Hacker News Reader

Comment by alexvr | Hacker News Reader

I personally think copyright law provides enough protection for software. I think patents would actually harm innovation in the software world.

Algorithms are inherently protected, anyway - they're not like clever mechanical solutions that can be reverse-engineered by anyone with eyes and a brain; they're usually close-sourced when the author wants them to be. Patents may be necessary to protect products with transparent functionality, but software usually doesn't fit that bill unless it's deliberately open-sourced.

Even if you developed a strong AI algorithm that will change the world, there should be competition. It's not like Microsoft will hire a few engineers to rapidly develop something that transcends yours; it would be a reverse engineering challenge to any competition.

What if algorithms behind network protocols were patented? Computers would still be novelties and corporate tools, I think.

View on HN

So patents solve the incentive problem, just like copyright. Copyright solves the incentive problem when "writing masses of software" is the hard part. Patents solve the incentive problem when "figuring out what software to write" is the hard part. Take things like garbage collectors. Even the most complex are only 15,000-20,000 lines of code. Writing those lines of code was the easy part--it was figuring out what that code should do, and doing all the testing to figure out what worked and what didn't that was the hard part.

>Copyright solves the incentive problem when "writing masses of software" is the hard part.

That seems backward to me--when you've already figured out what to make, and the hard part is producing mass quantities of it--isn't that what patents are designed to protect?

>Writing those lines of code was the easy part--it was figuring out what that code should do, and doing all the testing to figure out what worked and what didn't that was the hard part.

Similarly, typing the words of a novel is easy; figuring out what to write is hard, which is a good reason for copyright protections, but I'm not sure if it has any bearing on the issue of patents.

> That seems backward to me--when you've already figured out what to make, and the hard part is producing mass quantities of it--isn't that what patents are designed to protect?

The archetype of what a patent is designed to protect is something like a drug formula. Manufacturing lots of drugs is easy--figuring out what chemical to manufacture is the hard part. That is both hard, expensive, and easy to reverse-engineer.

OK. It's probably more common in folklore than in real life, but I was thinking of the situation where mass production was the hard part, for example, where a lone inventor comes up with a process or design, but does not publicly disclose it because of fears that someone else (with better ability to mass produce) would capitalize on it.

So I agree that patents (and the extent of their protection) makes sense for things that are "hard and expensive", in software as in other fields. But it seems like many software patents that get discussed are for things that did not require a lot of work; the 20-year monopoly seems to be a disproportionate reward just for being first, regardless of the work involved.

Reply on news.ycombinator.com