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

I think those who argue for patents have to show how exactly patents promote entrepreneurship and innovation.

What patents do is to boost the value of ideas relative to execution. They are supposed to help individuals with great ideas but without access to the resources required to execute on them.

For patents to make sense, there has to be a very large gap between the capital required to work on the idea itself and the capital required to bring it to market. Something like building a factory or conducting 10 years worth of clinical studies. Something that enables a big encumbant to copy an idea and win every time against the original innovator.

I don't think that is generally the case in the software space. It is certainly more true for enterprise stuff that you need sales people for and that's why MBA types may think more favorably about software patents. But even for that kind of software, the gap isn't nearly as large as for inventing a new type of nuclear reactor or jet engine or cancer drug.

Another problem with patents is that they are supposed to make people publish their research instead of keeping it secret, but many small companies and individuals cannot afford to go through the patenting process.

I can say for myself that I do not publish my research nor do I open source it. Secrecy is the cheapest option to have a head start with exactly the kind of ideas that might really benefit from patents. Ideas that you don't think someone else could have tomorrow by accident or by gradually expanding on existing ideas.

I'd also like to add to this that in our world, it's not the idea or theoretical mechanism that matters, but how well it's implemented.

If I can manage to copy your idea and give it a better UX, more reliability, better prices, etc. and consumers choose me instead, should you really be able to stop me?

I think ideas do matter. Not ideas like the one click order or hey let's put a buy button in a game, but how to accurately recognise everything that's said in a room full of people, how to seperate sarcastic from factual statements in text, how to predict the number of bananas that need to be on the shelf next saturday.

The problem with you copying my invention is that you didn't share my research costs, so you could easily beat me on price. But as they are software solutions, I can bring them to market myself without much difficulty. I can hire or partner with the best UX designers and marketing people and I'm protected by copyright.

I wouldn't want to stop you from working out similar ideas and compete with me, nor would I want to become a law firm spending my time in court with people who had similar ideas. Patents are a danger and I don't need their protection because software ideas are either trivial and shouldn't be protected or they are complex and don't need protection beyond copyright and contract law.

But if I decided to create, for example, a voice-recognition API, I wouldn't be copying your method, I'd be creating my own, better method.

I think the issue is that your patent wouldn't cover your algorithm, it would cover "service which translates audio into text."

No, you can't patent a problem, only a particular solution to a problem.

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