tag:blogger.com,1999:blog-5246987755651065286.post8579731666392748023..comments2024-02-22T16:15:42.388-08:00Comments on cbloom rants: 05-25-15 - The Anti-Patent Patent Poolcbloomhttp://www.blogger.com/profile/10714564834899413045noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-5246987755651065286.post-65774512898252391692015-05-27T12:56:00.009-07:002015-05-27T12:56:00.009-07:00As far as I can tell, neither patenting *nor* publ...As far as I can tell, neither patenting *nor* publishing actually works that well at preventing stuff from being patented.<br /><br />Patent examiners (and patent attorneys) find related patents using a keyword search. It's easy to describe the same technique in different words so that two instances of the same idea are very hard to identify that way, and it doesn't help that the legalese patents are written in makes it even harder to check if two patents describe the same thing than it would be otherwise. (And it can be *really* hard even with the best notation, if two people happened to arrive at the same core idea in two different ways.)<br /><br />In patent litigation, it is rare for anyone to go for prior art. There are numerous reasons.<br /><br />One is that trying to overturn a patent on prior art grounds after it's been granted is hard. You have to prove that <i>every single claim</i> in it has been described by someone else before. If that description is in a patent or scholarly journal, your chances are good; other sources, less so. Either way, doing this requires a lot of time and effort for every claim you want to dispute, and if there's a single original claim left in the patent, the patent itself will not go away. (Although the non-original claims may get stricken.)<br /><br />The more important point is, "are we <i>actually</i> in violation of this patent"? Patents are written in legalese :), and it can be pretty hard to figure out what they actually do (or do not) claim. Some claims are excessively broad, but if you find claims that you're definitely not in violation of, you can throw out large parts of the patent (everything dependent on those claims) and focus your efforts elsewhere.<br /><br />It is usually much easier to prove that a particularly-worded claim does not apply to you, even on some minor technicality, than it is to prove that a claim is invalid. Prior art is a last-ditch effort; you only go there once you are pretty sure that you are in fact in violation of specific claims, and then you try to find prior art for those if you're sure they are not original. (Or, more typically, you just pay up.)<br /><br />It is quite rare for patent lawsuits to actually go before court. Taking it that far is a large money and time sink, and neither plaintiffs nor defendants usually want to risk it. Defendants usually can't afford the cost of a prolonged legal battle, so they really have no choice. It's usually a couple of salvos fired between lawyers, followed by a quiet settlement.<br /><br />Worrying about prior art and the exact details of patents and what they cover is a very engineer-y thing to do. As far as I can tell (and that might just be me being cynical), nobody else cares. You don't actually get patents to cover some Very Important Inventions that are the backbone of your business. That's the popular narrative, but that's not what happens. In reality, companies just collect patents like crazy hoarders and throw them around in large packs: quantity not quality.<br /><br />Then in the media, you'll see mention of the MS patent on FAT, and engineers will get upset because FAT is such a stupid thing to patent, but it's not like somebody went "let's sue them over FAT, FAT is our core tech". MS slapped someone else with 100 patent violations, probably three quarters of which were totally bogus and easily defused, and FAT happened to be the silliest one of those that stuck. Could have easily been something else.Fabian 'ryg' Giesenhttps://www.blogger.com/profile/13685994980026854143noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-66446831265928656302015-05-27T07:33:03.590-07:002015-05-27T07:33:03.590-07:00Is filing a patent to protect a free idea better t...Is filing a patent to protect a free idea better than writing a freely available publication? Publication should count as prior art, right? (I guess you could dispute anything in court - including a previously filed patent though...)Arseny Kapoulkinehttps://www.blogger.com/profile/18310595345818946666noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-9070651793522886302015-05-26T16:06:33.269-07:002015-05-26T16:06:33.269-07:00> What do you do if a "non-practicing enti...> What do you do if a "non-practicing entity" (patent troll) sues you?<br /><br />Right, the cross-licensing doesn't apply here. My idea is that once the APPP has enough money, it can act sort of like an ACLU for patent defense. When you're picked on by a troll, you can ask the APPP for help, and if your case looks good, we'll take it on and try to challenge that patent. The eventual goal being to create a new attitude about trolls and build a momentum for standing up to them.<br /><br />> What does this do about the MPEG-LA type of model? There's a bunch of patents on the specifics of $thing, and you need to do $thing because it's part of an open standard that you need to support<br /><br />I'm not sure the APPP can do much about this. I mean I have some far-fetched ideas but no. The problem here is just that we allow these patented standards to be created and popularized.<br /><br />Fighting that is mainly about having patent-free alternatives, and about building the public desire to stay away from the patented standards.<br /><br />One thing the APPP could be involved in is fighting the generalization of MPEG-LA type patents. Like you can patent your specific encoding scheme, fine, we can't stop that. But the more generic interpretations like "any mocomp + huff is patented" we would fight.cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-331660658619350122015-05-26T14:58:18.948-07:002015-05-26T14:58:18.948-07:00It's quite intentional that the APPP tries to ...It's quite intentional that the APPP tries to stop others from getting junk patents, but is happy to get them itself.<br /><br />A lesser goal is build its war chest of patents. The greater goal is that by getting junk patents yourself, you stop others from getting them, and build a bunch of documented prior art that you can use to stop future junk patents. (hey, you can't patent addition, we already did).<br /><br />It's my intention that the APPP does NOT try to be holier than thou and above the morals of the system, blah blah. We use the broken system to bring down the system. We aren't going to play clean when everyone else is fighting dirty.<br /><br />Perhaps the original post should say "It greatly reduces the issuance of junk patents (to anyone but the APPP)."<br />cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-6502706579761476852015-05-26T11:32:22.524-07:002015-05-26T11:32:22.524-07:00Re #4:
So the scenario you're thinking about ...Re #4:<br /><br />So the scenario you're thinking about is the "IBM vs. Sun" kind of deal: big company with patent war-chest extorts money from a smaller company. I can see this working in principle, although I'd want to know a few more details before I got patents and assigned them over, but let's suppose there's a credible way to make this scenario work.<br /><br />What do you do if a "non-practicing entity" (patent troll) sues you? What is the APPP gonna counter-sue them with? They're not engaging in anything other than litigation!<br /><br />What does this do about the MPEG-LA type of model? There's a bunch of patents on the specifics of $thing, and you need to do $thing because it's part of an open standard that you need to support - various possible reasons: for inter-op with existing software, because you're on a government (or similar) contract that requires standards compliance, because it's actually the best available thing at doing what you need, whatever. They're not gonna sue you, not if they can avoid it. They just want their cut. How does the APPP help at all here? How does it prevent people from encircling a whole general area with patents so that anyone who steps over the line has to pay?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-41482832784862625262015-05-26T11:15:39.994-07:002015-05-26T11:15:39.994-07:00Re #5: you said in the original post:
"The A...Re #5: you said in the original post:<br /><br />"The APPP pro-actively watches all patent submissions and objects to ones that cover prior art, are obvious and trivial, or excessively broad. It greatly reduces the issuance of junk patents."<br /><br />vs.<br /><br />"The APPP just wants everything. It wants to be the Borg of patents and gobble up all it can. Legitimate or bullshit. Use the fucked up system to patent every trivial thing you can think of."<br /><br />now which one is it? I'm confused.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-5570799200890987122015-05-26T10:02:52.658-07:002015-05-26T10:02:52.658-07:00Basically your #4 is the major power of the APPP.
...Basically your #4 is the major power of the APPP.<br /><br />Without the APPP, small company patents some idea. Big company sues them for a bunch of patent infringements. Small company can't do anything and folds.<br /><br />With the APPP, small company patents idea and donates it to the APPP. Big company sues small company for a bunch of patent infringements. APPP counter-sues for violation of a bunch of APPP patents. Suit is settled (*).<br /><br />(* granted details of this settlement are a big vague in my mind)<br /><br />In a way it provides a co-op so that all the individuals and small companies can pool their IP to fight the behemoths.<br />cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-24219123385721878102015-05-26T09:58:04.044-07:002015-05-26T09:58:04.044-07:001&2. Yes, the initial seeding is an issue. Ho...<br />1&2. Yes, the initial seeding is an issue. However I don't think it's really so dire. What you'd do is write op-ed pieces, do radio shows, get attention, try to get charitable funding. It only really takes one rich person who believes in the cause, or a good patreon-type drive. Raising $2M to fund the first 100 patents is a pretty small amount.<br /><br />3. The APPP is looking for donations from companies that are not interested in profiting from their patents, due to their beliefs. The $20k patent cost is negligible for any company. In return the APPP provides legal defense (once the war chest is accumulated) which is worth a huge amount of security to the companies that donate.<br /><br />4. The APPP provides a much larger war chest that you use by proxy. Some big company sues you for using some nonsense that you claim. The APPP acts as your representative. They counter-sue the company, claiming violation of many APPP patents. Instead of your one trivial patent that the company could easily ignore,<br />the APPP has lots that they are worried about, so they settle. A small company on its own basically has no standing here; by many of them acting together through the APPP their ability to force cross-licensing is greatly magnified.<br /><br />5. Why does the APPP prefer detailed and narrowly scoped? The APPP just wants everything. It wants to be the Borg of patents and gobble up all it can. Legitimate or bullshit. Use the fucked up system to patent every trivial thing you can think of.<br /><br />6. I was think of the MPEG-LA or the RIM/NTP where you manage to patent a base technology that covers someone's core business, so that they have really no choice but to settle. But yeah, the swarm of bees model of lots of tiny patents works too.<br />cbloomhttps://www.blogger.com/profile/10714564834899413045noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-649144549102589052015-05-26T09:42:14.356-07:002015-05-26T09:42:14.356-07:00Okay, here are a couple more pragmatic issues with...Okay, here are a couple more pragmatic issues with this:<br /><br />1. Obtaining a patent, especially a SW patent, in the first place is a costly process in both time and money. Expect to spend somewhere south of $15k and over a week writing one down.<br />2. Which means that seeding the APPP with even just a hundred such patents is quite a substantial initial investment with uncertain results.<br />3. It also means that in donating patents to the pool, you're asking companies to write off a lot of money, which is not a good sale.<br />4. The goal of having a large patent "war chest" as a strategic investment is, usually, mutually assured destruction: somebody sues you over a patent violation, you sue them right back for violating some of your broader patents, then everybody settles out of court for a reasonable licensing fee. You would have to explain to me in detail how this process would work with the APPP in the middle, and why any company would donate their patents giving the likely extra complications this is going to cause.<br />5. Part of the perversity of the current system is that the kind of patents you would prefer in the APPP - detailed and narrowly scoped - are precisely those least likely to make other "players" worried about it in the first place. See the "workflow" above.<br />6. In particular, I'm not sure I buy the assumption that having a few "key patents" for really important algorithmic inventions would make a big difference. As far as I can tell, what actually wins patent lawsuits is large collections of nebulous patents. If you get to the stage where you have lawyers worrying about the details of whether a particular patent is relevant or not, you're already bleeding money, so you generally try not to let things get that far.Fabian 'ryg' Giesenhttps://www.blogger.com/profile/13685994980026854143noreply@blogger.com