Day 2
Firsthand report from mtsugawa.
========================
Day 2 of Infringment Trial - Part 1
by: mtsugawa
Long-Term Sentiment: Strong Buy 03/16/06 08:44 pm
Msg: 923740 of 923758
O.K., here we go again, I'll have to make this shorter than yesterdays notes,
otherwise I'll lose my job.
Arrived 7:35AM to a long line (20 people) in front of the courthouse waiting to
go through security! The line consisted of jurors, lawyers as well as
observers. TDOX, SS's Advisor and another investor were a couple of people in
front of me. Stone was immediately behind me in line and Nissly was behind him.
I made some small chat about security measures with Stone, but didn't dare
bring up Rambus or the trial.
Once past security and redressing, rode the elevator up with Stone and another
investor. When the elevator arrived on the 4th floor, the doors opened to a
completely packed hallway! This was at 7:45AM, 15 minutes before the trial was
to begin! I was now next to TDOX at the end of the line. Our chances of getting
a seat weren't looking too good.
Finally, they opened the courtroom doors and allowed the gallery and lawyers
in. By the time TDOX and I got into the courtroom, all the bench seats were
filled to capacity. The same people were present as yesterday (with the
exception of SDHawgMan). There were also a couple of new faces in the gallery.
The Bloomberg reporters were there as well.
So, TDOX, myself and three others were resigned to the fact that we'd probably
be asked to leave the courtroom once the trial started. TDOX went and asked the
bailiff if we could get some chairs and sit in the aisle. The bailiff went and
asked the court clerk's assistant and she gave him permission! So, I ended up
with the best seat in the house! A nice comfy chair all to myself right in the
front middle with a good view of both the jury, the video screen and the
charts/billboards used by the legal teams!
Rambus and Hynix legal teams same as yesterday. Stone and Farmwald for Rambus
and Brown, Nissly, Furniss, Chong (Korean attorney) and Jones(?) for Hynix.
8:00AM, clerk enters the court and asks the parties if there are any issues to
discuss outside the presence of the jury? Stone says that he has an issue that
needs to be addressed before Hynix's opening remarks. Clerk returns to
anti-chambers for a minute then returns and asks Stone if it can wait until
after he has finished his opening remarks. Stone says, 'yes'.
Jury enters the courtroom. Same nine as yesterday. Most are now smiling and
even chatting with the court clerks and amongst themselves. They seem to be
much more comfortable today than yesterday.
JW enters the courtroom asks Stone to finish his opening remarks.
Stone quickly reviews his presentation from yesterday. Mentions that the his
infringement analysis from yesterday (where he compared Hynix's product
specifications to Rambus's patent claims) showed 'literal' infringement. Stone
said that during the trial, Rambus will also show infringement using the
'Doctrine of Equivalence' theory.
Picks up where he left off yesterday, talking about what arguments he thinks
that Hynix will use to defend themselves - obviousness and inadequate
description. Shows the jury the same slide from yesterday containing the Patent
Examiner's (PE's) notes to show that the PE's realized that the description
described independent inventions. Highlights PE's note stating, 'access time
register is independent of the multiplexed bus'...).
Stone puts up a board titled, 'Patents Are Valid'. I couldn't read all the
bullet items, but here's the ones/parts I could read. You (or, better yet,
another attendee) can fill in the gaps.
- Patent Examiners are required to ...
- Need to solve ...
- Mark Horowitz & Mike Farmwald...
- Industry recognized that...
- Many companies signed licenses
As to obviousness, Stone shows some document from the prosecution history that
shows that the PE's cited 394 references to prior art. This means that they
looked at 394 different pieces of prior art and still decided that Rambus's
claims were novel and not obvious.
Hynix has listed 52 references to prior art in support it's claims of
obviousness. 30 of these 52 have already been looked at and rejected by the
PE's. 19 of them are redundant (same as others that have been rejected by the
PE's).
That leaves 3 references from the 52 that Hynix lists. One, called IAPX 432, is
owned by Intel (I think). Hynix will claim that this prior art solved the
memory bottleneck problem. So why was Intel, in 1999, still searching for
faster memory and it's engineers writing that the memory bottleneck was a 'key
problem' going forward?
1 of the references is to prior art owned by NEC. But NEC took a license from
Rambus. Why would they do so if this prior art rendered the Rambus patents
invalid?
1 of the references is not related to memory at all, so we can discard that
one.
The last reference is cited against only one of Rambus's patents regarding
'dual-edge clocking'. [I didn't catch Stone's reason for disregarding this
reference, he was going through this stuff pretty fast.]
Of the 52 prior art references cited by Hynix in support of their
non-obviousness defenses, none read on the 6 patents involved in this case.
[My comments: I didn't really buy into Stone's quick dismissals of these prior
art references. It seemed as if he was indiscriminantly eradicating whole
swaths of references (30 were already looked at by the PE's, 19 are redundant,
etc.) with little substantive support. Hopefully, the substance will be
provided during the trial. However, IMO, Stone's presentation here wasn't very
persuasive at all.]
Stone then says that the PE's job is to ensure that the claims cited in the
patent application are novel and not obvious. Hynix challenges whether the PE's
did their job correctly. If Hynix really believed that those 52 references read
on Rambus's patents, they could have taken up the issue with the PTO and filed
a petition...
Furniss objected. If this is going to the issue of repetitioning the PTO...
JW: Sustained.
Stone: The PE's did everything that they were supposed to do.
Stone then says that the industry recognized that F&H were pioneers in their
field. He shows an excerpt from the recent award given Horowitz from the IEEE
organization. "His research changed the way an entire industry thinks about
memory interfaces and..."(I didn't get a chance to jot down the rest of the
excerpt before Stone changed the slide).
Stone, this trial is about patents and claims and a lot of technical stuff, but
it's also about people. All trials are people. This is a trial about Mike
Farmwald and Mike Horowitz. Why would this college professor (MH) leave his
comfortable job to take on this huge and risky challenge? When it comes to
mental challenges, MH is fearless. He loves to solve problems people say cannot
be solved. Mike Farmwald is a creative thinker who thinks outside-the-box. [He
said more about MF, but I missed it.]
Stone then jumps to another topic and starts revealing some details about the
negotiation process with Hynix in 2000. This was interesting and new to me.
When Hynix received our notice of infringement, Mr. Chung (and Stone points to
one of the lawyers at the Hynix table) responded and asked for more information
on Rambus's patent licensing terms. There was much communication between Rambus
lawyers and Mr. Chung and other Hynix lawyers over the course of several
months. Never once did anyone mention that they thought that Rambus patents
were invalid.
Then, on Aug 26, 2000, Mr. Park (high-level exec at Hynix) sent a message to
Mr. Tate to arrange a meeting between him and Tate to 'finalize our business
agreement'. This was in reference to the SDRAM/DDR license. Tate agreed and the
meeting was set for Sep 6, 2000 in Korea. However, sometime before Sep 6,
another meeting took place, attended by 4 people. Who were these people? Well,
Mr. Park was one...
Furniss objects.
JW calls for a side bar.
JW, Furniss and Stone confer for a long time...4-5 minutes. [My notes: During
this break in the trial, I was watching the jurors and noticed that they were
much more comfortable today, talking and laughing amongst themselves. They had
bonded in a way. The two older ladies on the left end talked to each other and
the young woman and middle-aged man on the other side were talking and
laughing. The older man and young woman in the middle seemed to be a bit more
introverted. Back to the action.]
Stone returns and states, 'You will learn that during the negotiation process,
they never said that 1% for SDRAM or 4-4.5% for DDR DRAM was too much. Noone
ever said that during all the negotiations. Mr. Chung will testify that during
the months of negotiations, he never asked his superiors if they thought the
asking price was too high.
On Aug 29, 2000, with no warning to Rambus, Hynix filed this lawsuit against
Rambus. A few days later, Mr. Park sent another email to Mr. Tate simply
stating that "We reject your patents." [My note: That may not have been the
exact quote, but it something short and sweet like that. Basically, Park was
telling Tate to F^$k O#f! 'in the nicest sort of way' as TDOX would say.]
Stone then wraps up his remarks quickly, thanks the jury and sits down.
[My note: it was obvious to me that the sidebar had thrown Stone off of his
prepared presentation and the big closing to his opening remarks. Instead, his
opening remarks just kind of ended without any fanfare. Later, we come to know
the reason for this spiritless ending. (another tease)]
JW then dismisses the jury to the jury room. After the jury is gone, JW says,
"O.K. we have an objection from Rambus to a rather inflammatory exhibit that
Hynix has added to their opening remarks which had not previously been
disclosed. Mr. Furniss?"
Furniss: Your Honor, the reason we added this exhibit is to counter something
that Rambus said yesterday during their opening remarks that we were not told
about and they did not disclose. They are now claiming that they invented
synchronous DRAM. Rambus's ever-expanding inventions now include all
synchronous DRAM. They made that statement yesterday that we had not heard
before and was not disclosed.
This exhibit is deposition testimony from a witness that will be testifying
live later in the trial.
Stone: Your Honor, I never said that Rambus invented synchronous DRAM. I only
mentioned synchronous DRAM to set the context for dual-edge clocking.
JW: O.K., here's how I think it should be handled. You can't use the exhibit
but you may tell the jury what you think will be said during his testimony.
Furniss: Thank you, your Honor.
Stone: Thank you, your Honor.
There was another issue regarding Rambus's renegotiation of the Elpida license
in early 2004 (may have been 2003, since my notes show to references, one to
2004 and the other to 2003). Stone said that both parties stipulated that the
relevant negotiation time period was in year 2000. This renegotiation took
place in early 2004. JW asked about patents that issued after 2000? Stone said
those patent are not relevant. I think JW said that he would consider the issue
and let the parties know his stance prior to the damages phase of the trial.
I didn't really understand what this was all about, but Infringeon and another
investor explained during the break that it had to do with lower royalty rates
for the newly issued patents that were included in the renegotiation.
Time is now 8:45AM. Furniss will give Hynix's opening statements after the
break.
9:00AM. Jury is seated. JW enters and immediately begins, "Mr. Furniss, all set?"
Furniss approaches the jury box and say's "Good morning" directly to the jurors. They respond back. He then introduces himself. "My name is Dan Furniss representing Hynix". He actually bows to them as if he were in Korea or Japan. He then goes on to introduce each of the four
other Hynix lawyers, one-by-one. As each lawyer's name is recited, they stand and acknowledge the jury. It was very polite and kind of over-the-top. I'm not sure what the jury thought of it. Some probably were flattered while others probably thought he was pandering to them.
After the introductions, Furniss continues. You've heard Rambus's version of what this trial involves. I'm here to tell you "the rest of the story" (like that guy on the radio). He immediately establishes a connection with the jury.
[My note: Before I get into Furniss's remarks, let me give you my impression of Furniss' style and demeanor. He uses a lot of home-spun tales and familiar analogies to keep them interested and a bit entertained. He's a big, rough looking guy with a red, round face and slicked down hair and no neck, but he talks very softly and politely. I think sometimes he tried too hard to get the jury to react to his jokes/anecdotes and may have come off a bit condescending. It's was hard to read the jury. I definitely think he kept things moving much better than Stone did, his presentation materials were better and easier to follow than Stone's. All this had the effect of keeping the jury's attention throughout most of his presentation. His presentation was also shorter than Stones (about 1:30).]
Furniss continues his opening remarks. Tells the jury, "I have to be honest with you, you're going to have to answer some pretty tough, technical issues. In fact, because this case is so complex, I've done something that I've never done before...I've created a table of contents."
With that he pulls out this huge billboard listing all the major points in his presentation:
>Table of Contents
>> The Parties
>>> Hynix
>>> Rambus
>> Issues
>>> Infringement
>>> ??? [I missed this one]
>> Patents
>>> Specifications
>>> Claims
>> DRAM Industry
>>> DRAM Evolution
>> Rambus's Inventions
>> Rambus's Claims
>> Non-Infringement
>> Invalidity
>> Hynix
Shows a photo of a large memory fab in Eugene, Oregon. Tells jury that Hynix spent $1.7B to build this plant. This plant makes all the Hynix DRAM sold in the U.S. Hynix has other 'fabs' outside the U.S., but this one makes all the DRAM sold in the U.S.
Then shows a photo of two employee's from this plant. They are working in the 'clean room' and have on their 'bunny suits'. Tells jurors the number of employees, but I didn't catch it. I think it was in the thousands. [My note: He's trying to show that Hynix is an 'American' company that employs lots of Americans, not just a foreign company selling their products into America.]
>> Rambus
Rambus has no 'fabs'. They are an idea company. They sell their ideas. They have ~200 employees. Shows a slide from some Rambus financial presentation that states their business model is to not make the products, but to patent them and then sell licenses on the patents. Furniss says this is a very profitable model.
>> Issues >>> Infringement
Inventions cannot be infringed. Patent specifications cannot be infringed. Only patent claims can be infringed.
[I missed some stuff here.]
>> Patents >>> Specification: Sets scope of the invention.
>> Patents >>> Claims: Set boundaries of invention.
Builds on the analogy from the video shown at the start of the trial where they say a patent is like a 'deed of trust'. Says 2 realtors can make an agreement that says, "I'll pay you 2% for every house you sell in San Jose. If you sell a house in Willow Glen, Almaden Valley or (some other area of San Jose), you get your comission. But if you sell a house in Santa Clara or Gilroy or Milpitas (cities bordering San Jose), I don't have to pay you. It's
outside the boundary. Claims must be contained in the boundary of the specification." [I included this analogy to show Furniss's M.O. He uses familiar analogies to make the jury think that he's one of them. I thought it was effective at first, but got tired of it after a while.]
>> DRAM Industry >>> DRAM Evolution
Shows slides of a motherboard and it's components. Explains how DRAM works. How the data moves from the CPU, to the memory controller, to the memory and hard drive. Shows slide with chronological timeline:
- 1971 first synchronous DRAM invented
- 1991 EDO DRAM invented (JEDEC standard)
- 1995 SDRAM invented (JEDEC standard)
- 1998 DDR DRAM invented (JEDEC standard)
All of these DRAM's were standardized. Explains what standardized means. Says
standardization is needed in order to ensure one manufacturers memory will work in any
computer. Guarantees ample supply from multiple vendors. If you produce a non-standard memory, you wouldn't sell much of it. The computer makers determine what memory they want to standardize on, not the memory manufacturers. Napa Auto parts doesn't tell GM what engine parts they're going to sell. GM tells Napa Auto parts what parts they need to make if they want to stay in business.
Rambus claims that Hynix stole this invention, took that invention. Well, that's kind of half-true. Hynix simply made standardized DRAM. Talks about JEDEC. SDRAM was adopted in 1993. First introduced by Samsung in 1995. DDR adopted in 1997. If you wanted to sell any memory, you had to build JEDEC standardized products.
>> Rambus Inventions
Rambus did invent something. They invented RDRAM. They took a lot of existing technologies and added a couple of new ideas to create a pretty good product, RDRAM. It's like they built a new type of house and then tried to get patents for all hammers, nails and saws.
They believed that 70% of the industry would adopt RDRAM. RDRAM is still being used today. Rambus has made a lot of money off of this invention. The inventors have done very well.
Rambus real invention was a narrow, high-speed, multiplexed bus. Shows a diagram contrasting Hynix's old DRAM product (using a wide, non-multiplexed bus) with the Rambus RDRAM. Says that SDRAM and DDR DRAM do not have the narrow, multiplexed bus shown in the Rambus diagram, but an improved version of the old DRAM product.
>> Rambus's Claims
Reminds jury that you can only infringe a claim. Shows a slide titled, "Rambus's So Called Inventions"
Alledged Invention: Dual-Edge Clocking. Rambus didn't invent; Hynix doesn't use it. Quotes Redwine patent as prior art.
Alledged Invention: Programmable Block Size: Rambus didn't invent. Cites IAPX 432 and Budde prior art)
Alledged Invention: Access Time Register: Rambus didn't invent; Hynix doesn't use it.
Alledged Invention: DLL w/ Variable Delay Line: Rambus didn't invent; Hynix doesn't use it. Cites Lofgren patent as prior art.
You won't see any of these terms in Rambus's patent specification. Why? If these inventions were so radical and worth so much money, why are they not mentioned in the patent specification?
Rambus did invent RDRAM. RDRAM is simply a compilation of a bunch of existing technologies combined with a couple of new ideas. This time Furniss uses a high-speed train analogy.
Shows slide of high-speed train that's supposed to represent RDRAM. Says Rambus's so-called inventions are akin to the high-speed train inventor laying claim to patents on the Alloy Wheels, Rotating Headlight, Turbo Charger and Disc Brakes.
Furniss now shows a chart depicting the prosecution history of the Rambus patents-in-suit. This was a pretty powerful slide. It starts out on the left showing the single patent application from which all the other 54 patents were derived. As you move to the right, the more branches are added as the original application is split into multiple applications and each branch is further split by more splits and continuations. Some branches are eventually abandoned and others are created by filing more continuations. Furniss notes how the tree branches exploded once Karp and Steinberg joined Rambus.
Furniss: Why not just create a new tree instead of filing all these continuations? Because then they would lose their priority date. Explains why the priority date is so important.
>> Non-Infringement
Before beginning, Furniss asks jury if they'd like to take a break. I thought this was out of line since usually only the judge may ask a direct question to the jury. However, JW didn't say anything and the jury members didn't need a break so Furniss continued.
Used the Access Time Register claim to show how Hynix plans to prove it's products don't
infringe. Shows the Markman definition that states 'the amount of time a device must wait
before reading the data'. On the slide, the term, 'must wait' is highlighted. Furniss says that Hynix DRAMS do not have to wait the exact amount of time stored in the register. This is just one example.
>> Invalidity
- invention must be novel and must be adequately described in the specification
Let's look at dual-edge clocking or DDR. Ask yourself the following two questions? Is it
new? Is it described in the spec?
No mention of the term, 'dual-edge clocking' in the spec. Rambus points to one sentence where it states (something like) 'we get 500 bits from a 250 bit bus in one clock'. That's it. And they claim that that means the device has dual-edge clocking? If this invention was so important, why wasn't it described in more detail?
Is it novel? No. Shows slides of Redwine patent '852 from 5/18/1992. (Furniss, "Redwine is the name of the inventor, not a preference.") Shows that the patent specification explicitly uses the terms, 'rising clock' and 'falling clock'.
Shows the Novak patent from May 1987. Shows diagram clearly showing data being retrieved on both the rising and falling edges of the clock.
Furniss then mentions that the dual-edge clocking feature was used in other devices, not DRAM, but that Farmwald and Horowitz got their idea for dual-edge clocking from these patents while they worked at MIPS Semiconductor.
PLL & DLL. Furniss first shows a typical DLL diagram. Then shows a diagram from Rambus's
specification that supposedly contains the DLL invention. Shows that the DLL circuitry is actually a subset of the circuitry in the Rambus application.
Furniss says he's almost finished, but must address one more thing. He then produces the same chart that Stone used yesterday that shows eight DRAM manufacturers that chose to license Rambus's patents and Hynix as the lone company that chose not to license them. Furniss says Rambus wants you to believe that we're the only company that has not licensed their patents. They said, 'Hynix refuses to pay.' Well now I'm going to tell you the rest of the story.
Let's start with Hitachi. Hitachi also refused to pay Rambus, so Rambus sued Hitachi. Well, Hitachi doesn't like to pay lawyers, and I can understand why. So, they caved in and bought a license. However, they got out of the DRAM business within 6 months, so the license didn't really cost them much at all. Furniss crosses them off the chart.
He continues this way around the chart, crossing off the licensees. NEC, no longer in DRAM business. OKI minor player. Toshiba, big company, but not a big DRAM player. Samsung, they're no longer licensed. Rambus has now sued them using many of the same patents at issue here. Elpida, Mitsubishi, small players.
Let's look at the top 5 DRAM producers in 2000. Furniss lists Samsung, Micron, Hynix,
Infineon and one other (he couldn't remember the company) on the margin of the chart. Says that none of these companies have a license for these patents. So, in reality, we're not the only ones that refuse to pay Rambus. In fact, only the smaller manufacturers have licensed these patents. [The inference here is that Rambus bullied them into signing by threatening litigation.]
Furniss then talks about Karp and Steinberg. Tries to make it seem as if they just wanted to create as many claims as possible from the original patent application, even if the
specification didn't support the claims. Says, "these claims were afterthoughts". He called them, "inventive claims" invented by Karp and Steinberg. Karp and Steinberg never even spoke to F&H about their patents.
This case is a tough case and, as you know, a very important case. It's not trivial. Blah, blah, blah. Hynix deserves your verdict. Thank you.
JW: O.K., let's take a short 5 minute break. Jury starts to leave when Stone gets up and asks if he can approach the bench. JW waits until the jury leaves and then listens to Stone.
And this is where the fireworks start! Unfortunately, my notes are very sparce here. The exchange was going along at a very fast pace and I was somewhat in shock at how angry and animated Stone had become. I'll just try to give you the gist of the exchanges.
Stone: We agreed not to discuss why RDRAM failed. We stipulated that we would not discuss the JRA documents and then they go and kick the door wide open by telling the jury that RDRAM failed due to high royalty rates! And, he (pointing to Furniss) set me up during that last sidebar! He wanted to prevent me from telling the jury about the evidence we have that these guys colluded to drive Rambus out of the market! And then he goes and says RDRAM failed due to high royalty rates! That's not why it failed! It failed because they decided as a group
to boycott RDRAM and fix prices on DDR DRAM. I have evidence to prove this! We know that Mr. Chung and Mr. Brown (again he points directly to the Hynix lawyers at the table) were planning litigation to kill Rambus for months! We know that they all met on Aug 28th and then they both filed litigation the very next day, one in Delaware the other here. They were performing a 'pincher' on Rambus [My note: forcing Rambus to manage multiple simultaneous lawsuits at the
opposite sides of the country]
There's nothing you can say now to undo what he has told the jury. He has kicked the door
wide open! They tried to starve us out of existence!
Furniss has now moved over to Stone's side to try to intimidate him. Furniss is a fairly
large bulky man while Stone is stout, but short.
Furniss: RDRAM was not successful due to problems with manufacturing. Intel had problems getting their products to work correctly with RDRAM.
Stone is pacing along the jury box he is so angry!
JW: Well, it seemed to me that the inference you made was that RDRAM failed due to the
royalties.
Furniss now begins to shuffle his feet and stammer a bit. "I did not say that RDRAM failed. I was very careful in that regards. In fact, I said that RDRAM was a good product."
JW: Why shouldn't I allow them to explain why these guys didn't sign a license? Your
inference was that they didn't sign a license due to the high rate. It was an independent
decision by each company.
Furniss: What? Do you want to turn this into an antitrust case? There's no proof of any of that, your Honor. Just because there were some meetings and then they both filed suit around the same time doesn't prove there was collusion and a group boycott. Rambus has already filed an antitrust suit in San Francisco. These issues will be decided in that case.
Brown now gets up to support Furniss and hand him a note. They talk briefly.
Stone: We know that Chung wrote a letter to Steinberg on Aug 26, 2000 stating that they were ready to come to terms and setup a meeting for Sep 6 to finalize the agreement. So on Aug 26th, they're ready to come to terms and then they have this meeting and all of a sudden, without warning, on Aug 29 they file suit on multiple fronts.
Nissly now gets up and hands Furniss another note.
Stone: Appleton had also setup a meeting with Tate in Sep 06.
[My note: I wasn't able to capture a lot of the dialogue that went on here. It went
extremely fast and back and forth. I was a bit in shock at Stone's candor and unabashed
anger. He was almost yelling at Hynix's legal team. He kept saying, 'They (or he) set me up!' Basically, Stone wants JW to allow him to use the JRA documents to rebut Furniss's inference as to why RDRAM failed . Furniss claims he never said nor implied that RDRAM failed due to high royalty rates. Stone says, just review the transcripts.]
JW: Here's what I'm going to do. No reopenings at this time. We have 3 days now before the trial resumes. I'd like to review the transcripts to see exactly what transpired. I'd like see some short briefing on what evidence Mr. Stone would use to rebut this inference and Hynix's reply. My inclination at this point is to allow some sort of rebuttal.
JW then called for a short 5 minute recess. Stone, still steaming, was the first person out the door of the courtroom and immediately headed to the restroom to cool off. The Hynix team of lawyers scurried off to a private room to discuss their strategy.
This ended the opening statements phase of the trial. What an ending!
Next up, Rambus presents it's case. Their first witness is Mike Farmwald.
I wasn't able to stay for the witness testimony and probably won't be back until the closing arguments. I'd really love to be there on Monday if Stone is allowed to present the JRA evidence, but it just takes so much time and energy, I don't think I can do it. I'm exhausted.
So, good luck to everyone and I look forward to reading others court reports in the coming days.
Mark
===============================
