European rejection of software patents is a victory for open source [LWN.net]

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 12:50 UTC (Thu) by liljencrantz (subscriber, #28458) [Link]

This theory would hold if someone constructs an algorithm that is so hard to reverse engineer that it would take even a large group of people decades to do it. In such cases, patents could mean that after a limited set of time, everyone can benefit from the breakthrough research, while the author can be safe in knowing that he has a limited time of exclusive ownership of the algorithm. Sounds all nice and reasonable to any layman.

The only issue here is that almost every single software patent is over the obvious solution to a simple problem. There are very few, if any, algorithms suitable for patenting. I've been asking a lot of people for a single reasonable software patent, so far I've found none.

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 13:54 UTC (Thu) by rjw (guest, #10415) [Link]

That is not the only issue. This theory would presuppose that providing an algorithm as software is distinct from providing it in a human language.

Any encoding of software, no matter how obfuscated, is still decipherable. Otherwise it would be unrunnable... any software is an expression just as a work of human language is. It is covered by copyright, and this is sufficient, because there is no difference between it being possible to run and possible to comprehend.

On another note, this reverse engineering argument certainly makes the case for all patents to have shorter terms, in line with the increased speed of reverse engineering in all fields... I think this would actually be a much fairer patent law : "How long would an optimal sized team of reasonable men, experts in the field, take to reverse engineer this method or apparatus?" Unfortunately, this would require competent patent examiners.

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 19:21 UTC (Thu) by AJWM (subscriber, #15888) [Link]

I think you're interpreting the reverse engineering argument totally backwards.

Patents protect what is easy to reverse engineer, not what is hard. For that latter, trade secrets work almost as well, and in some cases better (the specific formula for Coca-Cola, for example).

Now, there's a difference between that which is easy to reverse engineer and that which is obvious to anyone "skilled in the art". Unfortunately, the latter seems to be no longer a bar to patenting. As an algorithmic example, suppose somebody comes up with a method for easily finding the prime factors of very large numbers. That is hardly obvious -- skilled mathematicians and cryptologists have been working on it for years (and who knows, NSA may even have figured it out). However, that algorithm implemented as code might be quite easy to reverse engineer. Should it be patentable? Well, that comes down to whether one thinks any algorithms at all should be. From the difficulty or non-obvious standpoint though, it would certainly qualify.

European rejection of software patents is a victory for open source

Posted Jul 15, 2005 15:08 UTC (Fri) by rjw (guest, #10415) [Link]

In your specific example, no patent system in the world (currently), even the USPTO, would protect it (if they abided by their own rules, a most unlikely prospect) - it would be a clear consequence of a mathematical fact (some pattern to the distribution of primes), and therefore unpatentable as it would be a discovery not an invention. So the algorithm(s) would be obvious once the mathematical fact is known. I wouldn't expect an actual real live patent examiner to reject it, but by the rules it certainly should be.

My jab was really at patent examiners : I can think of no possible patent system for which it is economic to employ examiners to do a proper job.

My point on the reverse engineering argument is : why should every invention be afforded the same length of government enforced monopoly? This is clearly discriminatory against inventions which are harder to reverse engineer. The people who derive the most benefit from the patent system are those who hold a trivial patent legally thought to be valid, because they gain the most additional value that they would not have in a free market.

So yes, I agree - people with easy to reverse engineer, trivial patents are the most protected. But I don't agree that is a good thing economically in any field, software or not.

Where patents advance the useful arts

Posted Jul 15, 2005 16:20 UTC (Fri) by giraffedata (subscriber, #1954) [Link]

I think you're interpreting the reverse engineering argument totally backwards.
Patents protect what is easy to reverse engineer, not what is hard.

No, you've got it backwards. The original, essential purpose of a patent is to protect what is hard to reverse engineer. A patent is a deal the public makes with an inventor wherein the inventor agrees to divulge how his invention works and the public in exchange guarantees him the exclusive right to profit from the invention for a while.

Patents were created in a time when reverse engineering was hard and people protected their inventions by trade secret. Sometimes, they died with those secrets. While alive, they had to be very careful to whom they licensed the secrets. It was bad for everyone.

Note that "patent" is the latin word for "lying open." A patent is primarily about disclosing what you invented, not laying claim to it.

Today,, we see a whole other function of a patent: to protect what is easy to reverse engineer. For this kind of invention, the inventor does not have trade secret as an option -- if he sells the invention, the first buyer will know exactly how it works. But he has another alternative: abandon the idea, and don't bother to develop it into an invention at all.

Drug patents are a good example of this. A drug company can't keep its formula secret, but it doesn't have to develop the drug at all. Developing a drug into an invention costs millions of dollars. The patent is a deal the public makes with the drug company: you create a drug for us, and we will let you profit from it exclusively for a while.

So, in conclusion, the pro-patent theory of which ESR speaks is two-pronged: Where an invention is expensive to develop or hard to reverse engineer, the patent system can bring valuable new technology to the world. Software patents tend not to fall into either category.

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 13:43 UTC (Thu) by rmstar (guest, #3672) [Link]

The theory is that if development costs are huge, but copying an invention is cheap, then patents really protect the innovator's investment and thus are an incentive.

I would say that in software that point is moot, because if the development was expensive, but copying (not in a bit-for-bit digital copying sense, but copying as recreation) is easy, then the inventor is incompetent. Consider sudo, progress bar, one-click, etc.

In biochemistry/medicine the situation is somewhat different, because selling even a simple molecule as a remedy for something includes expensive large scale field studies. However, I have heard about trivial patents and trivial patent based extortion in that field too.

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 17:52 UTC (Thu) by flewellyn (subscriber, #5047) [Link]

I can think of ONE software patent that was actually novel, non-obvious, and a true advancement of the state of the art. That would be RSA.

Obviously, that's expired. And that's one out of...how many hundreds of thousands granted? Not a very good "hit" rate there, USPTO.

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 19:49 UTC (Thu) by vonbrand (subscriber, #4458) [Link]

To be fair, you would have to consider only those patents you have checked yourself. And that would have to be a random (statistically fair) sample.

RSA patent good?

Posted Jul 15, 2005 16:33 UTC (Fri) by giraffedata (subscriber, #1954) [Link]

I'll buy that RSA is novel, non-obvious, and a true advancement of the state of the art, and therefore refreshingly within the letter of patent law, but: Would you say RSA was expensive to invent? Did it take so much work to think of that its inventor wouldn't have bothered if he didn't get a piece of the action every time it's used?

If not, then it still falls outside of what ought to be patentable, in many people's value system, since the patent only serves to transfer wealth to people who have great ideas from people who don't.

European rejection of software patents is a victory for open source

Posted Jul 21, 2005 7:12 UTC (Thu) by Wol (guest, #4433) [Link]

Actually, I *WOULDN'T* buy that ... :-)

It was discovered seven years before it was patented, and it was in (government) use as a trade secret pretty rapidly, again certainly long before it was patented.

Cheers,
Wol

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 20:55 UTC (Thu) by cdmiller (subscriber, #2813) [Link]

Software = mathematical algorithm => software patents = bad idea.

Has anyone patented the idea of the patent system yet? How about the "business method" of patenting something and suing people who try to use it?

Patenting patent suits

Posted Jul 29, 2005 20:11 UTC (Fri) by Max.Hyre (subscriber, #1054) [Link]

How about the "business method" of patenting something and suing people who try to use it?

Never happen: too much prior art. :-(

European rejection of software patents is a victory for open source

Posted Jul 14, 2005 22:33 UTC (Thu) by kleptog (subscriber, #1183) [Link]

Hmm, I thought the purpose of patents was to encourage people to publish trade secrets to advance the progress of science and get a limited monopoly on it in return. It was never about making money.

So if something is not or would not have been considered a trade secret, there appears to be no benefit to society to grant a patent it. By that reasoning business methods would be out since they're hardly something you can keep secret. Whether source code can be considered a trade secret is another question. As long as the patents don't include actual source I don't think you can grant them patents either...