
It is no secret that the folks at the technology news site
Ars Technica are no fans of patents in general, or of ‘software patents’ in particular. It was
Ars that got
stuck into Australian scientific research body CSIRO last year, because it had the temerity to invent something useful for wireless communications nearly 20 years ago, and then sue those companies that made millions from devices implementing the technology, while stubbornly refusing to enter into meaningful negotiations over licensing.
Last month,
Ars reported the donation of US$500,000 to the
Electronic Frontier Foundation (EFF), and the creation of the ‘Mark Cuban Chair to Eliminate Stupid Patents’, while bemoaning the fact that the EFF is looking ‘only’ to reform the patent system, while ‘[shying] away from calling for the abolition of software patents altogether’.
In our opinion, however, it is to the credit of the EFF that it is at least trying to engage with the issues around ‘software patents’ in a more nuanced manner. However, there is a far greater problem with almost all debate around this issue than whether or not it should be about ‘abolition’ or ‘reform’. The single most fundamental matter that must be resolved before we can have any type of meaningful discussion about ‘software patents’ is for us to agree on what we are talking about.
To put it simply, what is a ‘software patent’?
Not only has the EFF so far failed to address this question, it appears to have completely failed to recognise that it is an issue. To see this, you need look no further than
the EFF’s ‘Defend Innovation’ project site. Starting with the first point of the proposed reforms, we see that ‘a patent
covering software’ should have a shorter term than other patents. OK, this is something we could meaningfully talk about. Many software-based inventions do not require the same duration of protection in order to provide a fair return to their inventors when compared to, say, pharmaceutical products.
But what is a patent ‘covering software’? Does this mean a patent with claims that could encompass a software-based implementation of the invention, even if the patent specification additionally, or alternatively, describes a hardware-based implementation? What if part of the invention is implemented in software, and part in hardware? Should a competing product have some form of defence to infringement after the shortened period if it is implemented in software, but not if it is implemented in hardware? What if the invention
cannot be implemented in software at the time of patenting, but then advances in technology and processing speeds make this practical later in the term of the patent?