Slashdot Mirror
Debian, SFLC Publish Patent Advice For Community Distros
An anonymous reader writes "The Debian Project is pleased to announce the availability of the Community Distribution Patent Policy FAQ, a document meant to educate Free Software developers, and especially distribution editors, about software patent risks. The FAQ has been prepared by lawyers at Software Freedom Law Center (SFLC) at the request of and with input from the Debian Project. While the document does not constitute legal advice, it provides insights on dealing with software patents, which might be applicable to other community-driven Free Software distributions. The Debian Project maintains a critical stance towards software patents: we consider software patents a threat to Free Software and we believe they provide no advantages in promoting software innovation."
The basic idea behind the patent system is sound. There's no economic incentive for individuals and small to medium-sized businesses to invent things when a big company can just take the idea and easily outcompete due to greater resources. And without the patent system, there's no incentive to release inventions into the public domain rather than try to protect them as trade secrets.
This applies just as much to software as to physical objects. Suppose I came up with a method to dramatically increase a car's gas mileage. What's the difference if the method is a change in the physical structure of the engine or an improved algorithm in the car's software? The same logic applies: if my method is not protectable by patent law, I lack economic incentive to put the necessary time and effort into developing the invention.
I understand (and agree with) arguments that the patent time should be less for software, that the thresholds for patentability and enforcement are far too low, and that the whole system in general is being abused and needs major changes.
But I have yet to see a rational argument for why physical inventions but not virtual inventions should be patentable.
It's generally a good document answering basic questions on software patents, but the answer to the first question makes me cringe: "What is a patent? A patent is a state-granted monopoly ..."
A patent is not a monopoly. A monopoly, implies that a person have market power: that they can set the price of a product above what it would be in the free market. However, a patent does not grant such power. The vast majority of patents are for useless inventions and would not be able to demand any market power. But even useful patents don't necessarily confer monopolies. I doubt there is a single software patent that confers a "monopoly" on the patent holder. Any infringing software can be modified to work around just about any software patent. For example, if you have a patent on quicksort, I'll use merge-sort. If you have a patent on binary search, I'll write a ternary search algorithm. Patents do not confer monopoly power, they only confer the right to exclude others from infringing on the patent.
Just a bit of nit-picking in a generally good document.
Fast Federal Court and I.T.C. updates
Only the US appears to have really bad software patients and if you live outside the US EU and Japan chances are no one bothered to patent it.
I also find it funny that source code is considered a description of the patient. Does that mean for the likes of x264 you would have to sue each individual complier for the infringement (yes I know they distribute the complied library). Also does that mean that all Gentoo users are terrorists/communists?
I found their discussion around the difference between source and binary distribution interesting. I'd be curious as to whether you could even come up with a truly robust definition for those terms.
Is the distinction being human-readable? Well, if I have an opcode table binary executables are at least as readable to me as Chinese is.
Is the distinction being directly executable? Well, what about Java and any language that compiles to bytecode? What about a theoretical microprocessor that implements in hardware a C interpreter?
Machine code and C are just two languages - they are a bunch of symbols arranged in some syntax that coveys a series of steps to be executed.
I could see how you could distinguish between writing code and executing code, just as you can distinguish between talking about robbing somebody and actually doing it.
If it's not legal advice, then I would have to be some kind of idiot asshole to follow it, as patents are a legal fabrication.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Richard Stallman's lecture on patents is worth listening to get an understanding of the real issue and some history on how some software patent cases have played out. http://audio-video.gnu.org/audio/#2011 Scroll to 'The danger of software patents.'
I would think that free speech only applies to what humans say not the output of a complier and the output changes depending on settings. While you can imply or auto complete some things this is the actual work of a human. If someone writes a program in machine code so that it can be read then it would be source code as well.
Is the distinction being human-readable? Well, if I have an opcode table binary executables are at least as readable to me as Chinese is.
Stop trolling, if you did not speak English this would also be unreadable to you.
No, they are a threat to our very way of life and should be stopped.
Copyright, sure, patent, no.
---- Booth was a patriot ----
I don't think the Debian Project are the has the best ideas when it comes to software patents, seeing as their policy is 'assume the position for Uncle Sam'. We should instead follow the example of Canonical: 'incorporate somewhere under control of a more friendly government and waggle tongue'.
Machine code and C are just two languages - they are a bunch of symbols arranged in some syntax that coveys a series of steps to be executed.
This is the reason software shouldn't be allowed to be patented in the first place. It doesn't matter whether you write a linear algebra system or an iPhone fart program, in the end it's all math.
Their explanation is probably incomplete, as it could easily be read as meaning that patents themselves induce their own infringement. That would mean anyone could sue a patent holder for having the patent, and I suspect this is not the case in reality.
OTOH if it is the case, then let's all get a-suing. ;-)
Stop trolling, if you did not speak English this would also be unreadable to you.
Woosh!
Of course, but I do speak English, so I picked Chinese as a random example of a language I don't understand. And you're just proving my point - binary code is no less human readable than any language that you don't happen to know.
this learned study clearly shows why there should be no software patents.
http://www.bu.edu/law/faculty/scholarship/workingpapers/2011.html
Speech is communication that someone intends to use to convey information to another person (probably not an accurate definition) but if it is actually understandable by anyone or actually has any information is irrelevant. Legally speaking and by my definition you probably don't speak to your computer its just IO, its can become seeking if your communication goes to someone.
What is the relevance to your argument that someone cant understand something because they have no knowledge of the language? Its quite obvious and irrelevant. You were trying to make it support your argument buy picking a popular /. ethnic/racial marginalisation (that might be a little over the top and you may not have intended to). Is the sole inverter of a (programming) language not entitled to free speech because no one yet understands it (some lobby group convinces a country not to allow a new language to be developed)?
My whole point is that there is fundamentally no real fundamental difference between source code and binary code in the first place, so it seems a bit silly to treat them differently from a legal perspective. The difference at best comes down to intended use. If I email you a binary and tell you not to run it, does that make it non-infringing? If I email you source code and tell you to compile/run it, does that make it infringing? Where do you draw the line? I suspect it would come down to communication of intent, and if that is what we want to look at then why bother to distinguish between source and binary as it is completely orthogonal.
I was anticipating the argument that the difference between source and binary is that the former is human-readable. However, there is nothing that really makes source more readable than binary, other than the fact that it is usually written in languages more similar to European languages than the x86 instruction set or whatever. In fact, if you educated somebody in China WITHOUT any exposure to any European language (including programming languages that contain European words), and taught them to a graduate-level in Computer Science, they might find it easier to read binary code in a hex editor or dis-assembler output than in C.
And, for the record, I have spent time trying to learn Kanji and Hiragana, and the former is barely different from Chinese, having traveled to Japan several times. I am shocked as the next guy when people utter racial slurs, and I don't really consider picking a random language that not many who read Slashdot are likely to know as an example a slur. I could have picked French, but the fact is that French is fairly comprehensible to most of those around here. Maybe next time I'll pick Cuneiform, if that isn't considered a slur on archaeologists the world over.
Im still not sure your not trolling address my actual points.
You have to look at it from the other end speech its not defined by if its reverse engineer-able or not, you dont read binary code left to right or another direction you jump all over the place and re read sections. Speech is determined by being it spoken/written not by the recepient understanding. Computers are in no way legal entity to be compared to a human and though a hex editor gives a lot of context programs are still just a big number (this has its own arguments associated with it) that was worked out by a computer. If you want to hand write binary instructions (for anything large this is terribly inefficient to create and the number of people who could make the code run significantly faster than C without unreasonable bugs would possible in the single digits, its not practical to code say VLC in binary) that you want to show to someone else I would guess you could test it court.
I did not call it a slur, i don't think it is and you were not intending to, you compared binary to Chinese and said they were the same where it counts for the argument. The main reason i called you a troll is you put and example in to help prove your point that does nothing of the sort but may confuse people.
So, how is source code "free speech" but binary isn't? You can read either out loud. Either is understandable to somebody trained to understand it - as with any language. Either can be processed directly by a microchip as well...