The patentability of software inventions

You’re at:
News & publications >
Article

07 December 2009

The Patents Bill 2008 seeks to update and replace the Patents Act 1953. It is long overdue. It will bring patent law into conformity with the international norm. In some areas the law remains the same. One of these is the status of software inventions – they remain patentable subject matter under the Bill.

In the Patents Bill, yet to be reported back from the Select Committee, clause 13 of the Bill preserves the definition of “Invention” as used in the Patents Act 1953 and as interpreted by the case law under that Act.

Furthermore, while the Patents Bill in clauses 14 and 15 expressly excludes certain inventions from patentability, software inventions are not listed in these exclusions. This outcome in the Patents Bill is hardly unexpected, bearing in mind that in 2004 the Ministry of Economic Development when conducting Stage 3 of the law reform process discussed the issue of software inventions and called for submissions. The vast majority of those submissions were in favour of preserving the status quo. In addition when the exposure draft of the Bill was made available the issue of patentability of software attracted little attention.

However, after the Bill had been introduced into Parliament and referred to a Select Committee, a number of parties made submissions that software inventions should be excluded from patentability. Many of these, including InternetNZ, were not aware software inventions already were patentable. The issue has thus become controversial and will probably remain so irrespective of what recommendations the Select Committee makes when it reports back to Parliament.

Important Preliminary Points

1. Computer software can be analysed at different levels of abstraction. At a high level of abstraction is the procedure or “algorithm” for accomplishing a required task. This comprises a list of steps which must be performed. At the lowest level is the machine code which can be read and executed by the computer to perform the task. Above this level is the code which the programmer, using one of many programming languages, will write to implement the specified algorithms. This “source” code is subsequently converted to machine code.

As a “writing”, copyright is the intellectual property right (IPR) which will protect the latter. Copyright cannot protect ideas, only a particular expression of an idea. Ideas can only be protected, if at all, by patents. Patents are the IPRs for protecting software algorithms, independently of the code or language used to implement them. Patents and copyright protect different subject matter.

2. Inventions within any technology are only validly patentable if they are both novel (not published or used before) and non-obvious (possess inventive step) [1]. Some patents are “accepted” (or “allowed”) from time to time for inventions which are obvious. This is an unavoidable imperfection in the patent system. Such “bad” patents (or patent applications) are not confined to software inventions. To the extent that they are disruptive it is up to competitors or users to invalidate them. This happens from time to time in all technologies. A recent case where a software-cum- business method patent application was successfully opposed and on appeal declared invalid for obviousness by the High Court, was Cool 123 v Vodafone and Others [2].

3. There is a distinction between software inventions and business method “inventions”. Simply because the latter may be performed by a computer system does not make them software inventions – any inventiveness certainly will not lie in the software. While there is currently no law in either New Zealand or Australia to exclude software inventions from patentability, both countries exclude “mere schemes or plans”, which is often all a business method is. An example of case where a patent was invalidated on this ground is Grant v Commissioner of Patents [3], where the full court of the Federal Court of Australia held that “business, commercial and financial schemes as such have never been considered patentable”. They do “not produce any artificial state of affairs, in the sense of a concrete tangible physical or observable effect”. In the United States business methods may not survive the currently awaited decision of the Supreme Court in Bilski v Doll [4].

4. Software inventions may be sub-divided into different categories and some categories may be more readily considered patentable matter than others. Possible categories are:

(1) operating systems (software for managing computer hardware and software modules) such as Microsoft Windows, Apple Mac OSX, Unix, GNU Linux;

(2) application software such as: (a) business support software; (b) engineering and scientific software, and (c) games and entertainment software;

(3) embedded software, the software which derives “embedded systems”.

Embedded systems are special purpose computer systems for performing a dedicated function. Examples of uses of embedded systems include washing machines, aircraft (avionics), brushless dc motors, cars (even anti-lock braking systems), guided missiles and medical devices, including implanted devices such as Pacemakers. Microprograms which perform low level fundamental tasks in reduced instruction set processors in place of hardware can also be considered embedded software. It is in embedded software that the direct equivalence between software and hardware is to be found. The washing machine controller, for example, could be implemented using electronic circuits instead of a programmed microcomputer. The programs for embedded systems also (but not always) differ from programs for general purpose computers by the manner in which they are stored – in a chip (“firmware”) rather than hard disk.

History

Part of the history of patent law internationally is that when new technologies emerge, many become admitted as patentable subject matter, although often after a period of controversy and uncertainty. The Courts either hold that the new technology falls within existing definitions of patentable subject matter or alternatively the legislation is amended. Examples of such technologies are microbiological inventions, pharmaceutical inventions, and bio-technology inventions, with issues still arising from the latter.

In the 1960s computers and computer software emerged from the universities and research centres where they were conceived and entered the truly commercial realm. It was not long before the courts in the United States and the United Kingdom were faced with the issue of whether software inventions or software related inventions were patentable subject matter. The first case which went to the US Supreme Court was Gottschalk v Benson [5] in 1972. The claimed invention was “a method of converting signals from binary coded decimal form into binary�”. The Supreme Court held that such a claim would “be a patent on the algorithm itself” and it was well established law that mathematical formulae could not be patented. The Supreme Court took a different view in 1981 in Diamond v Diehr [6] where a claim to a process for curing synthetic rubber, which although determined by computer program, algorithm was allowed. This would be considered as an “embedded system” today as distinguished from software for a general purpose computer.

In the United Kingdom, the first case went to the courts in 1966 in Slee &Harris’ Application [7]. This seems to be the first time anywhere, where the patentability of software inventions was considered by a court. In that case a claim to “a method of operating a computer” was refused whereas a claim to “a computer when modified to operate according to a method” and a claim for “means for controlling a computer” were allowed. But the leading software case in the United Kingdom (under the Patents Act 1949) was undoubtedly Burroughs Corporations Application [7] in 1973. There the Court held that patent claims for “a method of controlling a system of computers” or “method of operating or programming a computer” were allowable.

UK Law changed when the UK ratified the European Patent Convention (EPC) and enacted the Patents Act 1977. The EPC required countries acceding to the convention to exclude from patentable subject matter “schemes, rules and methods for performing mental acts, playing games or doing business, and programs for computers” [9] , the exclusions to apply only to the extent that the patent relates to such subject matter “as such”.

The reason given by the framers of the EPC for excluding programs for computers was that a program was no different from abstract mathematical theories which were already excluded from patentability in all European jurisdictions and were expressly excluded from patentability under the EPC in the initial draft. Restricting the excluded programs to “programs as such” was designed so as not to deny patents for traditionally patentable products and processes simply because they incorporated programmed computers in their control systems. Thus, it was intended by the drafters of the EPC that patents for claiming “Programs for doing something characterised by the following steps�” should not be granted, but patents could be granted for claims to “A machine or process controlled by a program which executed the following steps �”. Thus, embedded systems constitute patentable subject matter in Europe.

New Zealand History

Up until 1995 the New Zealand Patent Office (as IPONZ was then known) refused to grant patents for computer programs despite such UK decisions as Burroughs Corporation’s Application, which in the absence of any New Zealand decision would be highly persuasive in a New Zealand Court. The Patent Office had been following the practice of the Australian Patent Office. In the 1991 Australian case International Business Machines Corporation v Commissioner of Patents [10], the Federal Court held that a claim for “A method of producing a visual representation of a curved image” on a computer display was patentable subject matter. The Federal Court followed the UK decision in Burroughs Corporation Application.

In 1995 the New Zealand Commissioner of Patents gave a decision in Hughes Aircraft Company’s Applications [11] that in view of the Australian IBM decision and the UK Burroughs decision that patents should be granted for Hughes Aircraft’s software related inventions. Hughes Aircraft’s patent applications typically claimed “A computer controlled process for determining en route air space conflict alerts status for a plurality of airborne aircraft”. This decision did not open the flood gates for all software inventions. As stated in the decision, it was essential they produced a commercially useful effect as opposed to a mere intellectual one. This decision marked a substantive change in the treatment of software related inventions by what is now the Intellectual Property Office.

It should be noted that there is still no decision by a New Zealand Court on whether software inventions are patentable subject matter.

Software as such and Inventions Incorporating Software

It will be seen from the discussion of the exclusions in the European Patent Convention that a distinction is drawn between the patenting of software or programs as such (pure software inventions) and the patenting of inventions which incorporate software control, that is embedded systems. The latter are normally allowed in Europe. The invention in the US Diamond v Diehr case also fell within in this category.

Patents for pure software inventions, that is, program algorithms as such which have no technical contribution outside the computer system itself have been allowed in the United Kingdom (see Burroughs case above) under the old 1949 Act, Australia in the IBM case and New Zealand in the Hughes Aircraft Company case.

The Effect Of Clause 13 In Excluding Subject Matter

The exclusions from patentability are not simply those set out in clauses 14 and 15 of the Bill. These are simply the statutory exclusions. There are also the exclusions established by case law under the 1953 Act and indeed its predecessors. This comes about because the definition of “invention” under clause 13, “a manner of manufacture within the meaning of section 6 of the Statute of Monopolies” is the same as that contained in the 1953 Act and its predecessors and also its English predecessors. All New Zealand case law on what is an invention and what is not, is incorporated by clause 13 and also English and Australian case law where there is no New Zealand law. Thus the definition of patentable inventions under clause 13 with reference to case law excludes mere collocations, inventions where the only material product is intellectual content, schemes or plans, mental processes, mere discoveries, and processes for the treatment of human beings (although the latter is repeated again in clause 15 of the Bill).

Submitters Arguments For Excluding Software Inventions

The following is a representative selection of arguments (in bold) used by those submitters to the Select Committee who oppose software inventions being patentable. The contrary arguments are in normal font.

1. Software patents only benefit large software companies, such as Microsoft. On the contrary, in the United States it is small software companies that have been benefiting through having software patents. Microsoft, by way of example, has been sued on many occasions by small software companies for patent infringement. There have been at least three US court decisions this year alone where the decisions have been that Microsoft has infringed. The first was infringement of a Uniloc patent (damages $388 million). The second was infringement of a Virnetx patent. The most recent win was by the small Canadian company, i4i which has been awarded $290 million in damages. Software patents in fact allow small players to sit at the same table as the large players.

2. Software is adequately protected by copyright law. Copyright (at least since the 1985 case IBM v Computer Imports [12] and the 1994 Copyright Act) only protects one expression of a software invention whether it be the code or perhaps a screen display. It does not protect higher level ideas such as algorithms or innovations in software engineering. It is a fallacy to view software as just programming.

3. Allowing software patents is a capitulation to the United States where they are allowed. On the contrary, the 1995 Hughes Aircraft Company decision to allow software patents in New Zealand followed Australian and United Kingdom law. It did not follow US law and nor was any political pressure applied by the United States.

4. Software development is handicapped if it cannot use an idea that someone has already patented. The first innovator who has patent has a temporary period of exclusivity for that innovation. That is his reward. But the patent conferring that exclusivity publishes the algorithm and gives an incentive to others to devise alternative and perhaps better algorithms for achieving the same result.

5. The acceptability of software patents was effectively enforced by a small change in legislation from the US Patent Office in the 1990s and the question has never been properly debated. In fact the change in the United States was brought about by the Supreme Court [13] in 1981 recognising that new and non-obvious developments in software technology were just as meritorious as a new and non-obvious developments in any other technology and therefore should be similarly patentable. Prior to when the USPTO had rejected software inventions since the mid 1960s. Further, the question has been debated extensively world-wide and there are volumes of publications on the topic. The debate did not start with the first US court case on the topic, but in fact commenced following the first UK case on the topic six years earlier in 1966 and following rejections of software applications by the USPTO around the same time.

6. The infamous DET patent for a method of selling goods internationally online would not have been granted if software patents were excluded. The fact is under present law a Court is very likely to follow the precedents in the Cool 123 case and the Australian Grant case and strike this or similar patents down. Furthermore, under the Bill IPONZ examiners will at last be able to examine for obviousness and a DET application would likely fail without any statutory exclusion.

7. Software is different from other technologies. New software uses or builds upon a multitude of software developments made in the past. Patents would inhibit this. All technological innovation comes about like this, there is nothing unique or special about software. Such an objection could be equally used (and equally ineffectually used) against pharmaceutical patents, electronics patents, mechanical patents and so on.

8. Software cannot be described as “in invention”. Indeed it cannot, anymore than electronics can be so described. But as opposed to commonplace development there can be new and non-obviousness innovations in software which are rightly called inventions. Submitters are viewing software as simply “programming” and that focus is too narrow. For example, the software engineering that gave rise to operating systems which allowed multi-tasking and multi-users was a classic instance of a patentable invention in its day.

9. Harmonisation should follow European rather than US law. The present Bill is an attempt to harmonise with a mixture of European and Australian patent law. It owes nothing to US patent law which has always been considered by the international patent community to be an aberration.

10. If one allows software patents, one is opening the door to patenting all thought. Well, New Zealand has done so since 1995 and thoughts are still unpatentable. It is a fundamental component of all patent law that scientific theories and mathematical algorithms are not patentable subject matter. And that includes the USA – see the Gottschalk v Benson case referred to earlier.

11. Software Patents will encourage more “Patent Trolls”. A “patent troll” is an entity which owns patents, but does not manufacture products. It seeks the derive income by “trolling” the business place for infringers of its patents. One of the most famous trolls is NTP, a very small company that successfully sued Research in Motion, the manufacturers of the BLACKBERRY handheld email device. Based on New Zealand experience with software patents since 1995, there may be a few more opportunities for trolls, but the impact to date has been insignificant. In any event one cannot be too critical of trolls – the Crown Research Institutes are potential trolls (they patent but make nothing) as evidenced by the Australian CSIRO who is suing many US companies for infringement of its Wifi patents. CSIRO has so far won or settled with 14 companies for $200 million.

12. Large entities register patents for software patents many years after the original work simply because no one else has. Patents for inventions of any sort can only be validly obtained if they are new and possess inventive step (non-obvious) at the date they are applied for. That is not the case years after the “original work”. As in other countries the Bill requires IPONZ to examine all patent applications for both novelty and inventive step.

Economics Justification

The rationale for the patent system as a whole is an economic one. It is to provide an incentive for innovation in all technologies and in particular an incentive for the investment required to commercialise such innovation.

This is needed because “The competition of the competitive model�almost completely precludes technical development” [14]. In a pure completion market, upon marketing a new invention competitors will quickly imitate and the fruits of the innovation will be spread over the entire market. The imitators, without paying any costs for development, profit along with the innovator.

A patent system of some sort has been in existence since 1474 (Venetian Republic) and in its modern form since 1852 (Great Britain). Between 1851 and 1872 two Royal Commissions in Great Britain investigated whether the patent system should be kept. Isambard Brunel supported abolition while Lord Kelvin did not. The retentionists won. It seems the Netherlands is the only country which has abolished its patent system (1869) only to re-introduce it in 1912.

There have been no recent large scale objective economic analyses of the patent system. In the 1970s, when such studies were undertaken sometimes at the behest of governments, the usual conclusion was that the arguments for and against a patent system was equally balanced. On that basis, governments always opted to maintain the status quo, that is to keep the patent system.

If a particular technological sector such as the software industry was able to demonstrate conclusively by economic studies that the net effect of the patent system in relation to that technology was economically detrimental then that would be a powerful argument for it to be excluded from patentability. However, no objective and detailed study in relation to software seems to have been undertaken and certainty no hard economic evidence has been presented to the Select Committee by submitters on the Patents Bill. More specifically no evidence has been given of any decline in the software business in New Zealand since 1995, nor any reduction in investment in this industry since software patents have been granted in New Zealand.

The position of software for embedded systems has hardly been recognised by submitters who argued for software inventions to be excluded from patentability. A controller for even a washing machine, with 16,000 lines of code will take 18 months to design and implement and will require considerable R&D capital. That capital is more likely to be found if there are patents in place to mitigate the risk of immitators causing prices to drop and minimising a return on investment. In fact venture capitalists will require patents or at least patent applications before giving funding.

As previously explained the real value of a new embedded system is likely to be determined by its software. Excluding all software inventions from patentability would mean funding for embedded systems would dry up and cause damage to New Zealand’s high tech export economy. It is noted for example that the Auckland Bioengineering Institute, while making its modelling software open source, believes patents are a necessary tool in successfully spinning off companies to commercially exploit narrow sector tools [15].

If and when a proper economic analysis is undertaken the chances are that the outcome may be different for each category of software invention.

Footnotes

(1) See Patents Act 1953, sections 13, 21 & 41, Patents Bill 2008 clauses 13(b), 60(i)(b), 106(i)(a).

(2) (2007) 78 IPR 653.

(3) [2006] FCAFC 120.

(4) On appeal from 543 F.3d 943 (Fed.Cir. 2008)

(5) 409 US63.

(6) 450 US 175.

(7) [1966] RPC 194.

(8) [1974] RPC 147.

(9) Art. 52(3)

(10) (1992) 22 IPR 417