My personal opinion is that if the patents required a pseudocode implementation using a common-standard language to represent the components and operations, it would be much easier to determine patentability, would prevent generic, wide-sweeping software patents and require the investment of actual invention rather than just an investment of dreaming to be patented. Physical device patents are required to be specific enough that it can be used to create the device using only the patent and references in the patent. The same should be applied to software patents, with the translation from the pseudocode components to whatever language or platform be the same as physically constructing the components of a machine.
My thesis is that even if I grant you a sane patent regime the macro level scales so fast and big that that changes the game and makes the concept impossible in practice.
EDIT: Reading your recent response to another commenter below, I don't want to give off the impression that I think any current software patents should be maintained. I think they should be all completely removed (or required to be resubmitted fitting new requirements that include specific details on implementation). My ideas are in the "dream world" where the rigor was applied since the first software patent was applied for.
> NO!
It's hard to believe, but it is true. Imagine simply replacing tiny transistors on a chip with big relays and physical switches that flop back and forth. Yes, it would take a long time and be a very big device, but it's a mathematically proven fact.
And it kills open source software, which doesn't have a model to pay license fees.
Plus there are so many not formally documented ideas that the Patent Office regularly issues patents covering well known prior art; surely you've heard of this? E.g. the XOR patent?
With any luck http://en.wikipedia.org/wiki/In_re_Bilski will drive a stake into the heart of this madness.
Anyway, we're still talking about adding so much friction to the writing of software that it would utterly change the field and curtail most development, especially in small firms ... well, I suppose a healthy underground would develop. But there wouldn't be much for YC to do.
No thanks.
Maybe the solution is holding software patents to the same rigors as mechanical patents. Why the distinction? That's what I don't understand. And still don't.
Why are they treated differently? By everyone. Even outside the courts.
2. Scale: just how many are used in any one system makes "deleting software patents" a good and probably the best outcome. Especially since there was no lack of health in the software industry before they were allowed, and there's no evidence I'm aware of that they now help it and plenty that they don't.
It also seems to break down, mostly due to scale, when you insert patent trolls into the game. In the physical world there's a lot of cross licencing, where one patent, revealed to the world by the very act of patenting it, often results in an improvement, whereby the two companies cross license.
Let's take a step back: the idea of patents is that for their absolute monopoly of limited duration you reveal your secret sauce instead of keeping it indefinitely hidden behind trade secret law. I fail to see how the publication of software patents helps the field in this way. Even if they weren't obscurely written, there are again just too many of them (too many ideas).
2. There are so many because it is easier to write code than it is to cut iron. That doesn't mean the value of the invention isn't worth just as much to the world and the inventor. Maybe more. We should see an exponentially increasing quality of life due to the exponentially increasing ease with which it is getting to innovate.
People can build a CNC in their garage with parts from the lumber store and a couple websites. That was unheard of 20 years ago. We have lasers in our junk that can start fires. Think of the leap fire gave us!
NO!
Outside of the lower end of the embedded realm, computer systems are fantastically more complicated than anything physical we've every tried to construct, and they contain so many ideas that could be subject to patent in the current regime that you analogy fails.