Alex's Kwiki

The Patent Office have written a glossy leaflet talking nonsense about "Computer-Implemented Inventions" aka CCIs aka Software Patents.

Thoughts:

Why do we need a Directive?

What a load of reactionary crap. Yes, we clearly could use a directive because we are drifting towards the US. The people pushing us in that direction are certain nations, though - the Parliament has agreed a text, the Council hasn't officially set out it's changes yet and is coming under pressure because it's measures are so pro-patent. The Parliament developed a perfectly workable directive; however, this leaflet does not mention it because they want to drift towards the US situation.

(What will it change - they keep arguing that they need new legislation to re-enforce the (legislated) status quo. Is that a sustainable argument?)

Why isn't copyright enough?

Leaflet says that copyright isn't good enough because "independently developed solutions" aren't covered. This is good though! Development of software is a low-cost effort and intrinsically non-inventive; blocking independent solutions is a bad thing.

Proposed directive - fact and fiction

It will make clear what is patentable

Rubbish. The "technical contribution" loophole is large enough to drive a truck through. The patentability of software will become a matter of draftsmanship, not "technicality". All software has a "technical effect": all software runs in the real world and interacts with the real world. To allow patentability on any software that has a physical effect means that all software is patentable.

The idea of "technical contribution" is woolly and unclear, and will only lead to further confusion about what can be patented (there is no definition of "technical", nor no easily-applied test). It's likely this would require further legislation - we would be in the same situation as the USA post Diamond v. Diehr which lead them to the situation they find themselves in today (the logical fallacy that there is such a thing as "technical software").

It would ban "business method" patents

Nope. They still rely on "technical contribution" - so long as you can find a technical contribution, you can patent the idea. Finding a technical contribution will probably be quite easy.

Open Source Software will survive

Yes, it will survive. No, it will not prosper, and it will be further stunted the more software patents are available. Already with free software you cannot:

• legally view any DVD you have purchased;
• legally convert CDs into mp3s for use with your mp3 player;

(similar problems to the above exist for other multimedia). Similar attempts have been made to patent: file systems used on digital cameras, plug-ins for web browsers, etc. etc. These are all things which affect home users; developers have an even worse time.

It would allow patents for the first time

It is true that Europe has been slowly drifting towards the US situation; this has been done via the backdoor in an attempt to make people think that it is inevitable. Patenting of software has increased vastly in recent times though, and promises to grow even further after this legislation has passed (which would surely encourage more patenting)

It would be harmful against SMEs

Yes. It does not protect them from big business. At best, and SME might be able to use their patent to enter into a cross-licensing system. There will always be a small number of people who will benefit from patents, so there will always be examples of people who have patents and benefitted from them. The overall effect, though, is largely negative, and mostly affects SMEs.

Europe would follow the US lead

Yes. We're just taking longer to get there. The position that "technical contributions" exist in software is completely fallacious; no such division between "technical" and "non-technical" software can exist. This puts us in a position that the clause can successfully be argued against in the future as meaningless, and then removed. This happened in the US.

Maths, symphonies, etc.

All software is a base a set of algorithms. Once you allow one set of algorithms to be patented (software), what logical argument can you use to prevent other classes of algorithm from being patented? At best, you can try to maintain the logical inconsistency in law, which would result in courts giving inconsistent decisions based on the unstable law. At worst, you have to admit defeat and we end up where the US is.

Case study aside

Radioscape appear to have ~60 patents, covering combinations of hardware and software. At least some of them (WO02067432 - a Viterbi decoder that has been speeded up by using lookup tables - blimey!!) look like pure mathematical patents. Are they really a good case study? Hmmm...

Examples of patents refused

Christ, you mean they actually refuse some patents?! Hmm, actually, they seem to invite you to look at their "ex parte" decision list. This decisiob is horrible:

http://www.patent.gov.uk/patent/legal/summaries/2004/o29204.htm

Maybe we ought to generate a list of things that they have allowed - maybe that would be a way to bring the Radioscape case study under the microscope... :). I'm sure there are plenty of patents which we can list. I did go through their complete ex parte case list 2003/2004 to look for patents that were refused then subsequently allowed (I thought it would be an easy way to fish for stuff close to the wire) and didn't find much, but I'm guessing that a lot of these would be European patents anyway?