Hacker News new | ask | show | jobs
by kevinpet 6106 days ago
I think it's interesting that everyone is assuming it's a bad software patent. Here's the chain of reasoning I see again and again, and I don't buy it:

1. Some software patents are vague, overbroad, or not novel. 2. This is a software patent. 3. This patent is vague, overbroad or not novel.

It doesn't logically hold up. You may want to argue that we shouldn't have software patents as a matter of principle, but pointing to examples of bad patents doesn't prove this. The guidepost for me is RSA. To me, if you have patents, then RSA should be patentable. It was novel, it was useful, and the fact that it could be implemented in software doesn't change that.

This being about software patents is coincidental. This same story could be told by a tool manufacturer, or an MP3 manufacturer, or anyone else. The real issues are whether the patent is valid, whether they are infringing, and then, if we find that we don't like the consequences when all is played out, figure out where the breakdown between what is really good for innovation and what is actually happening is.

If you want to say no software patents, then someone needs to come up with a strong argument why a new way to build an amplifier that improves sound quality should be patentable, but a new way to compress audio files that improves sound quality should not be.

There's a lot that could be done with patent reform to make things better, but doing away with software patents is orthogonal to this.

Business practice patents are a different story. Most of them don't patent a "how to do X", they patent a "wouldn't it be cool if we did X". My example is that Amazon One-click was trivial to implement. There was no "how". The proof of this is that B&N worked around it by going back in and adding a second click. You couldn't work around a good patent in this kind of way.

2 comments

As a developer, the main problems with software patents are that 1) pretty much any combination of things you're doing could currently be patented by someone else, 2) it's impossible to tell if you're infringing something until you get sued, and 3) things that should be filtered out by the "obviousness" principle aren't, such that you're constantly in danger of stumbling onto a patent just by solving a problem.

A patent is a very, very, very sharp tool: it grants someone monopoly rights to that "invention" for 17 years (or 20, if it's international). Anything that most people could come up with inside 2 days, or even 2 weeks, of thinking about a problem shouldn't be patentable; that's ludicrous, and counts as "obvious" in my book. Things that are obvious combinations or extensions of existing inventions aren't supposed to be patentable.

So a lot of this software/business patent nonsense would go away if the patent office would apply a more ruthless standard of patent obviousness. Most software patents are things that at least 10% of the engineering population would come up with inside a week of trying to solve the same sort of problem; many of them are the sort of thing that 90% of engineers would come up with inside 4 hours. That doesn't apply to RSA, or to MP3, or to most algorithmic things, but it does apply to things like "method for holding online auction" or "method for allowing purchaser to assemble a shopping cart" or the other ludicrous sorts of things people patent.

So I totally agree, RSA should be patentable: it's a non-obvious invention, and if you're using RSA without their permission there's a 100% chance it's because you stole it/reverse engineered it. But most software patents are the sort of thing you accidentally stumble onto in the course of solving the same sets of problems, and patents were never intended to (and should not) preclude other people from solving those same problems. You can't patent the idea of holding an auction in meatspace, but suddenly because it involves software it's patentable? Etc. Most software patents are much closer to business process patents than to mechanical patents.

Personally, I'd settle for simple ruthless application of the obviousness principle. The USO PTO granted more than 40,000 software patents last year. If you only included things that were really novel, that a decent engineer wouldn't come up with on their own in a few days or weeks of thinking about a problem, and that you'd have a hard time accidentally infringing, I bet that number would be more like 500.

Keep in mind, there are plenty of other things that aren't patentable: story ideas, for example. So yes, you're going to draw an arbitrary line about what's patentable, but that's the job of congress and the patent office: to draw that line in the place it maximizes utility.Remember, there is no natural right to a patent in the US. Congress has the power to determine what's patentable and what's not, and is explicitly directed by the constitution to do so in a way that maximizes progress. The way it's done for software hinders progress right now.

You say that you approve of RSA because if someone is using it, its inevitable that they learned of it from RSA. I don't think you'd have to go that far to say what's worth patenting. I think that's the purpose of patents -- to make people disclose worthwhile discoveries so that others can benefit from them at a fee -- but they can't be structured that way legally, because unlike copyright, it's very hard to show the origin of ideas.

What makes a patent a bad patent is usually that it's on the idea of what's a good idea to do from a business perspective, and once you've realized that, hey, online auctions are a great idea, the mechanics of writing the code is pretty straightforward, as you say.

I'm not clear how things are de facto in European law, but from reading http://en.wikipedia.org/wiki/T_258/03 it looks like at least some people are on the right track -- you can patent something involving software, but only if you innovation is technical in nature, not a business innovation that involves a lot of technical details to implement.

I'm asserting that it's a bad software patent because I'm familiar with the details. You obviously don't have to take it on faith that my assertion is correct. Believe me, I'd love to share the details--and I will when my friend gives me the go-ahead.

I agree with you that not all software patents are vague, overly broad, or anticipated by prior art. I've been on the other side myself.