FSF joins forces to fight software patents in U.S. Supreme Court | Hacker News Reader

FSF joins forces to fight software patents in U.S. Supreme Court

fsf.org

182 pointsconductor27 comments

What would happen if software were no longer patentable? Finance may offer us a guide. Financial innovations are not patentable [1]. This has two effects:

(1) Good ideas drawing on common knowledge and resources catch on fast. This compresses the time between invention and commoditisation. Those who rapidly implement ideas, whether their own or others', are rewarded. Scale and agility gain currency. Downside: the blind copying of competitors' ideas.

(2) Good ideas drawing on scarce knowledge or resources are protected fiercely. This comes in the form of secrecy - think hedge funds. It also comes as increased pay for producers. In the late 1970s Salomon Brothers helped invent the mortgage-backed security, giving them a de facto Wall Street monopoly. This made everyone involved on the mortgage desk incredibly valuable. Salomon did not pay them accordingly and so lost its storied monopoly.

[1] http://blogs.reuters.com/felix-salmon/2010/11/19/can-you-pat...

If one has the secret formula for Wall Street monopoly, would it make most economic sense to let the world know about it in a public patent?

I think this is a great analogy.

A hypothesis: If financial innovations were patentable we'd have a lot more transparency on the way the industry operates. Assuming non-frivolous patents, it's hard to complain about being unable to copy a sophisticated financial technique, because of the industries opacity.

Good parallel. One of the points for having patents was to reduce the amount of secrecy in technology IP rights allow companies to share knowledge without fear of competing against their invention. This is covered by a form of IP: trade secrets.

This is a legitimate argument for patents in several fields, and for allowing patents for software parts of machines (users of such machinery are better off as a group if they can find out about its workings). Perhaps users of securities are put at risk by the trade secrecy governing securities innovation and so financial patents should be allowed: certainly, failure to understand risks associated with financial instruments was a large part of the problem in the 2007/8 financial crisis.

For software it is not so useful: patented software typically is triple protected: by the patent, by copyright, and by secrecy, since the patent applies to ideas in the source code that are not visible in the deployed binary. So software patents do not promote transparency.

In anticipation of a slew of "this is so obvious" comments that quote from the brief, I urge you to remember that there's an opposing side to this argument: no matter how compelling or obvious this brief's arguments may seem on their own, the court will be weighing them against the arguments of the opposing side.

Without seeing those arguments, we can't really make any substantive judgements about this brief.

All the briefs, just for you: http://www.scotusblog.com/case-files/cases/alice-corporation...

The opposing argument is "we don't want to lose our ability to crush new innovations and keep lawyers rich".

Yeah, take a look at this pack of horrible deceptions stopping just short of lies that the BSA/Microsoft has submitted to the court:

http://www.bsa.org/advocacy/~/media/Files/Policy/Patents/BSA...

I have read the brief. It basically says to constantly apply the "machine or transformation" test to a patent to determine if it's standalone software or something substantive. It also mentions how some software doesn't rely on a specific type of machine to execute.

How is the PTO expected to be able to determine the difference between a "standalone software" patent from "software as part of a machine/transformation" patent? And if they can't and the law is only applied in retrospect (e.g. patent validity challenges), how does this abstract "machine or transformation" concept prevent litigation which seems primarily driven by intimidation not expectation of victory.

The PTO would probably revise the Manual of Patent Examining Procedure to include any new direction from the Supreme Court. See the current section on patent subject matter eligibility:

http://www.uspto.gov/web/offices/pac/mpep/s2106.html

Of course. But without a bright-line test that can easily be applied by the corps of non-lawyer patent examiners, what's the point? And does the Supreme Court ever really give us a bright-line test?

Some of this is verging on arguing how many angels can dance on the head of a pin. Trying to draw lines within the framework if section 101, despite the general inadequacy of that framework. The flaw of section 101 continues to be that the law is decoupled from the economic rationale for patent. To whit: patents exist to protect expensive capital investment in research from free riding, but judges are not allowed to consider the scope of that capital investment in determining patentability. The law continues to adhere to that outdated notion that innovation can happen in a "flash of insight."

I thought the Amicus from Schultz, Love, Bessen, and Meurer was a lot better than the FSF brief. It explains the issues much better and suggests more constructive conclusions that would lead us a lot closer to outright abolition of software patents.

http://marginalrevolution.com/marginalrevolution/2014/03/sof...

Software patents in general aren't game here. This is about how courts should draw the line between applications of principles, which are patentable, and the physical principles themselves, which are not. This has particular relevance to software patents, because it is in this area that there is the greatest charge that people are trying to patent general principles.

Can someone explain to me exactly what it would mean if software ideas were no longer patentable?

Ask 10 people this question and you'll get 11 different answers, but IMO the practical impact will be the lawyers will make less money and there will be less multi-billion dollar lawsuits, but the industry will keep trucking on as always.

The idea of software being patentable is actually a very new concept. Though there are some rare examples of software systems being patented back in the 1980s, it wasn't until the late 1990s that software patentablity was a widely accepted concept, prior to then software was seen as a mathematical construct and thus explicitly excluded from patents. The industry got along fine back then on copyright laws alone and will get along fine with or without patents in the future.

Worrying about the giant lawsuits is the wrong angle.

Any time MegaCorp 1 accuses MegaCorp 2 of stealing its stuff, it's going to be expensive to resolve. Even if you draw the line at "copyright, not patent" MegaCorp 2 is just going to push the line that much further in how closely they can copy MegaCorp 1.

It is way easier to avoid committing a copyright violation than a patent one because the USPTO has let way too many fundamental and obvious ideas be patented. As a developer, if I write 100 lines of code I have probably violated at least one active patent and probably more like 3, completely unknowingly.

I agree with both of you. On one hand, when you have a personal pissing match like Apple v. Samsung, if it wasn't about patents it'd be about something else. On the other hand, requiring a showing of willful violation would cut down on patent suits against smaller companies.

Also, a larger part of the Apple v. Samsung cases was about design patents rather than software patents.

It's so easy to accidentally violate a software patent. It's a two-character difference in this case:

http://graphics.stanford.edu/~seander/bithacks.html#IntegerA...

It would mean that IT innovation won't be threatened by anticompetitive forces which want to retain the market to themselves or trolls and racketeers who simply want to rip off some money without actually producing anything.

A good history of software patenting:

http://www.washingtonpost.com/blogs/the-switch/wp/2014/02/26...

For example, let's look at the heavily patented field of video codecs. Right now, anyone can read the H.264 specification, you just have to pay potentially exorbitant fees to the patent holders to implement it. Without patents, there's no monetary incentive to contribute to open standards, and the industry would move to either specifications under NDA or binary-only codecs.

Or so the pro-software patent argument goes. In practice, there have been lots of undocumented binary formats parts of which are patented. The old .doc/.xls/.ppt formats had that status for years.

I found this passage from WildUtah's link very helpful in understanding the subtle differences here:

"But crucially, copyright protection allows someone to independently develop software to achieve the "same overall result" as a copyrighted program. In contrast, patent law doesn't allow independent invention limiting the opportunities of future innovators and creating the risk of accidental infringement and wasteful litigation."

Personally, I hope at least that they declare that math is not patentable even if executed on a computer.

What is software if not math executed on a computer?

Reply on news.ycombinator.com