You might think that Windows 95 is well out of date. But how out of date will it be in the year 2070, when it finally comes out of copyright?
Computer software is protected by the law of copyright (which originated in England in 1710) and patents (which go back to the fifteenth century). That these systems are less than a perfect fit for the protection of software shouldn't come as much of a surprise. But are we stuck with what we have, or is there a real prospect that the law could be reformed to provide a better and fairer system for the protection of intellectual property in software?
Perhaps the loudest voices that are raised against the law as it stands are those who oppose intellectual property rights of any sort. They argue that unlike other forms of property, intellectual property can be reproduced at no cost (in fact, there is usually a cost involved in stopping people from reproducing it). They also note that one person's use of intellectual property generally does not detract from another's, and that works of intellectual property usually build on what has gone before, so that authors are able to charge for the use of their work disproportionately to the effort they put into it.
Although all these statements are true enough, the fact that there are problems with applying property law to software does not imply that we must accept that intellectual property is inherently impractical or immoral. For most people, that is an extreme position. Although intellectual property may be capable of being reproduced cheaply, it is certainly not as cheap to create, and without an economic incentive to do so, authors will be less inclined to produce as much. To put it crudely, we live in a capitalist economy, and we all have to eat.
Apart from that, there is also the practical point that the politically powerful software industry - not to mention the music and motion picture industries - are scarcely likely to sanction the abolition of the intellectual property regime, but they may cooperate with its careful reform. There are good reasons why they might wish to do so, too, as the problems in the application of intellectual property law to software are certainly not one-sided. A more appropriate intellectual property system for software should benefit developers as well as users.
There are two paths that law reform could take: firstly to work within the bounds of the existing intellectual property regime by tinkering around the edges of copyright and patent law, or secondly to overhaul the existing law completely as it applies to computer software and to replace it with something more appropriate.
Before we can determine which path to take, let's take a step back and consider exactly what the problems of the existing law are. The inadequacies of copyright law are actually quite different to those of patent law, but even when the effect of the two is combined, we are left with an ineffective and anachronistic patchwork that satisfies the interests of neither developers nor the general public.
Copyright law treats software as a literary work, the same as a book, a poem or the script of a play. Presently software remains protected in Australia for fifty years after it was published or fifty years after the death of the author, whichever is later.
If we are to attempt to reform the existing law, one of our first ports of call should be to look at the term of copyright protection that software attracts. Given the rate of development in computer software and hardware engineering, this term is clearly excessive, as a 50 year-old software program has no conceivable commercial value, but could be of historical or educational interest to other developers and the public at large.
One of the lessons of the Internet is that when the law falls out of step with the community, the community will rewrite it. Thus, the concept of "abandonware" has sprung up. Abandonware is software that is at least five (some say three, or seven) years old and is no longer available to be commercially licensed, and which users copy freely on the assumption that it has been abandoned. Although this may sound reasonable, it has absolutely no basis in law; copying of such "abandoned" software is a breach of copyright to exactly the same extent as if the software copied was the latest version of Microsoft Office.
But perhaps to legitimise the concept of abandonware would be a sensible concession to public interests in the reform of copyright law as it applies to software. If an abandonware amendment to the copyright law were to be introduced, the questions to be considered would include:
Apart from the length of time for which copyright enjoys protection, another issue is that users and developers are prohibited from modifying the software for their own use unless they are specifically licensed by the copyright owner to do so. Admittedly, the law does now allow software to be copied for the purposes of interoperability, error correction and security testing, but without access to the source code these rights are of limited use. Of course, this is exactly the problem that open source software addresses, but although it goes without saying that a software developer should be entitled to put their work into the public domain or to release it under an open source licence, it is quite another thing to require them to do so.
Perhaps, then, it should be a condition of copyright protection that the source code of software is made available to those who have licensed the object code – or even, that it be made publicly available for other developers to learn from, but not to copy. This may sound radical, but really it isn't. Patent law already requires that a detailed specification of the protected intellectual property be made publicly available. In the United States, it is already necessary for a partial deposit of source code to be made before a copyright owner can sue for damages for the infringement of copyright in that software.
Patent law is also in need of reform as it applies to computer software. There are those who criticise software patents for stifling innovation by locking up inventive ideas in private hands, but in many ways patent law is a closer match for the needs of developers and the public than than copyright law is. After all, a literary work as we conventionally understand it has little in common with software source code, and much less in common with object code; but for purposes of patent law an idea is an idea, no matter whether the method of operation it describes is found in a software program or a mousetrap.
Interestingly, though, patents never used to be available for computer software, or (more generically) for business processes. It used to be necessary for an idea to interact with the physical world before it could be protected. As soon as that requirement fell away, there was an explosion in the grant of patents for software process ideas, and unfortunately the standard of assessment of some of these patents was fairly low. Although patentable ideas are supposed to be novel and to involve an inventive step that advances the state of the art, some of the highest profile patented software has been criticised for being obvious and involving very little inventiveness at all.
Australia does already have a system of 8-year “innovation patents” that is similar to what Bezos describes, but unfortunately it is not compulsory. Innovation patents are granted more easily than regular patents, because they require a lower standard of inventiveness, and are not assessed by a patent examiner until they are challenged by a competitor. Many software authors will opt for an innovation patent because it is cheaper and quicker to obtain, but if they want their software process to remain protected for the full 20 years, the original patent system remains available to them.
Copyright law doesn't protect ideas, it protects the form in which ideas are expressed. As long as you don't copy or adapt the source code or documentation directly, simply rewriting it is usually enough to avoid the copyright. And whilst patents do protect certain types of ideas, they don't protect the results of those ideas, so if you are able to re-engineer a software process to achieve the same result by a different method, the patent can be overcome.
Are patents and copyright enough, then, to protect the interests of software developers? Although this is open to debate, there is certainly a school of thought that says they are not. In particular, they do not prevent a “clone” of a software application from being developed, particularly where the original application is insufficiently new and original to qualify for patent protection. To give a trivial example, the classic computer game Tetris has been cloned many times over, and its author has no recourse against those who have done so unless they used a confusingly-similar name for their program (and only because that would be an infringement of his trademark in the name).
If neither copyright nor patent law quite meet the needs of either developers or users of software (since, to be fair, neither were originally intended to cover this field), the alternative is to develop an entirely new form of intellectual property protection for computer software. This is not such an unlikely idea; there are already special systems of intellectual property protection for plant breeders and circuit designers.
In fact, the idea of introducing specific intellectual property rights for the protection of software goes back thirty years, when IBM sponsored the introduction of a special intellectual property regime for software into the US House of Representatives. The failure of this bill is now history, and it is unlikely that any of the major software houses or copyright collection agencies would now resurrect a similar proposal.
There is, however, still a fair amount of academic interest in the development of a new regime for the protection of software, with a 1996 paper recommending the development of a law covering “misappropriation of software behaviour”. Supplementing the existing copyright and patent regimes, this would protect developers from the appropriation of the distinctive behaviour of their software, even if this did not involve a literal copying of code or the infringement of a unique patentable method of operation.
Although such a new right would be useful to developers, it wouldn't much help the general public while software also remained subject to excessively lengthy terms of protection under copyright and patent law. What is really needed is for software to be exempted from the operation of patent and copyright law, so that the new regime that is brought in its place does not subject users to the same restrictions. In particular, it would be important to ensure that users had the right to use the source code of software to correct bugs, and that they could adapt it for their own needs or for purposes of engineering interoperability with other software.
This is all well and good in theory, but what are the real prospects of introducing such radical changes to the law? One barrier that we do face is that Australian intellectual property law, and in particular copyright law, does not operate in isolation from the rest of the world. We are a signatory to a number of international instruments such as the Berne Convention for the Protection of Literary and Artistic Works, and the World Trade Organisation Agreement on Trade-Related Aspects of Intellectual Property Rights, which require us to observe minimum standards of copyright protection. Most relevantly, the Berne Convention requires copyright protection of literary works to subsist for at least the life of the author plus 50 years, and the WIPO Copyright Treaty requires us to regard a computer program as a literary work. (It could be said that this only applies to source code rather than object code, but that is rather a debatable point.)
It is not impossible to conceive that Australia could, in cooperation with other countries, change the rules of the game – it was only in 1996 that the most recent amendments to international copyright law were made – but this is an ambitious and lengthy process. In the meantime, there is some wriggle-room within the international constraints under which we operate, for example, we can place whatever conditions we wish (such as the deposit of source code, or the payment of a registration fee) on the ability of a party to sue and recover statutory damages for copyright infringement.
Happily, now is exactly the right time for interested parties to make thoughts like these known, as the Commonwealth government is currently reviewing its 2000 “Digital Agenda” reforms to the Copyright Act, which introduced what few software-specific provisions we do have (such as the right to copy software for error correction and interoperability). Even if we are stuck for the time being with a couple of square pegs for a decidedly round hole, there is at least scope for us to shave one of those pegs around the edges.