Hacker News Reader

Comment by BrentRitterbeck | Hacker News Reader

As a community that does a lot of work dealing with algorithms, I'm curious as to how others feel about this. I think it is important to protect trade secrets (and algorithms would certainly fall under "trade secrets" for some products), but I also believe that hindering anything mathematical or scientific is, for lack of a better phrase, negative.

If your company depends heavily on a huge body of code, are you in favor or against software patents?

View on HN

Once you patent something it ceases to be a trade secret. This is an important feature of the patent system; to ensure that what would otherwise stay as trade secrets (eventually) become available to the public.

My company depends heavily on a huge body of code, but copyright protects us more than enough. The ideas are cheap, the code to implement it is hard. I think it is fair game if somebody came along and implemented the same ideas we do. Patents stifle this kind of healthy competition.

Thanks for the response.

I'll now pose a different question. Do people in this community keep their ideas out of the patent records, or do they patent? I come from a finance background and have recently returned to school to pursue an MSFE. In the financial world, new strategies, often implemented through algorithms, are the lifeblood of a lot of quantitative firms. Do people think it is better to not publish a patent if your business depends heavily an a very unique idea?

If you sat down everyone in the industry who had the same amount of education and professional experience, how many of them would have come up with a similar system? Sure there is more than one way to skin a cat, but after years of cat skinning school and years of professional cat skinning everybody comes to do it in close(maybe not identical) to the same way. I would step back and try to consider how unique the idea really is, and how easy it would be for someone with a similar background to figure it out just by looking at inputs and outputs.

Testing obviousness with the benefit of hindsight isn't always fair. :-)

One test for non-obviousness is if the problem has been known for a long time, but no one has solved it; or (even better), the standard approaches that are taught for it lead away from your solution.

Trading Technologies has in the past been aggressive about defending patents on their trading software.

On the other side of the aisle, many approaches are under trade secret protection.

I think operationally patents don't do what we would like, and that is to give a small innovator some room to profit from an invention. In real life, it is large companies that produce, acquire, and use them in a defensive fashion.

Our product relies on novel algorithms. That's our "secret sauce," if you will. We publish the algorithms we develop on peer reviewed conferences and journals. There have been improvements made to our work by others and we encourage that. What prevents our competitors from copying our work and reaping the rewards? So patents seems to be a necessary evil from our perspective. It's not the patents that we shall fault it's the patent holders.

Reply on news.ycombinator.com