Hacker News Reader

I was asking this question because protection against others independently developing a similar algorithm is what the patent system provides, and that's why many people, including myself, are opposed to it.

I wasn't assuming that you personally or even your employer actually wanted that kind of protection. So my rather rhetorical question was intended to highlight that discrepancy.

You want to prevent others from copying your work but what you do by patenting it is more than that. You prevent others from making or benefiting from their own original invention. Of course, as long as software patents exist, you don't have much choice.

What about this is unique to software patents?

I don't understand the "destroy software patents" crusade. "Destroy all patents" makes sense, and "improve all patents" makes sense, but why are software patents any different than device or chemical patents?

My main point is not limited to software patents, it's just more striking in this area because the drawbacks of patents are so clearly demonstrated every day. And contrary to other disciplines, copyright covers a lot of ground when it comes to software.

But to understand what might make software special you have to look no further than the things that are currently not patentable and never have been, like mathematical approaches and business processes.

So, for instance, if you invent some data mining approach based on graph theory and you write your formulas on a piece of paper or into your matlab or R worksheet you cannot patent it. If you express the same ideas in a couple of C functions you can patent it.

Another example. Say you are the first to have the idea of selling coke and sandwiches at gas stations. You cannot patent it. But if you are the first to have the idea of selling avatars from within some computer game, you can patent it. Why is that?

The differentiation between patentable and non patentable types of approaches and ideas is not something that was invented by software patent crusaders. It's right there in the law already.

The trouble is that it's a lot easier to copy an algorithm without in any way violating copyright than it is to copy it without violating a patent on it.

I'm not so sure about that. If the license prohibits reverse engineering and someone still does it, that's clearly a violation of the law. If a commercial company does it, it's going to be difficult for them to keep secret as some employees will surely know about it.

I agree that part of what patents intend to do, protecting original invention from copycats, is legitimate. But if it's done in a way that punishes original invention that seems counter productive to me.

If the law allows reverse engineering (like here in Germany) your license can prohibit it as much as it wants. It's still legal and no violation of the law.

I believe even in Germany you cannot use the results of reverse engineering to copy the product and sell it commercially without violating some copyright law. But I'm not denying that patents can sometimes be helpful to protect against copycats.

I just think that prohibiting original independent invention in order to prevent copying is absurd because it punishes exactly the people who are supposed to be the beneficiaries of patents.

>"protection against others independently developing a similar algorithm is what the patent system provides, and that's why many people, including myself, are opposed to it"

Flip this around. It is protection against others wasting research on something that has already been invented and [supposedly] enables them to add their research to your own and thus add effort rather than waste it. Is that not a noble cause?

Patents have a double purpose: One is to ensure implementation details of inventions enter the public domain and the second is to provide a limited monopoly as a bribe to ensure that the public domain will ultimately benefit.

In this way patents promote innovation, encouraging the release of new technological developments quickly to other practitioners in the art and so accelerate said development.

The only (large) part I see broken in the current patent system is that the duration of monopolies hasn't reduced to keep pace with the acceleration in the life cycle of new technologies.

We're not concerned about other people independently developing a similar algorithm as it's rather non-trivial and as I mentioned, it took an awful lot of effort to develop it. We're much more concerned about the very real problem that people (e.g. competitors) will use our research and work.

Then what you do is keep your algorithm as a trade secret. The point of a patent is to publicize the invention so that others may benefit, and in compensation to the inventor they get a period of protection on its use so that they may license the invention and recoup R&D costs.

To that same token...

Let's say company A and company B create the exact same alogorithim in the exact same time. Company A markets it much better and effectively "wins" the market. Should their success be determined by purely R&D alone? Since this is a rather unrealistic (albeit becoming more and more realistic) situation. Imagine a slightly more realistic scenario. Company A starts R&D day 1 and finishes day 100. Company B starts R&D day 50 and finishes day 125. But Company B has better management, marketing, etc. and (again) effectively wins the market.

Who gets precedence?

> We're much more concerned about the very real problem that people (e.g. competitors) will use our research and work.

So why don't you just keep it secret? Is this ultimately the question?

"I'm not sure if that's due to a misunderstanding on your part or a lack of clarity on my part."

These are merely two perspectives from which we can view the same phenomenon.

We're much more concerned about the very real problem that people (e.g. competitors) will use our research and work.

How is that plausible? Aren't these trade secrets? Is your research/work publicly visible?

Seriously?

They sell the product to customers, other people spend a small amount of time reverse-engineering how it was done, and copy it.

I bet he assumed a product as a service, where the code isn't in the hands of the customer.

If your algorithm is non-trivial, wouldn't it be better protected by trade secret law? Patent law ostensibly requires you to disclose the algorithm (though patent attorneys seem to be very good at avoiding that), which would allow your competitors to develop competitive, but non-infringing algorithms. If your algorithm is so far ahead of everybody else, they don't stand a chance if you just keep it to yourselves, right?