Software should be copyrighted, but not patented.
I wonder what would be if a writer could patent story lines: "It is like a like a novelist, patenting all action thrillers, that involve a guy saving a girl, who is kidnaped by the bad guys.
Watch out, any other books that have a bad guy kidnaping the girl, can be infringing this patent."
There are many industries that flourish without the needs of patents. How did we end up in this mess?
So physical stuff like valves can be patented, but not software? That seems arbitrary. After all, it's usually the design of the valve that contains the invention. What difference does it make if you print one out? If you mean "nothing should be patented," just say that. One could make a case for that argument. But there's nothing special about software.
How did we end up in this mess?
What evidence do you have that we're in a mess? Have you ever been prevented from doing something you wanted to because of patents?
The fact is, for all the noise people make about them, software patents just don't matter much. Large companies accumulate patent portfolios to protect themselves against patent suits. Startups file for them because investors and acquirers are impressed by them. But they don't decide which startups win.
I've had to take software out of production and replace it with non-infringing software.
What evidence do you have that we're in a mess?
EOLAS v Microsoft, Amazon v Barnes&Noble, Sun v NetApp..
If you don't mind my asking, where were you working when you had to replace this software?
As far as my own situation, the details are here: http://news.zdnet.co.uk/software/0,1000000121,39194883,00.ht...
Ironically it was another "ARC" patent: http://en.wikipedia.org/wiki/Adaptive_Replacement_Cache
From the article, it doesn't sound like you were required by an injunction to remove any code. Someone just read in the news about some newly issued patent and panicked. What's the plan for dealing with the 1000 other patents PostgreSQL is violating?
The PostgreSQL project tries to stick with things for which there is plenty of prior art and thankfully most of the best stuff in databases is old enough so that patents aren't a big issue.
Unfortunately some of the things people do with PostgreSQL's extensibility aren't as old. I'm sure there are some people out there using custom UDF/UDA functions that may infringe on patents.
1. Every year at CMU's Human Computer Interaction Institute, it appears someone reinvents something patented though not deployed by Xerox PARC.
An example is a "looking glass" feature for something like photoshop, where you can preview the result of an action like a filter -- or something like that.
From the anecdotes I've heard, PARC just sits on it, doesn't deploy or license it, and acts to block others from using it.
2. SIFT is a new object recognition technique built upon years of research. A recent patent on the best implementation has restricted groups from using the techniques. They grant exclusive licenses on the level of "for all robots in America", but that's pretty useless for a startup, isn't it?
I'll say it: nothing should be patented. Trade secrets, in the literal and legal sense, make much more sense. If a system can be reverse engineered and copied, you don't have a secret, and it isn't in the best interest of society to block copy cats from beating the originator.
The idea that makers wouldn't create due to the fear of competition is just wrong -- though it is the most often used argument to defend patents.
[edit: And another thing: people respect and fear patents, so despite my opinions, I'm going to act in my best interest to acquire any patents possible on my work. My goal is notably against the best interest of society as a whole: create unfair advantages for my company.]
The rate of innovation constantly increases, and for patents to keep working as they should, the term should get shorter and shorter. Even if the system is made to be "fair" in this way, eventually the term will be made so short as to not be worth the effort of filing the patent. Maybe we are past that point already.
Seems that something that doesn't matter shouldn't be legislated. Do patents matter in other industries? If patents drive innovation in some industries, but have no effect on others besides putting money into lawyers' pockets (and out of taxpayers', entrepreneurs', and corporations' pockets), that would be a case against software patents but not all patents.
Btw, your writer analogy is a poor one, since prior art exists for that story line and thus it would not be patentable. There are also industries that would collapse without patents (pharmaceuticals for one).
The sad truth is, I really believe this would what would've happened!
The one example I've found is, embarrassingly enough, Yahoo, which filed a patent suit against a gaming startup called Xfire in 2005. Xfire doesn't seem to be a very big deal, and it's hard to say why Yahoo felt threatened. Xfire's VP of engineering had worked at Yahoo on similar stuff-- in fact, he was listed as an inventor on the patent Yahoo sued over-- so perhaps there was something personal about it. My guess is that someone at Yahoo goofed. At any rate they didn't pursue the suit very vigorously.
I can't help but wonder if pg thinks the same way about software patents now that there is a very real chance that his own patent could be used against him at the whim of someone at a company he's been publically bad-mouthing for some time.
If someone comes to YC with an idea you like but is incumbered by a technology and/or business model that you know infringes on specific patents held by a company like Microsoft, do you just fund them anyways and tell them it doesn't matter?
Are you saying it would have made no difference to Yahoo if Microsoft somehow patented the continuation-server idea Viaweb depends on before you did?
And yes, it would have made no difference to Yahoo. As far as I know, they didn't do any patent searches prior to acquiring us. In fact, I've never heard of patents torpedoing an acquisition. Acquirers (rightly) have a fatalistic attitude to patents.
Curiously enough, Yahoo did get sued by a patent troll over Viaweb. It was after my time, but it was something to do with shopping carts. A totally bogus claim, but I believe Yahoo settled, because they were able to pay them out of the money they'd held in escrow for us. (We were not that unhappy though, because the money held in escrow had increased something like 15x.)
Tivo v Dish - http://www.law.com/jsp/article.jsp?id=1201779834291
Forgent v Audiovox, ePlus v Ariba, Immersion v Sony/Microsoft, EMC v HP, USA Video v Movielink, BroadVision v Art Technology, Electronics for Imaging vs everbody .... http://answers.google.com/answers/threadview?id=492018
Pharmaceutical research in the US 2005 34.8B USD federal funds 41M USD. Thats less than 0.12%.
Pharmaceutical research in all likelyhood includes the cost of bringing drugs to market, currently around $800M per drug. Because of this high barrier, much of the Pharma energy goes into 'me too' drugs of a higher risk/reward ratio, again inflating the true spending on new research: http://freakonomics.blogs.nytimes.com/2008/01/24/what-dont-w...
The me-too phenomena is a problem but only because of two factors, pressure by pharma reps and the fact that the US is the only country I ever heard of that allows advertising of non over-the-counter meds.
Hiring doctors to travel the country, give lectures explaining how it works, renting conference centers, providing lunch, etc adds up. Learning how to use a new drug is not easy like learning software; trial and error doesn't work.
A drug is worthless if your doctor doesn't know what it does.
America is the one of the only countries that allows pharmaceutical companies to advertise directly to consumers. http://tinyurl.com/27y5df
>How did we end up in this mess?
Politics, money and special interest is a dangerous confluence.