Hi Zeev Hed; Re: "The court will indeed make the decision on validity of later "Sdram covering" patents, but whether the teachings were intrinsically in the 1990 specification (not the claims since that one never went to "claims"). Since neither you or I have access to that 1990 specification, neither of us can really say with any expectations of rationality what the court will decide."
First of all, since the claims in the 1990 document are not public, Infineon (or you and I) cannot be held to infringe them. The only purpose of the 1990 filing is to establish a time against "prior art", not to establish the boundaries of the invention. So the fact that we cannot see the 1990 filing doesn't prevent us from analyzing the limits of Rambus' patent territory. It does prevent us, perhaps, from seeing what Rambus in 1990 felt that their patent territory was, but if it literally included SDRAM in 1990, and then they had to file the 1992 patent which literally does not include SDRAM (because of the Bus definition, along with the clocking and other definitions narrowing the claims), then prosecution history estoppel would prevent them from later claiming that territory they had given up to get the patent. But I don't that that's the case, otherwise Infineon would have argued that, as they did in the case of the device ID (see later in this post).
Perhaps it would be good here to quote from Infineon's Markman argument, maybe you never bothered to read it. You should note that this is the statement by the winning party on the Markman decision, not that of the loser:
...
The claims of a patent, like land deeds or public laws, serve an important public notice function by defining and limiting the patentee's statutory rights to exclude. Patent claims must unambiguously put competitors (such as Infineon) on notice of what subject matter is covered by the patent and what subject matter is not. See Sage Products. Inc. v Devon Indus., Inc., 126 F.3d 1420,1425 (Fed. Cir 1997) (the patentee, not the public "must XXX the cost of its failure to seek protection for this foreseeable alteration of its claimed structure"). Given the important public notice function of patents, the patent statute requires that every patent contain a "written description" of the invention 35 U.S.C. $112, P112; see also Tronzo v. Biomet, Inc., 156 F.3d 1154, 1159-60 (Fed. Cir. 1998).
The relevant sources to be considered by the Court when construing a patent claim are (i) the language of the claim, (ii) the specification, and (iii) the prosecution history Markman, 52 F.3d at 979. These three "intrinsic" sources form the public record upon which competitors are entitled to rely to determine the meaning and scope of patent claims. In that regard, the Federal Circuit has explained that "it is only fair (and statutorily required,) that competitors be able to ascertain to a reasonable degree the scope of the patentee's region to exclude. They may understand what is the scope of the patent owner's rights by obtaining the patent and the prosecution history - 'the undisputed public record'" Id. at 978 (internal quotations and citations omitted). Expert testimony and other extrinsic evidence may be used as an aid to understanding the patent, "but not for the purpose of varying or contradicting the terms of the claims." Id.
...
Rambus -- not its competitors -- prosecuted the patent ex porte at the Patent & Trademark Office ("PTO"). Thus it was Rambus' obligation to provide claims and a written description of the invention that adequately satisfy the notice function. See 35 U.S.C. $112, P1,2. Indeed, the Federal Circuit has repeatedly held that where two reasonable interpretations of a claim term exist, the narrower interpretation favoring the competitor must prevail in light of the notice function of claims. See, e.g. Digital Diometrics vs. Identrix, Inc., 149 F.3d 1334, 1344 (Fed. Cir. 1998); Athletic Alternatives, Inc. v. Prince Mfg., Inc., 73 F.3d 1573, 1581 (Fed. Cir. 1996) (to conclude otherwise would "undermine the fair notice function" and $112, P2.)
...
In this regard, the Federal Circuit has routinely rejected attempts by patentees to construe patent terms broadly during litigation where the express disclosure of the invention in the patent specification is narrower. For example, in Toro, the patentee urged a broad dictionary definition of the common terms "cover" and "including" used in its patent claims, arguing that the defendant's narrower definitions read limitations from the specification into the claims and improperly limited the claims to the preferred embodiment. 199 F.3d at 1301-02. The Court rejected the patentee's argument ...
...
The specification may also limit the scope of a claim term where the written description "expressly distinguishes over prior art" solutions, as in the Rambus patents. O.I. Corp. v. Tekmar Co., Inc, 115 F.3d 1576, 1581 (Fed. Cir. 1997). In O.I., the Federal Circuit construed the term "passage" to exclude smooth-walled, completely cylindrical structures even though the limitation was not expressly in the claim. Id. at 1581.
...
Finally, the prosecution story can likewise assist in construing the scope of patent claims. Southwall Tech., Inc. v. Cardinal IG Cu., 54 F.3d 1570,1576 (Fed. Cir. 1995) ("Arguments and amendments made during the prosecution history of a patent application and other aspects of the prosecution history ... must be examined to determine the meaning of terms in claims."). Distinguishing remarks may have a limiting effect regardless of whether they are directed to the identical claim at issue, or whether the remarks were the reason for any PTO action during prosecution. Id. at 1578-79 (noting "clear distinction" between relying on prosecution history to construe claim and prosecution history estoppel); see also Watts, 232 F.3d at 883 (concluding arguments made to distinguish other non-asserted claims also applied to asserted claim) The touchstone is again what a reasonable competitor would understand from the remarks. See Digital Biometrics, 149 F.3d at 1347 ("The public has a right to rely on such definitive statements made during prosecution and noting that "[n]otice is an important function of the patent prosecution process.").
...
IV. Infineon's Proposed Claim Constructions For The Patents-In-Suit
A. Background Of The Patents-In-Suit
All four of the patents-in-suit originate from a common 1990 patent application, and consequently share the same specification and written description. For that reason, the description of the background of the patents-in-suit is combined in a single section. To the extent that certain features described below relate exclusively to the asserted claims in a particular patent, that patent is identified.
...
The patents all distinguish the disclosed invention from prior art devices that do not interface to the new bus of the disclosed invention (the multiplexed bus). The patent begins by describing that traditional (prior art) computer memories are designed to work in parallel, with each memory connecting to only a small number of the data lines ('918 patent, col 1, lines 29-47). In contrast, the disclosed invention of the Rambus patent uses a multiplexed bus where every device is connected to all the bus lines. In that regard, the Rambus patents make great efforts to distinguish prior art devices that interface to conventional buses having separate address, control and data lines, and/or having point-to-point connections such as device select or chip select lines. ...
...
Tellingly, nowhere in the specification does Rambus disclose that the invention can or should be used with the prior art dedicated bus or point-to-point line architectures. In fact, the Rambus patents expressly distinguish their invention from these prior art architectures and stress the advantages of the Rambus multiplexed bus over this prior art. ...
...
During the prosecution of the '525 Application, Rambus expressly limited its claims to only those devices that contain device ID registers to overcome the PRO's prior art rejections. The claims originally presented in the '525 Application were rejected by the Examiner several times as anticipated by or obvious in view of prior art patents. (See Exhibit XXX.) To overcome these rejections, Rambus presented new claims -- including application claim 211 (which ultimately issued as claim 26 of the '804 patent) -- during an examiner interview and in a subsequent amendment. (See Exhibit J.) Application claim 211 included the new limitation "integrated circuit device." In its remarks to the PTO with its amendment, Rambus argued that the new claims -- including application claim 211 -- were distinguished from the prior art on the basis that they all contained the device ID register and relevant interface and comparison circuitry limitations:
The new claims submitted in this Supplemental Preliminary
Amendment have been added to more definitely and fully protect
Applicant's invention. These newly submitted claims are directed
to a memory device (or an integrated circuit having memory)
having (1) an internal register for storing an identification value.
(2) interface circuitry to recieve a request on an external bus,
and (3) comparison circuitry to determine whether the
identification information in the rquest corresponds to the
identification value in the internal register -- wherein when the
identification information corresponds to the identification value,
the memory device responds to the request.
...
"[B]y distinguishing the claimed invention over prior art, an applicant is indicating what the claims do not cover." Spectrum Intern., Inc. v. Sterlite Corp., 164 F.3d 1372, 1378-79 (Fed. Cir. 1998) (internal quotations and citations omitted). Here, by distinguishing its claims over the prior art, Rambus has indicated to the public that its claims do not cover devices without a device ID register and relevant interface and comparison circuitry. Indeed, the prosecution record showing Rambus' intent to define the '804 patent claims with reference to the device ID register could not be clearer. Although Rambus in its remarks recognized that some of the '525 Application claims may be distinguishable from the prior art on other grounds, Rambus explained to the PTO that it was not necessary to distinguish them further in light of the device ID register and relevant interface and comparison circuitry limitations in all of the claims. (See Exhibit J at P. 12, fn 1.)
...
rambusite.com
The fact that no one got patents on SDRAM is due to a combination of things, in my opinion. (1) Since it was a JEDEC, you couldn't get royalties on it (if you were an honest company), so there was no financial reason to get the patent. You couldn't use the patent to stop competitors from making the stuff, and you couldn't charge for it. (2) All the individual elements of SDRAM were covered by prior art, as was the case with RDRAM, for that matter. Consequently the only patent you could get was a patent for the whole thing (just like with RDRAM), but since that was invented by JEDEC, rather than a group of individuals, you wouldn't have been able to properly list the inventors.
Now if SDRAM was an "obvious extension" of Rambus' 1990/1992 patents, then why did the 1990/1992 patents so repeatedly disclaim the difference between what was being claimed and other, non multiplexed, designs? Or, for that matter, why did Richard Crisp not believe that SDRAM was covered by Rambus patents? If Rambus' patents so clearly covered SDRAM, why didn't Rambus' employees think so at the time?
-- Carl
P.S. Where I couldn't make out the original, I've used "XXX" to describe the missing word. |
