Sorry to see you go Nocturnal mouse catcher. You and CB and other dear departed souls used to liven up the place. I think this page has lost a lot of that fun-factor-thing.
Today, I went to the yahoo site and saw the following post. I think our discussions here are pretty boring and bland without CB's colorful metaphors. The day when you can read a more informed post in yahoo than in SI's famous penguin/eagle page is the day when you know we suck big time. Can't blame you for giving up the ghost. Heck, I'm spending more time reading the joke page than this one now. Tumbleweeds keep rolling on. Maybe I should just check back every saturday or sunday from now on. It's not like people still post here in any decent volume.
.......
Re: Rambus' Dependent Claims
by: investor_in_red 03/23/02 09:24 pm
Msg: 368253 of 368260
bump_n_fuzz,
You obviously don't understand patent law and claim construction. Once a person invents a concept of an electric motor, you cannot patent an electric motor in a blender just because you limited the scope the use of a motor. Likewise, SCI is a general-purpose interface that is meant to connect multiple processing units with multiple memory and peripheral devices. Once this GENERAL idea that applies to ANY electronic interface is discussed or published in a public forum, the idea becomes public domain. Only the originators of the idea can apply for the patent within one year of the publication. What Rambus did was to apply the ideas described in the SuperBus/SCI programme (e.g. use of both edges of clock, current steering drivers with voltage reference, programmable registers, use of PLL/DLL for signal de-skewing, etc.) adding highly multiplexed bus and a serialized data transmission scheme. What they ended up with is a DRAM implementation of SCI specification, more or less. Is this simple application of the public domain ideas from SCI to a DRAM interface a revolutionary invention? I don't think so. Then, there is this issue of intentionally omitted reference to SCI information in the original and earlier patent applications. Had the USPTO known about the SCI effort, I seriously doubt the Patent Office would have granted many of Rambus' patents.
But for the sake of argument, let's assume the Rambus' patents can survive both the prior art challenge and Walker Process fraud for intentionally withholding a prior art reference.
And let's look into claim construction.
Re: Patent Claims.
All of the patent claims you listed are "dependent" claims. These dependent claims are not new inventions by themselves - they merely describe possible extensions to the independent claim that they are built on. To infringe on a dependent claim, the offending technology must also infringe on the independent claim that the dependent claim is based upon. If you are familiar with the Rambus application PCT/US/9102590, you'll see that all the independent claims contain the limitations such as:
"...said bus carrying device-select information without the need for separate device-select lines connected directly to individual memory devices."
(See page 65: l2.espacenet.com 116680A1+I+ )
Guess, what.
Both SDRAM and DDR SDRAM have separate device select lines connected directly to individual memory devices.
Therefore, the independent claim does not apply; hence the dependent claims do not apply either. QED.
I would love to end the argument right here and ride into the sunset. Except that PCT/US/9102590 is not a granted patent, but merely an abandoned original application. In the patents in suit #5953263, #5954804, #6032214, and #6034918, these limitations were moved from the "claims" section to the written description section. Not being privy to the prosecution file history of these patents, I don't know why they were moved, but I am fairly certain they still limit the scope of most of the Rambus' patents. (Remember, Judge Payne's Markman decision?)
One last thing. You may be interested in this article on the recent trends in patent litigation at appellate level:
alston.com |
