« Bad Deceiver | Main | Senate Patent Gutting Bill »

August 5, 2006

Selective Memory

Reportage is a tricky business. Situations are often not as they first appear. Corruption being endemic, intrinsic to human nature, news and history often prevail from power, either to enforce or deny. On that happy note, we turn to the domestic computer DRAM market, which seems to be a large-scale exercise in manipulation all around.

Last week, Rambus was being hung out to dry by the Federal Trade Commission (FTC), reversing an earlier finding. The issue is whether Rambus was obliged to tell the computer memory standards committee at the time, JEDEC, that it had applied for patents which, ultimately, the standard would embrace, thus offering Rambus the windfall of extracting patent royalties from memory manufacturers.

The title of my first blog entry on the second Rambus FTC decision was "Sneaky Rambus." That was contradicted in a knowing comment by an inside observer, who thought "sneaky FTC would be a more appropriate title," outlining political pressure exercised by other memory makers to rig the gallows for Rambus at the FTC on its second bite.

The FTC is currently preparing to concoct a remedy for Rambus' transgression of non-disclosure to JEDEC, which could be forcing Rambus to a big payback of reaped royalties.

But the FTC has no hold on the courts. It's not axiomatic that the FTC ruling destroys Rambus' juggernaut of patent enforcement.

The judge in the Rambus-Hynix case, Ronald M. Whyte, noted in his January finding of fact that there was no reason for Rambus to think it had been obliged to disclose to JEDEC its then-current patent applications. In 1993, Gorden Kelly, JEDEC committee chair, stated that JEDEC did not require disclosure of patent applications, and that Kelly's employer, IBM, had no intentions of disclosing its patent applications to JEDEC.

In Rambus v. Infineon Technologies, Infenon had argued that Rambus' patents were unenforceable due to Rambus duping JEDEC, but the CAFC threw cold water on that. In Rambus v. Micron Technologies, the special master to the Delaware district court trying the case agreed that JEDEC policy didn't require patent application disclosure.

As noted earlier, antitrust and price fixing in the memory market are currently being investigated by the U.S. Justice Department, with price-fixing guilty pleas already extracted from former Hynix Semiconductor executives, Hynix being one of the companies Rambus successfully put the squeeze on for patent infringement. Rambus has stated that the JEDEC issue would have limited impact owing to earlier summary judgments in the case.

The third phase of the trifurcated Rambus v. Hynix trial looms. In this phase, Rambus defends itself against Hynix's allegations of monopolization, unfair competition, and outright fraud, in the DRAM market.

Micron is also going after Rambus for monopolization & unfair competition, in a complaint filed in 2000, headed for trial in December, in Delaware district court. Micron's case is firmly rooted in the JEDEC affair, as Micron accuses Rambus of only being able to charge for patent royalties resultant from deceiving JEDEC. Naturally, Rambus' counterclaim is patent infringement.

Rambus had sued Samsung for infringement, but backed off when Infineon managed to get a ruling against Rambus, for unclean hands in shredding documents, in the same district (Northern California) where Rambus was tackling Samsung. Weirdly, during the Samsung spat, there were allegations that Rambus had hired a Samsung in-house attorney to advise it about suing Samsung, while the attorney still worked for Samsung. In the aftermath, Samsung bickered over having Rambus pay its attorneys fees, but the district court judge scotched that last month.

Collusion is common in oligopolies, as is intense backbiting. If one accepts a Darwinist viewpoint, the legal imbroglios in the memory market seem, perversely, natural. Makes for hellishly dizzying blogging though.

Posted by Patent Hawk at August 5, 2006 1:01 AM | Litigation