The techdirt author obviously didn't bother to read the comments on the slashdot article outlining the second claim.

The IBM patent lists the coder's software as prior art. If they were stealing the idea, they obviously wouldn't do that. They're patenting enhancements that aren't in the coder's software, although the difficulty of reading patents makes that difficult to distinguish.

"It's funny how often we hear patent system supporters tell us that if you haven't actually gotten a patent for your invention, it's perfectly reasonable for someone else to go and patent it instead."

I'm not sure I've heard anyone say this, ever.

This is particularly troubling in the US, where the company could file for a patent after seeing the commit, but claim to have "invented" the idea before. IANAL, but I think the burden of proof would be on anyone who challenged the patent.

If the patent is granted, then any litigant bears the burden of proof in disproving it. The cost of patent litigation is sufficient that any institution would have difficulty pursuing such a case to that point, much less a group of volunteers that maintain an open source project.

I was going to argue why I think you are wrong but I am also not a lawer so instead, I will point out that IANAL comments in legal threads generally provide more noise than signal. Maybe we should strive to avoid them?

A reply from a lawyer would be ideal, but since the rest of us are affected by the legal system, I do think some attempts to understand it are worthwhile.

Here's a link to argue against me, but does it apply after the patent is granted, or only before? There's presumption of validity after. http://patents101.com/2009/03/date-of-invention-first-to-inv...

What are your arguments? He is right the US is first to invent. I think he is wrong in the claim to have invented it earlier bit is easy, the repo commit time line(and mailing list emails etc.) would give a firm date which the claiming party would have to have proof of inventing before. That is to say it isn't like a criminal case where only one side has to prove anything both parties have to prove invention dates.

Can one patent the method of reading open source code and patenting what's been found?

That's it. I'm patenting patent-related recursive jokes. If I see one more of these on HN, I'm taking the poster to the Supreme Court.

The Supreme Court only has original jurisdiction for a few very rare types of cases (which are outlined in the US Constitution) and discretionary jurisdiction for everything else, so they have no obligation to grant a writ of certiorari and hear your case.

You have to start with one of the lower federal courts, instead.

I refer you to this comment on reddit: http://www.reddit.com/r/AskReddit/comments/edxyk/submitted_a...

(and its responses, if you think I'm serious :) )

A meta-patent? Would still be vulnerable to meta-meta patenting tactics.

It's all patents. All the way down.

Halliburton actually tried to patent patent-trolling: http://www.techdirt.com/articles/20081107/0118162765.shtml

The state of software patents continues to deteriorate without any hope of reform in sight. I'm starting to suspect the situation will only improve once it has become bad enough that major players in the game begin to ignore litigation involving software patents. Perhaps defendants will not show up for court en masse, causing the current corrupt system to starve itself out of its patent attorneys, 'IP holding firms,' and the judges and districts willing to pander to them. What kind of trouble for the system would be caused by a large number of parties failing to appear?

Not quite sure I follow. If you don't show up for your court date, a default judgement can be entered against you and the bad guys win.

I think you started off with a good idea: major players beginning to ignore litigation might improve the situation. But I think that would be more likely to come from judges throwing out cases than defendants not showing up for court, for reasons that eli stated. You also kinda lost me at the word "corrupt," because you took on a sensationalist tone. I think that "corrupt" and "in need of reform" are not the same problem. The patent system was designed with noble intentions. Its major fault (IMHO) is that it hasn't kept up with technology.

Oh well... It's clear the USPTO isn't doing any reasonable homework in approving patents, but this issue in particular isn't that troubling because there is clearly prior art. Sounds like these companies are just wasting their money on something that will never actually be useful.

The vast information required to do "reasonable homework" on software patents makes me hesitant to blame the USPTO. I blame the rest of the system for not making it clear that such patents are nonstatutory.