Hacker News new | comments | show | ask | jobs | submit login

Honestly, I think software people have a largely skewed view of the patent system in general. I used to be in traditional engineering (mechanical/electrical to be precise), and the amount of R&D dollars that has to go into any major innovation is astronomical.

The amount of testing, regulatory costs, tooling costs, etc etc, makes bringing an innovative thing to market very, very expensive indeed. Most of these things are also trivially easy to reverse engineer - the patent is the only thing preventing a company's competitors from cloning the tech, avoiding all of the costs associated, and discouraging innovation overall.

The trick is software patents - software largely doesn't cost a lot, not compared to big heavy machinery requiring onerous certification and government authorizations. 3 guys hacking for a month straight is peanuts compared to, say, the amount of money it would cost to bring a new fuel injection system to market, yet the software these guys produce can be equally valuable from a marketability perspective. This is where patents fall apart - the point of the patent system is to encourage R&D spending by allowing protection of its results, because being simply first to market is not a large enough advantage in many industries to be profitable. It is, however, true in software, so I'm personally dubious of software patents in general.

Just some perspective - I dislike the "patents are evil, always, every way!" kneejerk reaction. It has a very valid role in a lot of industries that aren't well represented or understood by people on HN.

Here's an interesting Apple hardware patent:

"It is somewhat surprising that Apple received some of the patents in question, such as the patent on "Conserving Power By Reducing Voltage Supplied To An Instruction-Processing Portion Of A Processor". When you peel away the technical language, the patent basically is talking about saving power by supplying less voltage to a circuit and some common strategies to do so. Not only has then been seemingly done before (prior art), but it also is inherently given by laws of nature (power = current * voltage). If that's patentable, the general concept of die shrinks should be patentable, overclocking would be patentable (watch out Anandtech.com!), and a whole host of other processes made possible by laws of nature."


Everything patented or implemented in the history of the world is possible under the laws of physics; the argument that because it is possible, it is not patentable makes no sense.

As must be stated anytime there is an article on here concerning patents, read the claims, not just the title. Most of the titles are so broad as to be meaningless, but the claims must be made specific, as these are what can be defended against infringement.

In this particular case, the claims themselves ( http://www.google.com/patents/about?id=7CCWAAAAEBAJ&dq=7... ) are vague and extremely obvious, and I would be shocked if every claim hadn't already been implemented and/or patented by someone else. In fact, if you read the application, they were forced to retract the first twenty of their original claims.

I know it has become fashionable to say that the concept of patents in general is flawed, and that patent examiners are sub-human imbeciles. I disagree on both counts; software patents are generally crap and should be done away with to a large extent, but as another commenter has pointed out, many inventions require a great deal more investment and patents provide a great incentive to develop such things. It is unreasonable to expect patent examiners (engineers, not lawyers) to do a perfect job of pattern-matching each application with every similar previous application, considering that patent applications are deliberately worded in a language which is impenetrable to anyone trying to figure out what the hell something does. Reform at the policy level is needed, and acting as though the problem is that patent examiners aren't smart enough or that patents in general are a bad idea tends not to be very persuasive.

Which is a software patent.

No one ever said the Patent office was useful, competent, or smart.

Software is also somewhat protected because unlike a physical device reverse engineering is far more difficult. I can open pretty much anything and have a CAD model in less time than the original engineer took to design it.

True, I have learn programming and experience disassembling programs, and is really really difficult to understand how a program work just this way. If the program has millions of lines of code it is nearly impossible.

Patents are a contract, the state gives you monopoly, you give the state the knowledge(with plans, drawings and explanations) of how your thing works. Without showing the code, you should not be granted a patent.

The problem is ... all innovation is moving in the direction of software innovation.

For any given field, it is simply much easier to have a single, general purpose machine into which special instructions for special circumstances are programmed. You can see everything from 3-d metal printers to synthetic genetic computers on the horizon. Software will fuel all of these.

So if any future patent system is going to be a software patent system. And we've seen how evil software patents are.

There isn't an easy middle ground.

Touché sir, I should have narrowed my words a bit, I did not mean to claim that the entire idea of patents is bankrupt, I'm sure there are scenarios where it's helpful.

But as you state, in software, it's very detrimental, and I'm sure software isn't the only industry where that's the case. That's why we need a far more nuanced approach. We could start by not handing out 20-year monopolies for ideas that take only a few months to implement and bring to market.

In general, I think that it would be useful for all legislation to carry along with it legally binding "original intent" written down in broad, layman terms. That way a legal framework would exist to protect people from legislative hypocrisy.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact