Current UK position on Software Patents

Posted Dec 4, 2004 0:51 UTC (Sat) by wookey (guest, #5501)
In reply to: Linus Torvalds gets software patents wrong, says attorney (Out-Law) by sepreece
Parent article: Linus Torvalds gets software patents wrong, says attorney (Out-Law)

OK, I've been following this closely for the last year or so, so I'll see what I can do.

Software has been patentable in the UK and Europe for 20 years or more

This is fairly contentious. I could write a long spiel about the UK patent law the European Patent COnvention and the case law that has gone on since the 1980s progressively allowing more and more 'softwarey' things to be patented, but instead I'll point you to a couple of URLs which give a reasonable summary.

UK software patent history

And this page from the UK government sums up the current position reasonably well: http://www.intellectual-property.gov.uk/std/faq/patents/computer_software.htm

So EPO (european patent office) and UKPO (UK Patent Office) practice are similar but not identical. In general the UK is more restrictive about what can be patented due to the limitation to technical fields (i.e. image procesing algorithms are technical, but language processing algorithms are not). Most programmers think this is a crazy distinction.

An example of what is at the edge of patentability in the UK at the moment is this patent from ARM limited - which covers the use of copying a pointer rather than the data it points to, when doing CPU simulations (for switching the register bank). This patent went through the special process of arbitration for ones they aren't sure about 'ex parte', but it did get granted. Again, any C programmer is pretty surprised to find that you can patent this highly unoriginal use of pointers in 2002. The point of course is that this is 'only' in the field of CPU simulations, but still - some of us might quite like to write a CPU simulation using this technique.

There is a very cogent explanation of how this demonstrates the empiness of the 'technical contribution' test here: http://www.softwarepatents.co.uk/current/technical-contribution.html

the proposed Directive restricts patentability and grants de-compilation-for-interoperability rights that previous patent law did not

The proposed directive would restrict patentability if it said 'Computer progams are not patentable', or 'data processing is not a field of technology in the sense of patent law', but it doesn't say either of those things. It appears to in places by saying things like 'computer programs are not patenable unless <something longwindeed which turns out always to be true>'. There are several instances of this sort of language which seem pretty clearly intended to confuse anyone not intimately familiar with the debate into saying 'well that sounds OK, I could support that'.

Asking what 'the proposed directive' says is more complicated than you might think - there have been several versions, and the degree to which they restrict patentability, or offer defences, varies enormously. The original Commission draft basically allowed complete patentability. The amended EU parliament(that's elected representatives) version essentially added clauses like the above offering genuine restrictions and and interoperability defence. The current EU Council (that's the collection of government representatives for each country) version took out nearly all the parliament amendments and added a few even more egregious clauses (such as claims on the actual media containing the prgram as well as the program when run, making copying and quoting infringements as well as running) so this allows complete patentability.

I must admit to not being sure if the interoperability clause has survived in any form, but I have read plenty suggesting that it will probably be a very weak defence in practice, depenmding on what the court deem 'interoperability' to mean. We think it means 'being able to play my DVDs, and read/write file formats', but 'they' think it means 'can write a plug-in for a patented program'. i.e. a piece of software which is not talking directly to the patented program is not 'interoperating'.

Hope that helps. If you haven't already read http://www.nosoftwarepatents.com/ , then do so, as that is the clearest site on the subject by a long chalk. It doesn't go into the UK-specific questions too much, but it is the european level that really counts. It is safe to say the the UK is at the 'most pro-software patents' end of the discussion at the moment, and few cracks in that government position have opened up so far, although there are signs that they might be beginning to.