In an interesting twist last week, patent owners are being accused of stealing inventions (rather than having their inventions “stolen” by others vis-a-vis infringement). In a Techrights.org article, IBM and Tandberg are both accused of taking innovations from open source software repositories and claiming them as their own inventions to the USPTO. Indeed, this is a very serious accusation, since inventors are required to sign, under penalty of perjury, an oath stating that he or she “believes the named inventor or inventors to be the original and first inventor or inventors of the subject matter which is claimed and for which a patent is sought.” In addition, an attorney who knowingly allowed an inventor to falsely sign such an oath would likely be suborning perjury.
Taking a closer look at the IBM accusation, Big Blue is accused of stealing US Patent 7,552,305, titled “Dynamic and real-time management of memory,” from a software developer who dropped a tip over at Slashdot.org. The disgruntled slashdotter points to a program called HeapCheck as the source of the invention, and also feels slighted for having been included as a prior art reference:
to add insult to injury, [the IBM patent] even mentions my site (via a non-working link to an old version of my page) in the patent references!
Sure enough, the ‘305 Patent lists as a reference:
Tsiodras, Thanassis, Heap Check 1.2, Feb. 2, 2002 Available at: http://web.archive.org/web/20020202030317/http://www.softlab.ntua.gr/ttsiod/HeapCheck.html.
If IBM really did try to steal the Heap Check invention, this would be a bit like illegally squatting a house and then throwing a house-warming party with a classified ad that said, “Come check out my new stolen house!” So why is this reference in the patent? Well, another famous rule at the USPTO is known as rule 56, which tells us that:
Each individual associated with the filing and prosecution of a patent application has a duty … to disclose to the Office all information known to that individual to be material to patentability
In other words, the IBM patent attorneys cited the Heap Check application out of an obligation to make the PTO aware of any relevant prior art. Meanwhile, the invention that IBM claims to own is most likely an improvement to the Heap Check application. In fact, if Heap Check, as it existed in 2002, performed every single step of one of the patent claims, then those claims would be invalid for as anticipated. Granted, I don’t really know what Heap Check did or didn’t do, but I can tell that the underlined language was added to the first claim of the ‘305 Patent during prosecution:
1. A method of facilitating management of dynamically allocated memory, said method comprising:
havingallocating a dynamically allocated memory buffer for a process in accordance with an allocation mode of the process that enables the determining in real-time of an invalid access, wherein setting the allocation mode for the process to enable determining in real-time an invalid access is performed in real-time, and wherein the setting sets the allocation mode for an application executed by the process without requiring recompiling, linking or loading of the application to set, in real-time, the allocation mode for the application; and
determining in real-time that an invalid access of the dynamically allocated memory buffer has occurred.
Evidently, the Examiner believed something in that claim was not in the Heap Check application, or else the Patent wouldn’t have issued. If the Examiner is wrong, then the real beef is with the USPTO, not IBM.
Moving on to the accusations against Tandberg, the offending application is this international patent application (commonly referred to as a PCT) and not an issued patent, as is the case with IBM. According to WIPO records, the case hasn’t entered the national stage yet, meaning the owners of this application have no enforceable rights yet. (Of course, a quick scan of the search report, page 25 of the published application, shows that the closest two references fail to directly disclose all of the elements of the broadest claims). The original developer provides some interesting evidence that the invention may have been directly copied, including the fact that the application omits a feature of the alogrithm that was added shortly after filing of the initial application. He has since backed-down from his more aggressive stance , seeming to be content with the explanation that Tandberg independently created the algorithm, but the timing will undoubtedly raise a few eyebrows.
Elsewhere, Jan Wildeboer comments that “This could result in the true inventor being forced to remove his own code from his own software to avoid patent infringement.” Well, not necessarily. Certainly, not yet since the application is still a long way from maturing into a patent. But even more importantly, assuming the “true inventor” can demonstrate publication of the software prior to the date Tandberg claims to have invented it, they will have a tough time getting anything allowed other than any improvements they may have made to the software, similar to the IBM patent, above. In addition, Tandberg’s knowledge of a prior invention is now publicly documented. At least in the United States, if they try and obtain a patent without directing the Examiner to this prior work, they run the risk of having their patent rendered unenforceable for violating the duty of disclosure.
I’ll have more to say later about a potential solution that would help identify situations similar to these, even where the public doesn’t take notice.
Even if there does appear to be clear prior art, it’s not as if the Patent Office seems willing, capable or interested in actually understanding how these patents are simply copying what’s already being done elsewhere.
RESPONSE: Not really a fair characterization of the patent office. The majority of Examiners genuinely want to do a good job with the limited resources that they have, and tend to be very liberal in asserting similarities between teachings until they applicant gives them a good reason to believe otherwise.
Has this sort of thing happened before? Is this something open source developers should be worried about? How can developers stop patent trolls from stealing their work?
RESPONSE: Not sure what “trolls” have to do with stealing inventions, but how can developers prevent this from happening? First, you can’t always. Sometimes people do evil stuff… sometimes its an innocent mistake that no one catches. But when you are alerted to a pending application and you believe your own published work is material to the patentability of the claimed invention, one thing you can do is send the information and evidence of its prior publication to the inventor. This may obligate the inventor to alert the Examiner about the existence of the prior art, which could end up having the application rejected, or narrowed substantially.
At a time when governments around the world are trying to tighten the intellectual property regime, aiming to stop piracy and plagiarism online, it’s still ridiculously easy for corporations to steal some kid’s code and try to get a patent on it.
RESPONSE: In fairness, Dana, the “kid” did put the code up on the web for anyone to see, so from that perspective, yes, it was easy. But as I mentioned above, no patent has issued yet, and it sounds as if Tandberg is aware of the situation and can investigate whether this was a co-incidental innovation by unrelated parties, or a true derivation.