division 4A of Pt III of the Copyright Act 1968 (Cth) (the Act). Nevertheless he went on to hold that it was. Division 4A was introduced by the Copyright Amendment (Computer Programs) Act 1999 (Cth). It is headed "Acts not constituting infringements of copyright in computer programs" and consists of ss 47AB to 47H. Section 47C(1), which deals with "Back‑up copy of computer programs", is in part as follows:
[T]he copyright in a literary work that is a computer program is not infringed by the making of a reproduction of the work if:
(a) the reproduction is made by, or on behalf of, the owner or licensee of the copy (the original copy) from which the reproduction is made; and
(b) the reproduction is made for use only by, or on behalf of, the owner or licensee of the original copy; and
(c) the reproduction is made for any of the following purposes:
(i) …
(ii) to enable the owner or licensee of the original copy to store the reproduction for use in lieu of the original copy if the original copy is lost, destroyed or rendered unusable;
(iii) to enable the owner or licensee of the original copy to use the reproduction in lieu of the original copy, or of another reproduction made under this subsection, if the original copy, or the other reproduction, is lost, destroyed or rendered unusable.
45 Section 47C(2) provides in part:
[T]he copyright in a literary work that is a computer program, and in any work or other subject‑matter held together with the program on the same computer system, is not infringed by the making of a reproduction of the program, or of such a work or other subject‑matter if:
(a) the reproduction is made by, or on behalf of, the owner or licensee of the copy (the original copy) from which the reproduction is made; and
(b) the making of the reproduction is part of the normal back‑up copying of data for security purposes.
46 The primary judge expressed the view that the elements of pars (a), (b) and (c)(ii) and (iii) of s 47C(1) were made out, and that the testing of the DR Copy that RWWA had carried out was thereby authorised. His Honour was also of the view that the requirements of s 47C(2) were made out, on the ground that the reproduction had been made by a licensee as part of the normal back‑up copying of data for security purposes.
47 The primary judge made no orders consequent upon his observations about s 47C, and having regard to our conclusion that RWWA was entitled to test the DR Copy under the licence agreement, we do not need to determine whether he was correct in concluding that s 47C led to the same result. However, his Honour's third declaration states that RWWA is entitled to test the DR Copy pursuant to s 47F as well as pursuant to the licence agreement. Accordingly we need to consider whether the declaration was properly granted under s 47F.
48 Section 47F(1) provides:
Subject to this Division, the copyright in a literary work that is a computer program is not infringed by the making of a reproduction or adaptation of the work if:
(a) the reproduction or adaptation is made by, or on behalf of, the owner or licensee of the copy of the program (the original copy) used for making the reproduction or adaptation; and
(b) the reproduction or adaptation is made for the purpose of:
(i) testing in good faith the security of the original copy, or of a computer system or network of which the original copy is a part; or
(ii) investigating, or correcting, in good faith a security flaw in, or the vulnerability to unauthorised access of, the original copy, or of a computer system or network of which the original copy is a part; and
(c) the reproduction or adaptation is made only to the extent reasonably necessary to achieve a purpose referred to in paragraph (b); and
(d) the information resulting from the making of the reproduction or adaptation is not readily available to the owner or licensee from another source when the reproduction or adaptation is made.
49 The primary judge was of the view that all elements of s 47F(1) were established. His Honour said at [234]:
From s 47F…, it follows in my view that the process of reproduction for the DR Copy is authorised as … the reproduction is made for the purpose of testing in good faith the security of the original copy … In other words, the testing is necessary in order to be sure that if the System at the main premises at Osborne Park is destroyed, the security of the original copy of the System is tested.
50 His Honour then observed (at [235]) that while "testing in good faith the security of the original copy" in par (b)(i) might on one view be given a more restrictive meaning, that does not appear to be the legislative purpose having regard to its context, together with a passage in par 10.18 of the Copyright Law Review Committee's Report, Computer Software Protection (1994) which indicates that "security" in the expression "security purposes" includes security against disasters such as fire, earthquake and terrorist attack. He concluded at [236] that s 47F "applies to protect the testing process carried out by RWWA". He added at [237]:
'Security testing' is directed towards the possible failure of a system or network 'of which the original copy is a part'. Testing as to possible failure of a system (namely, the DR Copy) is the purpose of testing at the DR Site. Accordingly s 47F … also applies to protect RWWA.
51 The primary judge's reference to a possibly "more restrictive meaning" of "testing in good faith the security of the original copy" would appear to be to treat it as limited to "security against electronic or other invasion of a system" (at [232]).
52 His Honour's resort to par 10.18 of the Committee's Report is attended by some difficulty. At [232] he observed that the amendments to the Act in 1999 "use 'security' at s 47F in a broader sense than security against electronic or other invasion of a system". However, par 10.18 deals with what became s 47C and is directed to the expression "security purposes" in s 47C(2)(b). Section 47F did not derive from the Committee's Report. There is no doubt that "security purposes" in s 47C has the wider meaning his Honour had in mind. If that is not apparent from the provision itself - "back up copying of data for security purposes" - it is made clear by the reference in the Explanatory Memorandum to "security against the possibility of damage to or destruction of the computer or system", and by par 10.18 of the Report.
53 It is probable that the primary judge meant in [232] to refer to s 47C rather than s 47F. The context supports that theory. Paragraphs [221] to [231] all concern s 47C. His Honour did not turn to s 47F until [233]. Given that context and the fact that par 10.18 is concerned with what became s 47C, [232] should be treated as dealing with s 47C. However, on that view his Honour would appear to have used par 10.18, dealing with "security" in the expression "security purposes" in s 47C(2)(b), as elucidating the meaning of "the security of the original copy" in s 47F(1)(b)(i).
54 SAG submitted that RWWA's testing process is not "testing … the security of the original copy". The original copy is the installed copy at the Osborne Park site. Thus the testing that is permitted is of that copy and not another copy such as that installed at the DR Site. It is said that resort to the second limb of s 47F(1)(b)(i) does not assist RWWA because it too is directed to the computer system or network at Osborne Park - "of which the original copy is a part". Paragraph (b)(ii) is similarly limited to the "original copy" or a system or network of which it forms part. SAG took issue with the primary judge's reasoning, quoted at [49], that RWWA is ensuring the security of the licensed installation at Osborne Park by testing whether the back‑up site at KAZ's premises is going to work. RWWA did not grapple with this submission.
55 What s 47F(1) permits is the reproduction of the original copy for the purpose of testing the security of that copy. The original copy is the copy RWWA is licensed to use. The permitted testing is of the security of that copy. The passages from the primary judge's reasons quoted at [49] appear to us to be saying that the testing of the functionality of the DR Copy at the DR Site is the testing of the security of the original copy at Osborne Park. That, in our view, is not what s 47F(1) authorises. On the facts of this case, what it permits is the making of a copy of the installed copy at Osborne Park for the purpose of testing the security of the installed copy. As it seems to us, the primary judge's construction of the provision enables the DR Copy at the DR Site to be tested so as to determine its efficacy should the installed copy at Osborne Park for some reason be no longer available.
56 SAG propounded a further reason why RWWA's testing process is not "testing … the security of the original copy" in s 47F(1)(b)(i). It relied on Mr McKemmish's opinion at pars 6.6 and 6.7 of his report. Mr McKemmish's letter of instructions asked him to answer the following questions (amongst others):
15.6 what is the industry standard meaning of the expression 'security testing';
15.7 whether the Testing Process as set out above comes within the industry standard meaning of the expression 'security testing' …
The background information in the letter of instructions, and its summary of RWWA's and SAG's claims, is concerned mainly with issues under the licence agreement. The letter did not direct Mr McKemmish to s 47F or any other provision of the Act, although the language of RWWA's s 47C claims were incorporated into the summary in the letter. The agreement does not employ the words "security testing". The operative part of s 47F speaks of "testing … the security of the original copy". However, s 47F is headed "Reproducing computer programs for security testing" and the two questions set out about were probably asked with that expression in mind.
57 In any event, Mr McKemmish responded to the first question as follows:
6.6.2 The Oxford dictionary of computing defines 'security evaluation' as "the examination of a system to determine its degree of compliance with a stated security model, security standard, or specification. The evaluation may be conducted (a) by analysing the detailed design, especially of software, often using verification and validation, (b) by observing the functional behaviour of the system, or (c) by attempting to penetrate the system using techniques available to an 'attacker' ". Security testing is, in my opinion, a subset of the security evaluation process. Namely security testing can be considered the actual physical or tangible assessment or probing of a computer system's security.
6.6.3 Such testing would include, but not be limited to, testing the user authentication process, firewall and router configurations, intrusion detection system operations, system logging, open and closed ports, user account auditing and access control rules.
58 The dictionary quoted at 6.6.2 is the Oxford Dictionary of Computing (5th ed, Oxford University Press, 2004).
59 Mr McKemmish answered the second question, which he headed "Security testing and the disaster recovery testing", as follows:
6.7.2 In my view the Testing process as described in the material provided to me is not 'security testing'. There is no evidence to suggest that the logical system security of the RWWA system is assessed during the testing process. Rather the testing process is in fact a disaster recovery and business continuity test, the purpose of which differs from security testing.
60 Earlier in his report Mr McKemmish had described RWWA's testing process, based in part on information supplied to him by RWWA. The description is too lengthy to quote in full. Its essentials are these:
(a) The KAZ disaster recovery facility operates an IBM mainframe environment compatible with the RWWA IBM mainframe environment.
(b) RWWA employs volume mirroring as the primary form of replication, the net effect of which is to create an exact copy of RWWA's working production system on the KAZ mainframe.
(c) Where a disaster is declared, as simulated by the disaster recovery exercises, there are a number of steps that must be taken in order to make the KAZ system operational:
· KAZ readies its disaster recovery facilities
· the disaster recovery canisters are recalled
· external organisations are notified
· mirroring between RWWA mainframe and KAZ mainframe is cut (set to simplex state)
· the RWWA partition on the KAZ mainframe is powered up
· the mirrored copy of the production system is activated
· mainframe internet settings are changed to reflect the new disaster recovery site settings
· software keys are updated
· mirroring of distributed systems is cut
· the distributed systems' addresses are configured to the disaster recovery site
· firewall is activated
· business systems are activated, and
· synchronisation of current race data is performed.
61 Mr McKemmish states the effect of the steps described in [60(c)] as follows:
The net effect of this process [is] to make operational the mirrored copy of the production system. Given that it is a mirrored copy, the disaster recovery copy does not require installation of the object code files from the original distribution tapes. Additionally the mirroring process also eliminates the need for all of the compilation and linking processes typically required during the system installation, to be replicated.
62 Mr McKemmish was briefly cross‑examined about the security testing part of his report. It was put to him that while the primary meaning of "security testing" involves protecting the integrity of data from outside influences, an acceptable use of "security" in the industry is security against the possibility of damage to or destruction of a computer or computer system. He did not agree that "security" is used in the latter sense. He was then asked whether he had heard of testing the security of a system against unforseen disaster such as fire, earthquake or terrorist attack. He said he had not, and that he would not "put the two together in that context".
63 In his supplementary report of 19 May 2008, in response to a request for his comments on Mr McKemmish's opinions on "security testing", Mr Adams said nothing of present utility. In his further supplementary report, in response to another request for his opinion about the meaning of "security testing", he appears to have misunderstood the enquiry. In commenting on s 47C(2)(b) he said the expression "security testing" does not appear in the Act, and that Mr McKemmish's comments on the terms "security evaluation" and "security testing" "apply to a different context". He did not refer at all to what might be meant by "testing … the security of the original copy", the expression to which his attention had been drawn. However, in cross‑examination he was asked about Mr McKemmish's understanding of the concept. He agreed that security testing involved verifying the security of a program "in terms of hackability and firewalls and things like that, some sort of concept of integrity in the system itself". He added that it involved "[protecting] the integrity of the data from outside interference". He said it was a secondary and different sense of the word "security" when people talked about the security of a business overall - that was "business continuity rather than strict security testing". He then agreed that, based on everything he had seen and read thus far, he had not witnessed any security testing, in the first of these senses, conducted in relation to the DR Copy.
64 Mr Fink's supplementary report of 30 April 2008, in response to a request for his opinion on the meaning of "security testing", throws no light on the issues canvassed by Mr McKemmish. In his cross‑examination he said he was quite happy with what Mr Adams had said (recorded at [63]), though he added that security testing is not a term he had heard used in the mainframe environment, as opposed to the personal computer environment.
65 Mr McKemmish's evidence in relation to security testing, which was not seriously challenged and was agreed in by Mr Adams, and with the reservation referred to by Mr Fink, supports the view that RWWA's testing process at the DR Site was not "testing … the security of the original copy" within the meaning of s 47F(1)(b)(i). Their evidence also supports the view that the word "security" in the expression in s 47C(2)(b) - "the normal back‑up copying of data for security purposes" is not used in the same sense as "security" in the expression in s 47F(1)(b)(i) - "testing … the security of the original copy". The former would seem to contemplate the protection of a business and the latter the protection of a program.
66 Some support for SAG's contention that RWWA's testing process at the DR site is not "testing … the security of the original copy" is provided by Ricketson and Creswell's The Law of Intellectual Property: Copyright, Designs & Confidential Information (2nd ed, 2002) at 11.195. After observing that some of the language in s 47F is far from clear, and that the word "security" may not be used in the same sense in pars (i) and (ii) of s 47F(1)(b), the authors go on:
In ordinary parlance, however, the word "security" alone would refer to the protection of something, whether it be a building, chattel or program, against the possibility of unauthorised access by third parties, and accordingly there seems to be a redundancy of language used here, unless the terms "security of the original copy" or "security flaw" are intended to refer to something else. Possibly, this "something else" could be the vetting of the program for bugs and viruses where this has not occurred as the result of unauthorised access by a third party, for example, where an authorised user of a computer system uses data or a program derived from a defective file or disc.
At a more general level, however, it can be said that s 47F(1)(b) contemplates two distinct situations: subpara (i) is concerned with testing when the administrator does not know whether the system has a security flaw or is vulnerable to unauthorised access and wants to find out whether pre‑emptive action can be taken, while subpara (ii) is concerned with taking investigative and/or corrective action where the administrator has become aware that the program or system does in fact have such a flaw or vulnerability.
67 The Explanatory Memorandum to the Bill that became the Copyright Amendment (Computer Programs) Act 1999 (Cth) provides some assistance in discerning the intention behind the enactment of s 47F. It discloses that the copying predominantly envisaged by s 47F is that generated by, amongst others, decompilation or disassembly processes undertaken so as to test the security of the original copy or to investigate or correct security flaws in the original copy. See especially pars 1.1 and 19. In this connection we observe that the problem to which security testing in s 47F is directed is the need, without it constituting infringement, "to test [a computer program's] security to protect [it] against abuse (or 'hacking') and viruses": par 1.1. As there pointed out, testing can involve decompilation or making a temporary copy, which is an infringement if done without the copyright owner's permission. The section is not, by contrast, directed towards testing permanent back up or disaster recovery copies to ensure they will work if and when required.
68 For the above reasons we are unable to accept RWWA's contention, which the primary judge appears to have adopted, that "testing … the security of the original copy" extends to what was done at the DR Site, namely testing of the DR Copy to ensure that the System would be capable of being restarted and operated without the loss of data. In our view, "testing … the security of the original copy" should be confined to testing the original to ascertain its security from unauthorised access or against electronic or other invasion.