Monday, July 1, 2013

Property as Process: Nakamura's '839 Patent Family and What it Suggests about a Patent's Relationships to Time

A few months ago, I was doing some research into a particular LED patent issued to Shuji Nakamura (and assigned to Nichia Chemicals, his Japanese then-employer). This was one of the early blue LED patents (GaN, not ZnSe) and it played a role in Nichia's dramatic rise to Big 5 status in the LED industry. The patent is especially interesting because of how far it traveled (not that this is necessarily uncommon for patents nowadays). The first diagram below illustrates the connections between the '839 patent as found within the USPTO database. The patent was brought over from an original set of Japanese patent applications and therefore claims priority back to those JP applications (blue rectangles) for the various portions of its claims. Through a series of divisionals (based on the decision by the USPTO examiner that the claims in the US application actually refer to more than one "invention"), the original application resulted in a total of 8 US patents with the last one issuing in 2004.
 Zooming out from the USPTO perspective, however, we can see that this was not the extent of the patent's travels. By looking into the patent via the JPO, KPO, and the EPO's patent family search tool, we see the relationship between the original Japanese applications (along with the 5 patents they resulted in) as well as the 2 EU (taken into 2 German) and 4 Korean patents in its family.
 Finally, instead of taking the set of patents in a structural-style map, we can also see them over time in a way that gives insight into when Nichia's patent practitioners were working on which parts of this set of applications. The process viewed in this way reminds us how difficult it is for a patent that travels this far to be dealt with by a single patent practitioner. From the time of the first application (let alone the original technological work) in 1992 to the issuance of the final patent in its family in Korea in 2006 much had changed in Nichia. Nakamura, for instance, had long left the company for UC Santa Barbara, been sued by the company for trade secret theft, and won his own lawsuit over Nichia's failure to pay him adequate compensation for the value his technological work had brought to them.
Part of the take away from all of this is that we need to understand "property" not just in terms of what is owned or claimed, but also in terms of time. Patents especially are a sort of property with a very strange sort of relationship with time. Moreover, this is more than the difficulty of figuring out for certain when a patent's term is up. When I would ask my patent engineer informants how you could tell if a patent was a good one or not, they jokingly told me that you don't know until you use it in court, and win. Then how do you know if it has value? Well, you don't know that until it is used and upheld again. While the issuance of a patent (in most cases) means that the patent has already under gone (and passed) an examination to prove that it is "new," the interpretation of the patent law (in the US especially) may change several times over the course of the patent's life as the Supreme Court (or the Court of Appeals for the Federal Circuit) attempts to hone its guidance on what can and cannot be patented (and on a whole range of related rules and procedures). The value of a patent is thus predicated on a conditional future where it gets used and gets upheld. Yet, in the courtroom, the testimony that is most important for its validity (and thus its having or not having any value at all) is focused on a reconstructed past point in time where the technology was "invented" and the court's evaluation of the "state of knowledge of the art" at that time. Within itself, then, the patent references both past and future even as the process of claiming it may still be ongoing in jurisdictions in other parts of the world. Property is a process, not a state and understanding it takes time.
 


Creative Commons License
These diagrams by Matthew West are licensed under a