Licensed under: CC BY 3.0 AU
Dear Secretary,
I make the following submission. It primarily focusses on question 11: “Should the excluded subject matter for an innovation patent be amended to include computer software? Why or why not?”
Summary
Computer software should be excluded from the scope of innovation patents (and the patent system generally). More generally, the patent law ought to be modified such that, if it is possible to infringe a claim by running software on a general purpose computer no patent or innovation patent ought to be issued in respect of that claim.
About me
I am a solicitor. I practise in the area of technology and telecommunications law and have done so for about 20 years. In that time I have represented individuals and large and small companies with a broad spectrum of interests from start ups through large companies, and for both customer and vendor organisations. In the overwhelming majority of matters I have been involved in issues relating to copyright, patent and similar laws have been an important component.
Discussion
The US experience is that patents on software inventions have been very problematic. There is currently both legislative and judicial movement there to reduce their scope. Part of the problem is that it is difficult to maintain the quality of software patents with many patents being held invalid at trial (“… software patentees overwhelmingly lose their cases, even with patents that they litigate again and again. Software patentees win only 12.9% of their cases, …”[1]).
An additional problem is that multiple software patents may read against a single software implementation. While it is possible for multiple patents to read against the implementation of a physical article, the issue in relation to software is much more pronounced. By way of example, the MPEG LA list of patents which read against AVC/H.264 patents runs to over 60 pages (and they will not guarantee that this list is complete). [2] Allowing innovation patents in this area will make matters worse.
Patent systems impose compliance costs on all participants in the economy. Getting advice on codec licensing, for example, would be cost prohibitive for a start up. Start ups, rather, simply either avoid the issue or sign up for a licence whether or not it covers their proposed implementation. In one negotiation in relation to a software development project I was told by the other side that they costed the provision of a patent indemnity at 25% of the development cost per jurisdiction in respect of which an indemnity was sought. Therefore, if a patent indemnity was sought in respect of four jurisdictions, the software development cost would double.
These compliance functions do not contribute anything to the economy and kill marginal innovation – even where that innovation is non-infringing (since the infringement will not be known until after the compliance search is completed). Given that patent benefits accrue to only the patentee but compliance costs are multiplied by the whole community which bears them it is difficult to believe that the extension of patentability would result in a net increase in innovation. In this regard I note that I am yet to see a review of the patent system which takes account of the compliance costs it imposes.
There has been repeated and increasing litigation in the US against large companies in respect of software related patents. That large companies are unable to develop in a patent-compliant fashion is evidence that patent compliance is a difficult, if not impossible task. My belief is that it is most common for software patent infringements to arise as a result of independent invention rather than copying. If this is the case, the grant of patents over software serves no purpose in either promoting innovation or the dissemination of such innovation.
Finally, because software is implemented through source code and source code is typically not disclosed inventors of software related inventions already have sufficient protection for their invention through their ability to publicly exercise their invention without disclosing it. Indeed, for most software related inventions there is very little scope for proving secret use.
Notes
[1] Allison, Lemley and Walker, Patent Quality and Settlement Among Repeat Patent Litigants, The Georgetown Law Journal Vol 99:677 at 680.
[2] http://www.mpegla.com/main/programs/avc/Documents/avc-att1.pdf
Brendan Scott's Weblog 