Linux Scare Tactics (Forbes) [LWN.net]

Linux Scare Tactics (Forbes)

Posted Aug 2, 2004 21:02 UTC (Mon) by BrucePerens (guest, #2510)
In reply to: Linux Scare Tactics (Forbes) by adulau
Parent article: Linux Scare Tactics (Forbes)

So, tell me what you plan to do if you are a developer and you go to court? You can tell the judge software patents *are* incompatible with the author rights, which will not go too far. You can attempt to license the patent and be in conflict with the GPL (and despite what Dan Ravicher says, we will not be seeing many patents licensed for everyone's free use in GPL software, which is what the GPL would require). You can defend against the patent. OSRM would defend against the patent.

I am at real risk of being sued as are many individual developers. I would be able to afford a few days in court, and would then have to sign a settlement no matter how pernicious the terms. Those terms would probably make me sign over my copyrights and agree not to develop Open Source code any longer.

Thanks

Bruce

to post comments

Linux Scare Tactics (Forbes)

Posted Aug 3, 2004 1:37 UTC (Tue) by hs1 (guest, #23692) [Link] (2 responses)

brief question: afaik, patents still require to "significantly advance the state of the art", at least in theory. if you don't know about the patent, and come up with a patent violation when confronted with a similar problem, doesn't that already invalidate the patent?

No.

Posted Aug 3, 2004 3:24 UTC (Tue) by dwheeler (guest, #1216) [Link] (1 responses)

No. For patents, whoever registers first owns the idea. If you independently came up with it, it doesn't matter. And the patent office isn't really qualified to do this research, which makes the situation much worse. For LZW (the basis of GIF), the patent office actually gave the same patent twice to two different companies (IBM and Unisys)! In theory, if it's publicly known it can't be patented later by someone else, but court costs make that only theoretical. Because the costs are so large, only companies can practically participate in a patent court; even if you're right, you'll lose (and create a court decision making future losses more likely), because individuals cannot afford the fees to make it through a patent court case.

No.

Posted Aug 4, 2004 15:38 UTC (Wed) by raytd (guest, #4823) [Link]

No. For patents, whoever registers first owns the idea. (bold emphasis added)

May I ask what leads you to believe ideas are patentable? I have been taught that it is the implementation of an idea that is patentable. Not the idea itself.

Linux Scare Tactics (Forbes)

Posted Aug 3, 2004 14:35 UTC (Tue) by adulau (guest, #1131) [Link] (1 responses)

IMHO

The situation is getting worse if you acknowledge the existing patent. A simple example, a lot of patents are similar or related or cross-referenced. If you gain a non-revocable RF-License (for GPL only) from a "friendly" patenter, everything looks ok (a part the complexity layer this is not good but it's another story). Now, an unfriendly patent holder said his patent is similar to the patent where you have the RF-license. He starts a procedure against the patent holder to devalidate the friendly patent. The judge said that the unfriendly patent is valid and the other one is devalidated due to the anteriority. Now, you have publicly announced a RF-license for the friendly patent, the unfriendly patent holder can easily ask for more money due to the "ack" of the patent... The patent system is cleary flawed. Working with the patent system is more dangerous than not working with it.

Linux Scare Tactics (Forbes)

Posted Aug 3, 2004 18:31 UTC (Tue) by mmarsh (subscriber, #17029) [Link]

Not working with the patent system just isn't a real option for anyone in the USA. The system's there, and ignoring it doesn't protect you from it. What OSRM is doing is saying, "There appears to be a real risk of patent lawsuits against Linux users/developers. Your pockets might not be deep enough to defend yourself in court, but if you buy our insurance, we'll be able to defend you, and all of our other policyholders." If a patent gets smacked down in court from one case, everyone else is protected against further litigation using that patent.

We're talking about insurance here, which is all about risk management. If you're convinced that you'll never be sued over an alleged patent infringement, then don't buy a policy. If you have a stable of lawyers who can defend you and possibly launch a counter-attack, then don't buy a policy. If you want to pool your financial resources with others, you can either support the EFF and hope they defend you, or you can buy a policy and *know* that OSRM will defend you.

The triple damages aspect is somewhat unfortunate. Ideally, we'd have a group of lawyers and programmers looking at all of the patents and figuring out which ones are vulnerable. Since those programmers can no longer develop software without worrying about taint, I doubt we'd find many takers. Perhaps someone who is retired or has moved on to something aside from programming might volunteer, if any such qualified person exists. The next best thing might be a database of all the neat hacks people can remember, ideally with names, dates, and corroborating evidence, that the lawyers can then search through for prior art. That would be taint-free, but likely also a serious pain.