Hacker News Reader

Comment by billybob | Hacker News Reader

"The complexity of software and low standards for patent eligibility mean that software engineers produce potentially patentable ideas all the time. [Snip] Auditing 10 million lines of code for compliance with 18,000 patents is an impossible task—especially because the meaning of a patent’s claims are often not clear until after they have been litigated. [Snip] The result is a transfer of wealth from young, growing, innovative companies like Google to mature, bureaucratic companies like Microsoft and IBM—precisely the opposite of the effect the patent system is supposed to have."

Those three sentences together form the best explanation I've seen yet on why software patents are so problematic.

View on HN

low standards for patent eligibility

I'm not sure if the standards are low, rather that the examiners don't abide by the rules set out by the USPTO. Novel and non-obvious are both requirements.

because the meaning of a patent’s claims are often not clear until after they have been litigated

This is another failing of current examiners. The patent should be written such that one skilled in the art could reproduce it. The patents are convoluted and unclear enough that one likely couldn't produce much of what is patented, because its not clear what the invention is. The examiners should make people resubmit patents, under time constraints, so that they are clear. Much like submitting to a journal.

To be clear, patent "eligibility" usually refers to the requirements of 35 USC 101, i.e. any new and useful method or machine. Novelty and obviousness come later, hence OP's term "potentially patentable." The point is that pretty much any software process gets you past the initial hurdle.

And obviousness isn't as easy to show as you imply. Everything has to be based on prior art, which isn't always readily available.

Clarity is subjective, and even claims that seem clear can get twisted around once litigation starts. Attorneys are masters at this.

I'm just pointing out that it's not as easy as simply getting better examiners, although that would certainly help.

I've said it before, but I'd be really interested to see a requirement for getting chartered engineer status be to peer-review a certain number of patent applications. It might cut down on the number of frivolous patents a touch.

Is the "teaching" requirement for a patent really weak in court? It seems to me most patents are obfuscated works of legalese and ought to be invalidated on that basis, by a relatively clear reading of the law.

Or is it too easy to pay some expert witness $1000 to come in and say "yeah, that's totally clear to me and I'm an expert, so nyah nyah nyah."

Yea, there is a reason why math is not patentable, and it should not be patentable just because it is executed on a computer.

Reply on news.ycombinator.com