Patent granted for Distribution Testing

barbie on 2005-07-11T10:44:54

There has been a lot of discussion and opposition to software patents in recent times. The danger that concerns many software developers, is that innovation is going to be stifled. While true innovation is something the original author should have the ability to develop further, having a patent that prevents improvement of the idea, or even prevents the research into doing the same thing another way can be inappropriate.

Recently a software patent was brought to my attention, that on the surface might seems a great idea. And it is. EP1170667 - Software Package Verification. However, to patent it is just nonsense, as the technique has been used for many years prior to this being filed. Just about every IT company I have worked at for the last 20 years has used a home rolled software package verification application. And what about CPAN? Hasn't CPAN/CPANPLUS/Test/Test::Harness, etc been doing exactly this for nearly 10 years! Is this what would be termed 'prior art'? I doubt this patent is in any way enforceable, and is purely an attempt to increase SUN's number of patents to make them look good. Any lawsuit is likely to be as embarrassing as the SCO vs Linux debacle. The patent has been granted within a 9 month opposition window, where the grant date was 29th June 2005. Is anyone in a position to challenge them, or will it get through unopposed?

In other news the EU Software Patent Bill has been thrown out. Although this is a good move to stop software patents coming under the control of the EU governing body, it doesn't stop them being applied in individual countries. Hopefully SUN's patent and others will get challenged and those in power get to realise this isn't a good use of the patent system.


Too expensive to oppose

TeeJay on 2005-07-11T13:31:19

The major problem with the UKPO and EPO allowing through dodgy patents, is that it isn't one or two it is tens of thousands.

Baring in mind it is a lot harder for the same piece of code to be prior art than infringing, and the legal costs involved these timebombs just accumulate silently and unchallenged.

Worse still the patent offices refuse to clean up their own mess and expect other people to go to court to do what the patent office should have done in the first place.

Re:

Aristotle on 2005-07-12T00:56:24

HOWTO: invalidate a patent application with prior art

However, once the cat is in the bag, so to speak, ie. once the patent has been granted, it takes about $1,000,000 (yes, you read correctly) to invalidate it in court. Now imagine Sun Microsystems (or IBM, or Microsoft, or Apple, or Google, or…) has hundreds of bogus, invalidable patents like this one. You are a small company and BigCorpOfChoice wants to bring you down. They file a suit for infringement on 10 of their bogus patents. You could invalidate them all – if you spent $1,000,000 per patent in court.

Pop quiz: what is going to happen?

Okay, let’s say you actually braved this onslaught. Now BigCorpOfChoice brings out another 10 bogus patents…

Pop quiz: what is going to happen...

Nothing to worry about.

schwern on 2008-03-21T01:21:10

I know this is coming a little late, but I finally spoke to a patent attorney about this and he concluded that it's very specific to a very particular testing process having to do with test prioritization. It was rejected by the examiners and rewritten several times.

Also it looks like it's a purely defensive patent on the part of Sun and they have a clean track record.

So nothing to worry about.