Slashdot Mirror
Paul Graham on Patents
volts writes "The always interesting Paul Graham has a new essay, 'Are Software Patents Evil?'. "A few weeks ago I found to my surprise that I'd been granted four patents. This was all the more surprising because I'd only applied for three...""
i'd prefer to say that they are evil in their current form as they are implemented in the u.s. patent system.
if software patents would apply only to commercial software (e.g. gnu staff could always use patented stuff as long as they give away the product for free, and microsoft would have to pay nice $$$ for the same thing), the patents would be better.
but i think there's a higher chance of bush turning into a human being than this happening.
I'd tell you the chances of this story being a dupe, but you wouldn't like it.
How about this for a proposed solution:
1- You can't patent anything until it's been published in a peer-reviewed journal.
2- The duration of the patent is proportional to the clout of the journal.
Only novel ideas will get published in a scientific journal with any clout. And journals without a big name will only get you a patent for a few days. This way the drug companies can patent the fruits of their multi-million dollar investments. But when someone tries to patent something stupid like one-click shopping, they'll be laughed out of all the journals. So who will evaluate the clout of a journal? College professors do that already in order to establish who gets tenure and promotions, etc. Plus this would keep the decisions out of the patent office, giving them an opportunity to offload their problem of being overwhelmed with applications.
Of course the real solution is to abolish software patents, but that's a rather tough sell when lots of rich companies are lobbying against it, so how about seeking something that might actually succeed?
Graham has engineers disease: he believes that being an accomplished engineer makes him qualified to speak authoritatively on art, law, science, film... He probably caught it from ESR.
Anyway, I heartily recommend you read this fine demolition of Graham's opinions on painting before giving this dilletante blowhard any of your copious free time.
Athletic Scholarships to universities make as much sense as academic scholarships to sports teams.
First he says...
Unfortunately, patent law is inconsistent on this point. Patent law in most countries says that algorithms aren't patentable. This rule is left over from a time when "algorithm" meant something like the Sieve of Eratosthenes.
and then he says...
In 1800, people could not see as readily as we can that a great many patents on mechanical objects were really patents on the algorithms they embodied.
and using this argument he says that if you don't like software patents you don't like patents in general.
I think there is a big difference between the method of a process, and the mechanical objects that implement them. I have no problem with people patenting mechanical objects (that are non-obvious). I do have a problem with people patenting methods that would work with those objects. If he thinks that algorithms implemented in mechanical objects is the same thing as algorithms, he isn't much of a lawyer.
I believe Paul Graham might have been hanging out with Dvorak's crack pipe or something. Software patents are not like patents on mechanical inventions at all. It's the difference between patenting knowledge, and the application of knowledge. Software is not the application of knowledge, but knowledge itself. Knowledge is special. Knowledge can be freely shared, and should be. If it is not, we all suffer. Those high priests who think they have special title to knowledge and do not deign to bless the uninitiated help no-one but themselves.
... The solution is patenting as much as we can. A future startup with no patents of its own will be forced to pay whatever price the giants choose to impose. That price might be high. Established companies have an interest in excluding future competitors.
No intellectual endeavor procedes without building upon previous knowledge. This is particularly made manefest in the field of software, where there are untold constituent pieces of a system of any complexity. When we require the authors of such systems to seek out and license every brain fart anyone might have had that impinges on their work, progress becomes impossible.
If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today.
--Bill Gates
If we are to have a true and honest culture, we must be aware of the self-appointed tyrant who puts a fence around a painting or art or music or literature and shouts 'This is my preserve. Think as I do or keep out.'
--Walt Disney
Actually, there is a good distinction you can draw between software patents and conventional patents that is strong enough that you can't automatically infer that being against software patents means you are against all patents: Software patents are the only things I know of where the patented objects are also covered under copyright law.
I go more into depth about this elsewhere, but the short answer is that we shouldn't be surprised that patents, balanced for one type of use, and copyrights, balanced for another, make no sense when both are covering the same thing, since they were never designed to do that.
Obviousness is a real problem too, of course, but that's more a practical problem, one that could be corrected by more aggressive denials by the PTO. This is a fundamental conflict.
Reduce the length of the patents for software/business models to 2 years. That is more than adequate time to do something with it and doesn't lock things up. Patent trolls wouldn't have much time to work either.
"Boycotting Microsoft makes about as much sense as... I can't think of a witty comparison."
I agree the users can't boycott Microsoft, but I think the point that Engineers would start to leave is valid. The people I know who work for MS don't believe it is an evil company, a company with problems yes, but all large companies have that. If they did believe it had gone too far I don't believe they would stop any longer than the time it took to find a new job.
"The weirdest thing about a mind, is that every answer that you find, is the basis of a brand new cliche" -
Really? I think it fits perfectly. They haven't started suing yet because they are only on the brink of the downhill descent, where the talking happens. If Vista doesn't turn MS around, and they start to take a real slide in the sales and profts departments watch for them to follow this exact model.
If you think imaginary property and real property are the same, when does your house become public domain?
Potentially, also, the patent would be kept secret until approved. This would have the nice side effect of, if a competitor starts producing a similar widget in the time between when you applied for the patent and received it, it should be tossed on the grounds of being obvious--since the competitor clearly came up with the idea on their own.
No comment.
from my blog
I broadly agree with Paul Graham's essay on Software Patents, but I do think he underestimates the damage from patent trolls, and from what he calls the mafia-like behaviour of some patent holders.
Paul has been lucky in the field he has worked in, but in the Audio and Video area there are many patent thickets. Perhaps it is the history of Farnsworth's victory over RCA that makes video engineers patent hungry.
My first startup, The MultiMedia Corporation, was a spin-out from the BBC in 1990. One of our products was a program called MediaMaker that combined video from tape or videodisc, CD Audio, Pictures, digitised audio and Director animations into picture icons on a timeline for making presentations. It was demoed on stage at Macworld by the CEO of Apple, and we got Macromind to publish it.
Then the patent troll showed up. A company called Montage had made a video editing system that included several video monitors showing edit points from tape. The company had gone out of business but a lawyer had bought up the patents, including one on using a still image to represent a video sequence. The troll was working his way round the video companies, and he caused enough trouble to stop work on the product while we worked on a legal defence instead.
Later, while I was at Apple on QuickTime, there was a steady stream of patent trolls claiming that Apple should pay them royalties; enough to keep several lawyers busy, and a lot of engineers spending time working on prior art evidence demonstrations.
Several potential features were excluded from QuickTime due to patent thickets. The obvious one was the Unisys LZW patent that encumbered GIF, but there were other more subtle pressures that meant adopting open source codecs was discouraged. Working on the patent license agreements for MPEG meant that technology ready to ship was deferred pending legal agreement on more than one occasion.
So I'm much lass sanguine than Paul about this. I think software patents should not be granted, and the European Union's banning of them is the right decision. I hope the Gowers Review in the UK makes this UK law as well.
For certain industries you probably do need a patent to recoup your development costs. The drug industry certainly springs to mind.
I feel there may need to be a variation in patent length based on industry. 25 years is a lifetime in the high tech industry, but if patents were valid for say 4 years and only granted to non-obvious ideas then they could help the industry. As it happens right now, many technologies are obsolete before their patents are granted.
In August 2002, Paul wrote and published the article "A Plan For Spam". On December 13 2002, Networks Associates applied for spam filtering software patent that includes "Bayes rules". From the patent (#6,732,157):
"wherein the utilization of the Bayes rules further includes identifying a probability associated with each of the words; wherein the probability associated with each of the words is identified using a Bayes rules database; wherein the electronic mail messages are filtered as being unwanted based on a comparison involving the probability and a Bayes rules threshold; wherein the threshold is user-defined."
Maybe I'm just not wearing my tin hat today but I believe someone at Network Associates read "A Plan For Spam" and applied for a patent on it, every though it was not an idea created by them. That is sickening.