I am about to run a presentation at Linux.conf.au 2009 in Hobart. I want to particularly talk about Section 47D of the Copyright Act 1968 (Cth). Section 47D was a great affirmation of the right to reverse engineer computer programs for interoperability. One especially important feature is that 47D, thanks to s 47H, is not excludable by contract. In the United States, we have seen numerous examples of developers inadvertently waiving their fair use and statutory rights to reverse engineer software products by 'agreeing' to the EULA.¹⁾

The really big problem with s 47D, however, is that it only applies to computer programs. This is a really big problem for those who are interested in reverse engineering media rich applications – like computer games. Because games are not only computer programs but are also cinematograph films,²⁾ sound recordings, artistic works, musical works, etc, then the right to reverse engineer (and to backup) completely disappears.

If we believe that we really ought to have a right to reverse engineer computer programs, including games, the wording of s 47D has to be changed to immunise copying of works and other subject matter intertwined with software programs.

The inflexibility in s 47D also raises problems for the exceptions in anti-circumvention law. The definition of both Access Control Technological Protection Measures (ACTPMs) and the broader category of Technological Protection Measures (TPMs) in s 10(1) of the Copyright Act excludes devices to the extent that they ”[restrict] the use of goods […] or services in relation to the machine or device.” Now, this carve-out is quite important, as it effectively excludes restrictions like the anti-competitive encoded garage door opener device³⁾ from protection as TPMs. However, because these devices are built often to have more than one function, especially in gaming consoles, this carve-out will not always be effective. This means that developers interested in making products for locked-down devices will have to rely on the exceptions to anti-circumvention law itself, rather than the definitional carve-out.

Liability for actual circumvention (ACTPMs) or for making or distributing a circumvention device (all TPMs) does not apply where the device will be used to do an act that does not infringe copyright in the computer program and is
“done for the sole purpose of achieving interoperability of an independently created computer program with the original program or any other program.”⁴⁾

As you can see in that wording, the ability to rely on the exception is limited by what exactly will infringe the underlying copyright interests. So without a fix to s 47D, we generally can't feel comfortable relying on the exceptions to anti-circumvention.

This also raises an important point – manufacturers and distributors of modchips need to meet the purposive test to ensure that they fit within the exception. That is, they must be able to show that the modchip will be used for non-infringing reverse engineering purposes.

My slides are available here: 200901-lca-games.pdf

In late June a decision came down (which I appear to have missed) in the England and Wales Court of Appeal quashing the conviction of Mr Modchips (aka Neil Higgs) for providing modchips in contravention of the UK anti-circumvention provisions in s.296ZB of the Copyright, Designs and Patents Act 1988 (UK).

The case, Higgs v R [2008] EWCA Crim 1324 (24 June 2008), makes for interesting reading.

The Court of Appeal held that the wording in the UK act: “prevention or restriction of acts that are not authorised by the copyright owner of that work and are restricted by copyright.” is to be read similarly to the provision in the Australian act: “to prevent or inhibit the infringement of copyright”. Accordingly, the reasoning of the High Court of Australia in Stevens v Sony [2005] HCA 58 has some application in the UK.¹⁾

The Court of Appeal concluded that a technological measure must actually prevent the infringement of copyright:

In the end, therefore, one comes back to the UK Act. Is it enough if the technological measure is a discouragement or general commercial hindrance to copyright infringement or must it be a measure which physically prevents it? To our minds the position is clear – it is the latter. Neither the Directive nor the Act would have been drafted in the way that they are if such a general form of hindrance was enough.²⁾

This all sounds like good news. Especially for Mr Modchips himself. However, the court warned that future defendants may not be so lucky, especially given that reproduction into RAM is an infringing act:

It is for those reasons we quashed the convictions. Mr Higgs is a fortunate man in that it may well be that if the legislation had been less complex and/or the Crown had had greater opportunity to consider the details of copyright law the case would have been proved on the basis that merely playing a pirated game involves making a copy in the console and thus involves infringement. He may also be fortunate that, at least this far, he has not been sued in the civil courts. There the procedure is apt to be much faster, technical slip-ups in evidence can generally be readily cured before final judgment and the remedies of damages, an account of profits, injunction and legal costs are readily obtainable. Breach of an injunction, if serious, can of course itself lead to imprisonment.³⁾

This decision places the UK law in a very similar position to the Australian law after Stevens v Sony. There is still considerable doubt as to whether modchips are legal under Australian law after the changes to the definition of 'material form'. After the changes introduced by the AUSFTA, 'Material Form' now “includes any form (whether visible or not) of storage of the work or adaptation, or a substantial part of the work or adaptation, (whether or not the work or adaptation, or a substantial part of the work or adaptation, can be reproduced).”⁴⁾ This means that a device which prevents the playing of an unauthorised game can now be argued to prevent an infringement, rather than merely render an already infringing copy less useful. This means that the devices in the current batch of consoles may be technological protection measures, and this in turn means that modchips may now be circumvention devices, despite their usefulness for other, non-infringing, purposes.

In Australia, blind people are able to access texts in braille and books on
tape, but the demand for these media is decreasing. Blind people today are
increasingly reliant on texts in electronic form, and these are much less
readily available in Australia. Electronic texts are more portable and less
cumbersome than large braille volumes, and are much faster to navigate
than audio recordings. However, in Australia it is difficult for blind people to
get access to a wide range of electronic texts and there exists no scheme
enabling such access. At the same time sighted people are using electronic
text and other digital media at an ever-increasing rate. In order to
approximate the same level of access as sighted people, blind people
require access to accessible electronic versions of all published material.
The authors suggest that given the legal imperatives of Australia’s domestic
legislation, treaty obligations and social values, that there exists a moral
imperative to create a scheme providing blind people with access to digital
print media.

Thanks to the MALR and LexisNexis, it's available here under a CC BY-NC-SA 2.5 (AU) licence. Attribution must be to the authors and the publisher (first published by LexisNexis and MALR).

Published as Nicolas Suzor, Paul Harpur and Dilan Thampapillai, 'Digital copyright and disability discrimination: From braille books to bookshare' (2008) 13(1) MALR 1.

The seminar went very well last week. Below you can find the slidesets from the presentations.

• Bjorn Bednarek, Games Classification (8MB PPT)

• Nic Suzor, Virtual Liberties, (2MB PPT or 2MB ODP)

Presentations available under Creative Commons Attribution-ShareAlike 2.5 (Australia).

In 2005, the Australian High Court ruled that the combination of the boot ROM and region coding in Sony PlayStations and PlayStation games was not a Technological Protection Measure (TPM). Under the old law, a device had to “prevent or inhibit” the infringement of copyright. Because Sony's technology didn't prevent a copy being made, but only stopped the copy being played in the PlayStation, then it couldn't prevent or inhibit the copying that had already taken place.

This may have changed in the recent changes to Australia's copyright legislation. The definition of a Technological Protection Measure (TPM) was changed from a device designed to 'prevent or inhibit', to a device designed to 'prevent, inhibit, or restrict' the infringement of copyright.¹⁾ This effect of inserting the broader term 'restrict' seems to be to legislate around the High Court's decision in /Stevens v Sony/, meaning that a  PlayStation may well be a TPM under the new law, and dealing with mod-chips may soon be illegal.

The changes also introduce a new category of protected devices, called Access Control Technological Protection Measures (ACTPMs). An ACTPM is defined as a device used within Australia in connection with the exercise of copyright, which controls access to the copyright work.²⁾ For ACTPMs, unlike TPMs, it is illegal not only to manufacture or supply a circumvention device or service, but also to actually circumvent the measure. Previously, if you could obtain a circumvention device (mod-chip) from outside of Australia, there would be no restriction on you using that mod-chip yourself; but you couldn't have it done for you. Now, if the device is an ACTPM (but not a TPM), then actual circumvention may result in civil liability,³⁾ and, if done for commercial advantage or profit, criminal sanctions.⁴⁾ For both TPMs and ACTPMs, the manufacture, importation, distribution, or communication of a circumvention device, or the provision of a circumvention service, will attract both civil and criminal liability, with punishments up to 5 years imprisonment.⁵⁾

There is an exception in the new law which provides that a region coding device is neither an ACTPM or TPM, and a device or service designed to circumvent it will not be either a circumvention device or a circumvention service respectively. This means that a device which only has the function of preventing use in Australia of a film, game, or computer program which was purchased outside Australia, will not be protected. (It is important to note that this exception is very narrowly worded – for example, the law may not apply to allow circumvention of a device which restricts use of a game or movie to a particular State or city within Australia.)

The situation becomes more complicated when a single device has functions which both prevent use in Australia and either (a) control access to the copyright work (an ACTPM); or (b) prevent, inhibit or restrict the infringement of copyright (a TPM). Most technical locks on current game consoles arguably have both of these functions – the Sony PlayStation itself, if found to be a TPM under the broader definition of TPM, has both a TPM function and a region coding function.

The question becomes how the new exception will be interpreted, and how tightly manufacturers couple unprotectable region codes with protected TPMs or ACTPMs. In their submission to the House of Representatives Standing Committee on Legal and Constitutional Affairs (LACA) review into Technological Protection Measures, the IEAA stated that “[a]ccess controls used to enforce region coding are tightly coupled with additional and inseparable access controls that distinguish genuine from pirated games”.⁶⁾ The LACA Committee did not accept that this must always be the case, however, and noted that it should be practically possible to isolate region coding elements from TPM elements.⁷⁾

Where the manufacturer creates a dual purpose device, it will be an ACTPM or TPM, but not “to the extent” that it controls market segmentation. This can be read to mean that to the extent that it otherwise controls access, the device will still be an ACTPM or TPM. If a user can circumvent only the regional coding, and no more, then that circumvention will be permissible. A more difficult question arises when the user, in circumventing the region coding, also circumvents the TPM, and whether it was strictly necessary to do so or not. It is unclear under the new law whether circumvention will be prohibited in this case.

There is an argument that can be made that if it is not possible to separate the functions of a legitimate TPM or ACTPM from region coding functions, then none of the combination will be a TPM or an ACTPM. Neither the legislation nor the secondary materials, however, provide any clear guidance as to the extent to which this argument is correct. If this argument is correct, then it is still unclear that where it may be technically possible but extremely difficult and expensive to circumvent only the region coding, whether a 'reasonableness' test would be implied to excuse circumvention of both.

While it may be technically possible to circumvent only the region coding portion of a dual purpose device, it may be prohibitively difficult to do so. Most mod-chips bypass the entire device completely. In the Blizzard v BNetD case, the modders had to bypass a check to allow them to use their games on an interoperable server. In doing so, and because they didn't have the information required to implement their own checks, they also bypassed the CD-key authenticity check.⁸⁾) This was held to be an illegal circumvention. If a modder is required to bypass only the region coding portion of a dual purpose device, instead of a simple bypass, they may have to engage in very difficult low level reverse engineering, and may even require confidential information to reconstruct a working access control. By relying on this requirement, a manufacturer can construct a region coding which, while technically legal to remove, would be very nearly (but not quite) impossible to remove without also removing the access code.

When asking, then, whether mod-chips will be permitted under the new law, there are a number of questions:

• Does the device “control access” to a copyright work? If so, then the device may be an ACTPM, and circumvention will normally be prohibited, as will dealing with or providing mod-chips or circumvention services.

• If not, does the device “prevent, inhibit or restrict” the infringement of copyright? If so, the device may be a TPM, and while actual circumvention by people with the technical skill to do so will be allowed, dealing with or providing mod-chips or circumvention services will be prohibited. The new wider language of this test means that the device in the PlayStation may well be a TPM, where it wasn't under the old law.

• If the device also implements region coding, does the mod-chip only circumvent the region code and leave the access control or the TPM intact? If so, the mod-chip will be legal.

• If the device also implements region coding, and the mod-chip circumvents both the region code and the access control or TPM, the position is still unclear. A Court may ask whether it was reasonable for the chipper to circumvent the TPM or ACTPM in order to bypass the region code. Alternatively, a Court may take the narrower approach and ask only whether it was technically necessary to do so. This is a fundamental uncertainty in the new law, and one which may take another test case, many years, and potentially another High Court decision to resolve.

It is not clear whether mod-chips are legal in Australia. The new law is more restrictive than the old law, because the definition of TPM is broader. This means that a mod-chip which allows the playing of homebrew games may not be allowed, where it would have been under the old law after /Stevens v Sony/. A mod-chip which only allows playing of games from other regions will probably be allowed. What about a device which does both? The answer is unclear at best.

The great danger of this new legislation, in addition to the fact that the types of protected devices has been significantly expanded, lies in the ability of a copyright owner to design around the limitations of the wording. A TPM can be designed to also be an ACTPM, in order to obtain protection from actual circumvention. An unprotected region code can be rendered protectable by creating a dual purpose device in which the functions are separated but any method of circumvention necessitates circumvention of both the access control and the regional coding. The manufacturer is the only person in the position to confine a device to specific functions, but this legislation does not impose any clear obligation to do so. Instead, it provides an incentive to create dual purpose devices.

There remains much uncertainty as to whether the circumvention of region coding in Australia will be legal under the new legislation, and the ambiguity within the legislation provides a potential loophole for manufacturers to obtain legal protection for devices which would not ordinarily be protected. It will now be up to the Courts, again, to determine when Australians have the right to remove arbitrary or anticompetitive restrictions imposed upon their property by copyright owners.

Imagine that the makers of a board game only allowed you to play it on wet Saturday afternoons. Now imagine that the Australian Government made it illegal to play that game on any other day. If the board game is released on a DVD or another electronic format, the new proposed amendments to Australia's copyright legislation do exactly that.

From the first of January 2006, it will be illegal to access electronic material in a way that the makers don't allow. Since 2000, it has been illegal to make or sell devices which break digital locks, but not to break the digital lock itself. Following the Australia – United States Free Trade Agreement, Australia agreed to extend liability for circumventing these locks, or Technological Protection Measures (TPMs).

The 2005 landmark High Court case of Stevens v Sony highlighted that anti-circumvention law (as it stands before these proposed amendments) exists to “prevent or inhibit copyright infringement”. It does not, and should not, exist to prevent consumers from using material that they purchase in ways which do not infringe copyright. In that case, Sony could not prevent people from modifying their PlayStations in order to play games they legitimately purchased overseas.

Copyright owners may have legitimate reasons for placing locks on their material. These locks can certainly make it more difficult for users to make infringing copies of movies, music, or books in electronic form. But these locks also have illegitimate purposes when they seek to restrict use, not copying.

Anti-circumvention law has had a troubled history on this point. The High Court ruled that copyright owners couldn't control the use of their material. The House of Representatives Standing Committee on Legal and Constitutional Affairs agreed, and recommended that, in order to grant protection to these digital locks, we should “clearly require a direct link between access control and copyright protection”. The Committee also recommended that these locks should not be protected if they restrict competition. The Attorney-General's Department has clearly said that it accepts both of these recommendations. Indeed, the first draft of the copyright amendments clearly defined TPMs as devices which prevent or inhibit copyright infringement.

The first draft enshrined the principle highlighted by the High Court that Australians should have the liberty to use their legally acquired property as they see fit. In three short weeks and a legislative two step, that principle has been destroyed. The new copyright bill doesn't require protection of copyright interests. Instead, it gives copyright owners the broad ability to control use of their work, if they can merely show that the lock is used “in connection with the exercise of copyright”. Interpreted broadly, this potentially means that any lock will be protected from circumvention, as long as it controls access to copyright material.

This new broad language has enormous consequences for Australian consumers. The proposed law will give copyright owners wide ranging powers to restrict the use of copyright material any way they want. Subject to competition law and express exceptions for region coding, the copyright owner now has the unfettered ability to determine how you can use your legitimately purchased DVD, game, music, or electronic book. We're talking about CDs you can't use in your car, games you can't play with your kids, and books you can't read on Thursdays. This may seem far-fetched, but it is really a fundamental reshaping of consumer law in the guise of copyright legislation.

Copyright exists to protect people from unauthorised copying of their material. It does not exist to allow the authors of a legitimately purchased item to control how a consumer uses it. There is no justification for giving this power to copyright owners. International law doesn't require it. The Australia – US Free Trade Agreement doesn't require it. Even our High Court has said that we shouldn't do it. So why have we? What has changed so much in the last three weeks, that the Australian Government has suddenly severely reduced the rights of consumers without any public debate or explanation? Why exactly have we sold out our consumer liberty to the benefit of large media corporations?

Copyright law is important. If Australia is to remain competitive in the global marketplace, we need to ensure that hard work and investment in the knowledge economy does not go unrewarded. But that doesn't mean that we should protect the interests of copyright owners at any cost. This new bill destroys the balance between copyright owners and consumers. It gives protection to the most absurd of limits that can be imposed by technology. There is no reason why copyright owners need to be able to control how and where we interact with copyright material. This bill simply goes too far.

The only way that Australia will survive such a bold attempt to reduce our freedom and quality of life is for ordinary Australians to actively engage in this debate – go online, talk to family and friends, and most importantly, voice your opinions to your local member and your state Senators. We have little more than a few weeks before parliament will decide what law will be implemented. Educate yourself on these issues, as they promise to fundamentally alter our rights, our economy, and the way we use digital technology. Say it clear: technological protection measures should protect copyright itself, not the use of copyright material.

Blizzard make several popular games, including Warcraft, Diablo and Starcraft. Online multiplayer in these games is limited to using Blizzard's Battle.net service. Battle.net provides a mechanism for users to create and join multi-player games, to meet and chat with other users, and to record statistics and participate in tournaments. Battle.net functionality is built into the games. Blizzard's Battle.net servers check the validity of users' cd-keys when a user connects to the service from within the game. This validation is known as the “secret handshake” which allows only users with valid cd-keys to continue connecting to Battle.net.

The Eight Circuit Court of Appeals has determined that an open source replacement to Blizzard's Battle.net violates the DMCA by bypassing this 'secret handshake'.

Blizzard's End User License Agreements on the games themselves state that a user may not “in whole or in part, copy, photocopy, reproduce, translate, reverse engineer, derive source code, modify, disassemble, decompile, create derivative works based on the Program, or remove any proprietary notices or labels on the program without the prior consent, in writing, of Blizzard”¹⁾

Blizzard's Terms of Use on Battle.net state that a player may not

(ii) copy, photocopy, reproduce, translate, reverse engineer, modify, disassemble, or de-compile in whole or in part any Battle.net software;

(iii) create derivative works based on Battle.net;

(iv) host or provide matchmaking services for any Blizzard software programs or emulate or redirect the communication protocols used by Blizzard as part of Battle.net, through protocol emulation, tunneling, modifying, or adding components to the Program, use of a utility program, or any other technique now known or hereafter developed for any purpose, including, but not limited to, network play over the Internet, network play utilizing commercial or non-commercial gaming networks, or as part of content aggregation networks […]

(v) use any third-party software to modify Battle.net to change game play, including, but not limited to cheats and/or hacks;

(vi) use Blizzard's intellectual property rights contained in Battle.net to create or provide any other means through which Blizzard entertainment software products […] may be played by others, including, not limited to, server emulators.²⁾

The defendants were frustrated by the poor performance of Blizzard's Battle.net service, as well as cheating and otherwise offensive players. They subsequently began free development of bnetd, which would act as a replacement server for Battle.net which gave users more control over the games they played online. To create bnetd, the defendants had to reverse engineer the protocol spoken by Battle.net and the Blizzard games, and they also developed a small utility which was used to modify the Blizzard games so they could connect to other multiplayer servers. Notably, the defendants had no way of enforcing the cd-key validity check, and were forced to treat any cd-key presented as valid.

The district court granted summary judgment to Blizzard, holding that fair-use reverse engineering could be excluded by terms in shrink-wrap or click-wrap contracts, and that the reverse-engineering exceptions in the DMCA do not protect reverse-engineering in order to create fully functional alternative products, or where the program is distributed for free.³⁾

The Eight Circuit Court of Appeals affirmed the decision. Blizzard's EULA and ToS were enforceable contracts, and the defendants had waived any fair-use defence they may have had.⁴⁾ The 'secret-handshake' constituted an effective Technological Protection Measure (TPM), and bnetd circumvented that TPM by allowing all clients to connect. The 'interoperability' exception did not apply, on the basis that the bnetd emulator allowed unauthorised copies of the Blizzard games to be played on the bnetd.org servers. The court considered that this constituted infringement of copyright, and as such, the interoperability defence could not apply. The Court did not consider whether bnetd was a dual use technology which could have both infringing and non-infringing uses, or whether the playing of an infringing copy of a game on an internet server constituted copyright infringement at all.

In Australia, reverse engineering to make interoperable products is protected as an exception to copyright by Copyright Act 1968 (Cth) s 47D. Section 47H provides that section 47D, which was inserted by the Digital Agenda amendments, can not be excluded by contract. In Australia, Blizzard could not require that its users refrain from reverse engineering.

Reverse engineering for interoperability is also an exception to circumvention of a technological protection measure, in s 116A(3), where a 'qualified person' is permitted to circumvent a TPM for a permitted purpose, which includes interoperability from s 47D. A qualified person in this case would mean the owner or licensee of the copy of the game. Section 116A(4)(b) provides a similar exception for supplying a circumvention device.

There is nothing in the text of the anti-circumvention law that prevents the right to reverse engineer for interoperability from being excluded by contract. The exceptions to infringement in s 116A are not protected in the same way as s 47D protects ss 47B(3), 47C, 47D, 47E and 47F.

This case shows that this gap in Australian anti-circumvention law can have real consequences for Australian developers. Reverse engineering for interoperability is an important exception to the exclusive rights of the copyright owner, in that it provides developers with a mechanism to make competing products, or to adapt a technology product to work in new environments.

These exceptions are important – they concern not the piracy of games, but the right of players to make use of their lawfully acquired games in the way they want. A player who purchases a game which doesn't work satisfactorily with another product, like an internet game server, should not be precluded from seeking to play the game on another interoperable server. The right to use a game is a fundamental right of a purchaser of a copy of that game, and if the game must be reverse engineered in order to enable its use, then that reverse engineering should be permissible.

Both the Copyright Law Review Committee Copyright and Contract report and the Philips Fox Digital Agenda Review recommended that the Copyright Act be amended so that the permitted purpose exceptions in s 116A(3) cannot be excluded by contract.⁵⁾ If these recommendations are not followed, there is a significant risk that the ability to create interoperable software in Australia will be crippled, and producers of computer games will be able to require that purchasers of their games are tied to their other software products and services in order to make use of the games.