Monday, November 10, 2008

John Danforth

Posted with permission of John Danforth:

Let me say up front that I am a former Rambus employee – indeed, a former senior Rambus lawyer – and that I do have an ax to grind. That said, there are ongoing references to Rambus in the press and in fora like this that I feel ought to be corrected. I do not speak for Rambus, but I still have a sufficient interest in its future to want to begin correcting some of the misstatements I see. 

There are at least five things deeply wrong with the references to Rambus here.

First, Rambus invented technologies critical to modern computers and other electronic systems. Its efforts to protect those inventions ought to be praised, not derided. As our economy moves further and further from a manufacturing base, it is all the more essential that inventiveness be encouraged, rewarded and protected. 

Second, there is nothing wrong – in the law or morally – with doing what Rambus did with certain of its patents, i.e. amending patent claims to better cover infringing products of which one becomes aware after the relevant patent applications are first filed. Indeed, the law has long expressly approved of exactly such amendments. The key case is called Kingsdown.

I want to expand on this point because it lies at the heart of many of the misstatements that are made about Rambus

There are many good reasons for allowing inventors to amend patent claims after their original patent filing. One is that we do not want to discourage prompt patent filings by making initial patent filings unduly expensive or a trap for the unwary. Inventors are to be encouraged to disclose their inventions as fully and a soon as possible in a part of the patent application called a “specification”. This disseminates new knowledge broadly and quickly, which is what the patent system is, in part, designed to do. Inventors -- especially individual inventors, new inventors, or small or young companies -- may not always know at the outset what elements or combinations of elements in their specifications later prove to have been valuable or to have been novel when invented. They are therefore permitted to add new patent claims later. (Patent claims are distinct from the specification. They legally define what within a specification is asserted as being subject to patent protection.) The key limit on this is whether or not those new patent claims are adequately supported by the patent specification as originally filed. 

(I like to illustrate this concept with a very rough analogy. Think of the map of a newly discovered territory. It is logical that if a government wants new territories discovered and described as quickly as possible then it will therefore reward the discoverers with some form of protected ownership interest in the new area so long as its basic parameters are adequately described. Then, within those parameters, is also logical that a government (as with our patent system) permits smaller areas to later be claimed and protected if they fall within (and could reasonably have been predicted to fall within) the original map.)

Third, there is ample evidence that the infringement of Rambus’ intellectual property was not accidental or isolated. It was repeated and it has escalated (by more products and more companies over the past ten-plus years) – even as Rambus has brought patent lawsuits and its patents have been found valid and infringed. 

This record speaks volumes. It reflects the fact that Rambus’ inventions are the best solutions available. And there is ample evidence that the companies that created standard memory and memory controllers knew that they were using Rambus’s work – and knew that they were risking substantial patent royalties if Rambus was able to amend and/or defend its claims. This known, eyes-wide-open assumption of risk is documented by internal documents of the key standard setting organization (JEDEC) and key manufacturers. 

Fourth, it is fundamentally untrue to call Rambus a company that is simply in the business of creating patents. 

This characterization is untrue because Rambus designs (and was founded in 1990 to design) leadership products (like the XDR DRAM memory used in the Sony Playstation3 and the RDRAM used in Playstation2) that it believes are the best use of its inventions. Rambus has said that it wants to control the use of its inventions. Its strong stated preference is to have them used in its own designs – i.e. RDRAM, XDR and XDR2. Other uses of Rambus’ patented technologies (e.g., in SDRAM. DDR, GDDR, DDR2, DDR3, etc.) are a blatant taking of Rambus’s property and are, in fact, uses of that property that compete directly with Rambus’ own designs. (A computer, workstation, HDTV, game system or smart phone designed to use SDRAM, DDR, DDR2, DDR3, etc., is, by definition, not going to use RDRAM or XDR.) That explains, I believe, why there are now very solid, currently-pending requests by Rambus (in federal court and – as of last week -- also in the ITC) to enjoin the unauthorized use of Rambus IP

Finally, this characterization of Rambus as a patent-only company is also, given Rambus’ history, deeply unfair. Rambus must,to be sure, rely significantly on its patents, at least for the time being. But this is because its own leadership products have – strong evidence suggests -- been the target of a successful and illegal DRAM Company conspiracy to drive Rambus from the marketplace. That conspiracy appears to link directly to the now-admitted DRAM price-fixing conspiracy that recently rocked the industry, resulted in record industry fines paid to the Justice Department, and sent many industry executives to jail. Internal DRAM company documents – showcased on the front page of the Wall Street Journal -- show this connection. This history of concerted “Rambus killing” (to quote one DRAM industry executive) was also partly revealed in a four month FTC trial in 2003. It will likely soon be revealed much more fully in a San Francisco courtroom, in a major antirust case filed in 2004 by Rambus against key DRAM manufactures. That case is set for trial this coming March. Stay tuned.

John Danforth
Personal Blogs - Blog Top Sites