Trial Notes 4/13/06 Part 5 Do NOT propagate
Stone’s rebuttal (he jumps around a bit to cover all major topics):
Regarding performance gap:
Lee: we started at 30 and ended at 800.
Was he wrong?
Regarding Rambus IP:
Stone shows a 1992 R corporate background document describing their inventions. Reads a few terms from doc: clock skew- PLL, Architecture overview, high performance PLL.
Kil memo from Nov 1992: New Rambus architecture containing: R channel, R interface, and R DRAM to achieve 500 MB/s data rates. It has precise control of clock by way of PLL. Gets 500 MB/s by synchronizing the rising and falling edge of a 250 MHz clock.
[I call this section some of the best evidence of all. Stone shows R describing their inventions in 1992, and then Kil generates a Hynix memo explaining the Rambus features. And of course Taylor has testified H doesn’t deal with rising and falling edges of the clock. LOL]
Steinberg – unfair attacked- Did job well and efficient.
Tricks:
Taylor question regarding CAS =2, must wait at least 1 clock.
The Hennessey performance gap chart.
Read timing diagram shown to Murphy. He couldn’t tell the difference. The second command, second group of data, and the warning early termination is illegal were removed from chart.
JEDEC made us do it.
JEDEC is not prior art. Came in 1999. Not relevant to R infringement.
Assume other company infringes – two wrongs do not make a right.
DLL- not just voltage control. Also tap control is DLL.
Preliminary decision by PO. It’s not preliminary. It’s THE decision. 5 examiners
Taylor – Murphy comparison:
Murphy did his homework. Had 7 binders. If he didn’t know the answer he would look it up. Taylor just didn’t know.
Back to access time reg.
Number stored [CAS latency] is representative. DDR code is different (uses half cycles). The CAS code stored does not represent the binary number (there are no binary half cycle codes) but is representative.
Royalty burden:
Weinstein testified to 3%. Chung in previous deposition said it was 8%.
The trial is about the people. (That’s it for Stone)
Furniss:
F&H are wonderful people. Not attacking them. We’re attacking the claims that were written much later.
We’re not attacking Karp and Steinberg. Just what they did.
Murphy couldn’t answer the question how the patents increased DRAM speed. [There was a pained look on Whyte’s face at this moment as he looked up. Mine also. This was probably the weakest part of Murphy’s performance. Why the hell couldn’t he explain DRAM speed increases when using the various elements of Rambus IP??? It’s not that difficult!]
JEDEC – I didn’t tell you jedec was prior art. H didn’t decide (on their own) what the DRAM will look like.
Chung 8% royalty burden was error. 3% is correct.
F goes over verdict forms. Makes joke about double negatives (=pos) and double positives (=pos) – Yeah Right.
There was a bit of a jury reaction when he asked them to place a zero on damages.
On performance gap:
We’ve heard R solved the performance gap. Not one word is true. R tells over and over they solved the performance gap. It’s very effective, but not true. Henneccy said it, and R said it. The gap is widening.
Teece states R provides uncommon value – they solved the performance gap. His opinion has not weight. He bought marketing slogan. Wrong on perf gap assumption. His conclusions are likely to be wrong.
Pieces of invention are all prior art.
DLL- fundamental but shown only in a figure.
It had to do with making money. Want more money, write more claims.
Performance gap: a marketing slogan
Justice- R had revolutionary idea. It’s called RDRAM. They have what they’re entitled to. 3.5% is huge amount of money. Not justified. It’s not justice. Don’t buy marketing slogan.
Finished!
