Slashdot Mirror
Google Loses Bedrock Suit, All Linux May Infringe
blair1q writes "CNet reports that Google has lost the lawsuit brought by Bedrock, for infringing on Patent 5,893,120, 'Methods and apparatus for information storage and retrieval using a hashing technique with external chaining and on-the-fly removal of expired data,' and has exposed the Linux kernel, in which the infringing code reportedly appears, to liability for patent-license fees. Red Hat also participated in the suit, arguing that the patent was invalid, but the court decided otherwise."
We dont have software patents in the real world.
Apparently the company is owned by an East Texas lawyer, and does not do anything except file patent lawsuits against everybody.
Patent Litigation Weekly: PubPat and "Patent Troll" Make Strange Bedfellows
Meanwhile, it's not unusual that a patent-enforcement company recently set up in East Texas would file suit there this week against ten Internet-based and software companies. Nor is it unusual that Texas corporate records show the company is owned by an erstwhile big-firm lawyer, in this case David Garrod, formerly of Goodwin Procter.
http://thepriorart.typepad.com/the_prior_art/2009/06/pubpat-and-patent-troll-make-strange-bedfellows.html
Here is a sad, 2009 slashdot article about the company.
Patent Trolls Target Small East Texas Companies
> Posted by kdawson on Friday July 24 2009, @09:14AM
> from the patently-absurd dept.
> An anonymous reader writes "In a sign that patent trolls are getting desperate to keep their cases in East Texas — long known as the friendliest venue for their claims — some have taken to suing tiny, no-name companies that are run by East Texas residents. The hope is that, if at least one defendant is located in East Texas, the judge will keep the entire case there. Nate Neel, a Longview, Texas resident with a small open source software company called CitiWare, was sued by Bedrock Computer Technologies in June despite (he claims) having no customers or other meaningful operations of any kind. In response, Mr. Neel has posted a strongly worded letter to Bedrock's attorneys on his Web site. It will be interesting to see how East Texas judges respond to this abuse of process perpetrated against their own residents."
http://slashdot.org/index2.pl?fhfilter=bedrock
BTW: the F/OSS company, CitiWare, is no longer in business. Sued of existence?
Although I don't trust Florian Mueller any further than I can spit, I think it may be informative to read his blog post on the subject. Florian does provide court records, and good description of the patent in question.
http://fosspatents.blogspot.com/2011/04/texas-jury-finds-against-google-in.html
This one has *got* to find itself appealed, and that appeal will happen well outside of East Texas.
I for one do not see folks like IBM, RH, Intel, Oracle, or other huge companies simply forking over either, even if the "licensing fee" was something ridiculously low. IT would be the camel's nose in the tent, and they know it.
Quo usque tandem abutere, Nimbus, patientia nostra?
If only Florian didn't spin everything to his own personal vendetta... I might actually read his articles.
But he ducks and weaves and picks out all the wrong facts to focus on... And gets it entirely wrong.
As far as the linux kernel goes? They've picked a very specific release train. 2.4.22, which came out in 25-Aug-2003 .
So, there's really nothing prior to that which offends the patent? Given the generic elevator/tree description of the patent, i find it very hard to believe.
I also note that they don't call out what code actually offends, so there's no way to track down when it specifically came into being -
but of course, that's useful when you don't want to call attention to the fact that the code has been in place longer than the patent has been around....
The good ol' software patent, a method and apparatus for extortion and not much else.
Cancel that retirement party
"Waste not one watt!" - CZ
Combining a chained hash table with a priority queue was obvious 40 years ago. How can something like this be patented? It seems it is time for severe penalties for trying to patent obvious things. Like a few years in prison. These people do way more damage than terrorists ever did.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
For a layman's analogy, when rummaging through the fridge for something, they have patented noticing that the mayo is past it's expiration and so throwing it out while you're there. That is all.
Not only is it blindingly obvious to nearly anyone in the field, it's so blindingly obvious that a beginning student is fairly likely to think of it.
I won't comment on the validity, it seems pretty obvious to combine techniques for accessing/modifying a hashed/linked list with combing a list for items to delete, but there's a trivial work-around for it. Don't delete items as you comb through them, simply mark them as invalid and put them on a list of records to be recovered. Periodically, or when running low on storage, delete items on the to-be-deleted list. Might even be faster when multi-threaded if the invalidate can be done with a lightweight synchronization rather than locking the record(s) out while recycling them; can even keep a private list of invalidated records, then add that to a global list to be recycled. Claims 2, 4, 6, 8 are ridiculous on the face of it, though - using dynamic limits for ANYTHING is not novel unless you can show a significant problem that hasn't been solved before. Simply specifying a dynamic value that a routine uses to count the number of iterations of a process, length of time to spend doing something, number of things to do in a pass, etc, is 40 years old at least.
I assume this is tongue and cheek, but the Eastern District of Texas (in its entirety) has 10 Federal judges and 10 magistrate judges. In Marshall Texas, where a majority of the "trolls" file, there are three judges. (http://www.txed.uscourts.gov/page1.shtml?location=info)
Notwithstanding the sometimes (frequently?) wacky jury results, an interesting thing is starting to happen though with the Eastern District of Texas: the judges are getting very good at patent cases. The judges are very good at doing "claim construction" and other technical aspects of patent cases. The result is cases tend not to languish because the judge is confused or baffled by the process or technology.
By contrast, there are districts that it is virtually impossible to litigate patent infringement cases because there is NO expertise among the judges.
Yep that's exactly what happens when you allow patents on raw ideas rather than specific and detailed implementations. No software patent should be valid without the source code detailing the implementation where simply changing code to work differently completely avoids the patent. Of course if that was the case nearly every software patent would be gone with the exception of things like the RSA patents where changing the code would make it no longer work.
The problem with the patent office is that when they started allowing companies to patent raw ideas in the form of software and business method patents suddenly you could get a patent for an idea with no implementation, make it as vague as possible and then sue everyone. And that's exactly what's been happening and getting more and more common. It's the reason the courts are being overwhelmed with patent suits and it's also the reason the patent office is being swarmed with patent applications. Reinforce the old rule that requires that the patent detail a specific implementation (and in the case of a software patent that means the source code is part of the patent) and much of the problems with the patent system will go away. The side benefit is all but the most important and innovative software inventions are no longer patentable and business method patents go away entirely.
Ah but we can hope the upcoming patent cases before the Supreme reach exactly this conclusion and reverse the nastiness the court created when they essentially allowed business method patents by invalidating previous tests. The only valid patent system is one that deals in specific implementations.
On the flip-side, it might encourage big companies to believe that software patents = easy money for no work. They're already doing less than they need to be, the last thing we want is for their shareholders to insist they can make the same profits by doing less.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
Patent suits are are only decided by Jury at the first stage. This one gets automatically appealed. Just like all the other ones.
I wouldn't get too worked up about this.
Sig Battery depleted. Reverting to safe mode.
I'm certain this will get quashed in appeal when a real court hears it. :) (East Texas is really just Western Arkansas, only with more teeth, but the same amount of sister-love......)
It's the Stay-Puft Marshmallow Man.
methods and apparatus for information storage and retrieval using a hashing technique with external chaining and on-the-fly removal of expired data.
information storage and retrieval - nothing new here. using a hashing technique - nothing new here with external chaining - using what?
Using a linked list that the hash bucket points to -- hence, external chaining... Yep, nothing new here, standard hash table practice, as opposed to moving the collision to a different unoccupied bucket.
and on-the-fly removal of expired data - "on-the-fly" is a nice buzzword, but this is also nothing new.
I built a connection to player# firewall/database for a simple game server in 1996 (patent filed in 1997). I used a single hash table to store valid open connections and blocked connections. The packed IP address was used as the key (hashed), and an external linked list per hash bucket was used to resolve collisions. If more than one connection mapped to a bucket, it was added to the bucket's external linked list. Each connection (linked list item) had an IP address, port#, validity flag, and time-stamp; The later I used to automatically remove expired connections -- as I traversed the hash table. HOW IS THIS NOT BLOODY OBVIOUS to a professional skilled in the art of hash tables and faced with the problem of expiring bucket entries?! The patent is invalid.
The patent is clearly in violation of my prior art, and doubtlessly many thousands of other's. Fortunately, my source code for that BS game is closed -- no patent troll can figure out the method I'm using and sue me..... The patent (troll) system makes it more beneficial for me to hide innovation -- The opposite of it's intended goal. Linux source code is open, ergo, a broad target for trolls.
So apparently we can just mix and match random old techniques, and call it innovation and get it patented?
Yep. Apparently that's what they did. If someone else hasn't already patented it, you'll get a patent for it, no matter how obvious. The obviousness clause is not actually applied to patents. The patent database is searched for prior art, and applications granted if none is found.
What I find interesting is the "Method AND Apparatus" terms that all these software patents must use in order to link them to hardware and thus make them patentable -- Without the apparatus, they are not patentable.
A patent is public information, the detailed description is required to be published, no doubt in a textual form.
An "infringing" bit of open source software source code is simply a detailed description of a patent. Ergo, source code must not actually be infringing since it is not an apparatus, it is simply performing the same exact task that a patent claim performs -- describing the patented method in detail.
If a translation from the patent claim text to source code text is not an infringement, then a binary representation of such text must also not be an infringement -- A PDF displaying example source code that allegedly infringes a patent dose not itself infringe the patent -- It's just a document / file...
Compiling source code is merely translating it into another binary representation, just like translating a patent claim into example source code is not an infringement, and the PDF translation is not an infringement, a binary executable translation is not an infringement either!
Only when you add the apparatus to the method by running the software should the patent be enforcible. Prior to running the executable code the software is purely a description of the method employed.
A patent suit must not be able force the source code of a "patent infringing" open source project to be taken down -- Otherwise, we must also remove from public view ALL PUBLIC PATENT information.
Distribution of a PDF that descri
No. It looks like it describes something utterly obvious done to a hash table, though.
Let's say you have a hash table, where you resolve hash collisions by having a linked list of everything that shares that hash value. Whenever you traverse that linked list, whether you're adding another entry, or searching for one of the items on that list, or whatever, since you have to traverse the list anyway, you could examine each entry on the list and possibly throw it away if it's something that you think you'll never use. Like, say, if it's an expired cache entry.
If I'm reading this patent right, that idea patented.
Seriously. And that's outrageous. If stuff this obvious is patentable, then programmers simply have no chance at all. You would have to hire a lawyer to work a week for every hour that a programmer works. It's just bloody fucking insane.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
I believe that chaining hash tables is somewhere in Knuth. Which means that it came out before 1980. Yes, on page 507 of volume 3 he talks about "search methods commonly known as hashing or scatter search", so these were COMMON in 1973. And he talks about ways of refining, and how the keys need to be adapted when the contents of the table changes if you want to retain uniqueness (which he doesn't consider worth the effort), He goes on for a number of pages. But please remember that this was not original research on his part. He was merely reporting on commonly used techniques and analyzing them. And he often didn't go into details. He only had so much space. (And he was recoding everything to work in MIX, which made I/O especially problematic.) But Corman was FAR from being the first word on Hash Table implementations. (In college, sometime in the 1970's, we talked about chaining hash tables and various implementations. Given the limitations on storage [RAM & punch cards...not disk or tape] you can bet we spent a lot of time making sure dead data didn't continue taking up space...though we didn't always properly clear it, merely deallocated it.)
So the "patent", by this "Doctrine of Equivalents" and "After-invented technology" would be considered a mere refinement of standard approaches. If the judge and jury are idiots, or possibly just not interested in justice, one can hope that some appeals court will notice this. (And hope that Google already has it in the court records, because otherwise the appeals courts won't be allowed to notice it.)
But the existence of that patent is further evidence that the patent system needs to be totally scrapped, and all extant patents be considered invalid, because of the malfeasance of the USPTO. I don't think it possible that mere misfeasance could lead to the current mess, though I suppose that there are some other felonies that would also be applicable. Fraud comes to mind. Possibly accepting bribes, though that would need to be proven, as it's not directly shown by the evidence to hand. But "improperly performing their duties in a way calculated to unjustly benefit some parties and unjustly injure others" seems like a good characterization. It might, however, be difficult to prove that they intended the injustice.
All in all, I don't think filing appropriate criminal charges against the offfice-holders at the USPTO would be very successful, however justified it might be. Much better would be to just declare the entire mess corrupt and revoke all existing patents and patent rulings (on, e.g., what can be patented). And then re-write the law into something that can actually be honestly applied.
At that, patent law probably isn't as bad as copyright law. Patents still expire. They may not reveal anything useful, but after they've expired they form a valid basis for claiming that some new patent is invalid. So they do have some good features. Copyrights, however, essentially never expire. And they are allowed to be used to cover items which are protected by a DRM that will definitely render that material "protected" unreadable before the copyright even comes up for renewal. (OTOH, copyrights are granted freely, and without formally filing. But someone else having produced essentially the same work previously isn't protection against a modified idea being copyrighted. In fact, even being in public domain now appears to no longer be protection.)
Have I given the impression that I consider the legal system corrupt? Compared the the legislators it's pure as the driven snow. Then there's the president...
I think we've pushed this "anyone can grow up to be president" thing too far.
I think the biggest problem with outfits like Google is that they're populated by guys in their 20s and 30s who seem to think that everyone worked with abacuses and vacuum tubes prior to 1990, and little realize the monumental amount of R&D and computing theory done in the 60s and 70s. Guys out of labs at IBM, MIT and Bell pretty much thought up one way or another the vast amount of "new technology" that everyone uses today. Maybe instead of hiring legal guns and relying on the non-existent historical perspective of a lot of coders and software engineers being pumped out today, guys like Google when going against patent trolls should roll out the old guns from that heady era when so many of these algorithms and approaches were developed. That way, when some evil patent attorney declares "My clients invented this in 1996" some old dude from MIT can slap down his thesis from 1968 and say "Fuck you, you lying cunt, me and the boys dreamed that one up over a hash pipe."
The world's burning. Moped Jesus spotted on I50. Details at 11.
Ask John Carmack about that and he'll tell you about having to settle over a patent on "Carmack's Reverse" taken out by Creative Labs some time after he wrote it.
With software patents you can't win even if the technique is named after you before somebody else files the patent.
Knuth told us and the patent office how bad it would get with software patents in 1994.
I still would not be at all surprised to find prior art.
I wouldn't either. Hash tables with linked lists are in my undergrad notes from 1977. It's in every curriculum, I'm sure.
Oh, yes, the patent also involves garbage collection on the fly. That reminds me, Hans Koomen and I did a implementation of Interlisp right around then too. It had that. I forget where we picked up the algorithm, it was so long ago, but I remember thinking how great it was that the principles had already been developed by the time we needed them.
Those were the days when people were still using rotary-dial telephones, mind you. The patent in question was granted on April 6, 1999.
To summarize: according to the claim, this patent combines two known techniques in what I would regard as an obvious manner. The patent only covers garbage collection on a particular type of data object. Back in the seventies the existing art was already sufficient for managing all data objects.
Parity: What to do when the weekend comes.
And they will probably go the distance with it in a manner like Bilski was ran- this is the camel's nose in the tent and it's not like it's really valid on several different fronts. (Based on what I'm about to mention, it might be that this is the plan on this...I can't say...)
Besides, there's some very likely prior art. The initial release of NLANR Squid (v. 1.0) was on July of 1996. When it was released, some six months before the Bedrock patent was filed, it was the inheritor of the Harvest HTTP Cache research project's code and resources, which a reduction to practice places it at about a year and a half prior to the Bedrock filing. At it's core, there is the very implementation of what is described in Bedrock's patent as it uses a hash with external chaining and expiry on the fly- it's how the whole thing does fast lookup and aging on the contents in the cache to begin with. It's how it works so "nicely" and what made it special back when it was implemented.
I am not merely a "consumer" or a "taxpayer". I am a Citizen of the State of Texas
Steven J. Vaughan-Nichols wrote a great article about it:
Idiotic Anti-Linux & Google Patent Decision
Red Hat has sued Bedrock to get the patent revoked for numerous reasons. Among others, they point out that Linux, which dates to 1991, predates the 1997 patent; that no one has ever used the patent; and that in any case Bedrock has no claims to the patent
http://www.zdnet.com/blog/open-source/idiotic-anti-linux-google-patent-decision/8736
And PJ commented on Steven J. Vaughan-Nichols article:
"If I might remind you, Mike Anderer told us years ago, back in 2004, that this was Microsoft's plan, to see to it that FOSS companies got sued over and over again for patent infringement, until they gave out: "In a world where there are $500 million dollar patent infringement lawsuits imposed on OS companies (although this is not completely settled yet), how would somebody like Red Hat compete when 6 months ago they only had $80-$90 million in cash? At that point they could not even afford to settle a fraction of a single judgment without devastating their shareholders. I suspect Microsoft may have 50 or more of these lawsuits in the queue. All of them are not asking for hundreds of millions, but most would be large enough to ruin anything but the largest companies. Red Hat did recently raise several hundred million which certainly gives them more staying power. Ultimately, I do not think any company except a few of the largest companies can offer any reasonable insulation to their customers from these types of judgments. You would need a market cap of more than a couple billion to just survive in the OS space." So that's what he said. And 50 or more lawsuits lined up means using proxies, obviously. Remember Microsoft trying to sell some of its patents that read on Linux, or so they claimed, to patent trolls? OIN played man in the middle on that one, but who is to say that was the only one? That's why it's an antitrust issue, I'd say, using patents like this as an anticompetitive weapon. And if you want to know what is wrong with software patents, the damage they are doing, read the quotations from various business executives (like Andy Grove) in the footnotes to this article.]"
http://groklaw.net/
BTW: the F/OSS company, CitiWare, mentioned in the first slashdot article about Bedrock, is apparently out of business. I wonder if they were sued out of business?
I wouldn't get too worked up about this.
Except that is this is only one of several such lawsuits, and there may be many more on the way.
PJ, at Groklaw, suspects that Microsoft is behind these lawsuits, and she suspects that there are many more on the way.
If I might remind you, Mike Anderer told us years ago, back in 2004, that this was Microsoft's plan, to see to it that FOSS companies got sued over and over again for patent infringement, until they gave out
I suspect Microsoft may have 50 or more of these lawsuits in the queue.
http://groklaw.net/
As it stands, they did prove that it wasn't a valid patent- but the Jury believed "the little guy's attornies" all the same
The jury was never asked to rule on the validity of the patent.
The jury was asked three questions which must be answered consistently:
[Paraphrased]
1 Did Bedrock prove by the weight of the evidence presented here that Google infringed on its patent?
On Claim 1 - Yes. On Claim 2 - Yes.
2 Did Google prove by the weight of the evidence presented here that it did not infringe on the patent?
On Claim 1 - No. On Claim 2 - No.
3 If you find that the patent was infringed, what would be fair and reasonable compensation for Bedrock, based on the weight of the evidence presented here?
$ 5 million.
Bedrock v. Google
For the case to reach the jury in this form, Google must have lost every argument with the judge at every stage in the case where the validity of the patent could be contested.
That does not bode well for an appeal.
The jury trial was demanded in this case - and it is an expensive and high-risk proposition.
The appellate court judge does not second-guess a jury on matters of fact.
The most he is likely to allow is an argument that any damages awarded were "excessive."
Neither are you likely to get very far arguing that the jury was biased or incompetent.