Not so dead, *dead*, DEAD!

[George Macdonald]

I doubt it's very hard...I mean

What the hell does that mean? PCI-e is easy stuff and you don't need the
standards defn to design IP for it?... or you find other ways to get the
defn without paying and participating in the standards process?

I've talked personally with a fair number of folks at Rambus, and it
might surprise you at how good they are at circuit stuff. There were
several IHVs who worked with Rambus, not for their IP, but for their
design expertise to do DDR2 stuff. Now you have to sit back and ask,
why this would be...the answer is that they have tremendous experience
with high speed circuit, and a lot of folks don't. The company makes
money and has very good products; they wouldn't have been selected for
the PS1-3 if they just made crap.

"made"?? They don't "make" anything. The word out of Intel was that they
were good at designing "negative margins".

Now, their products are specialized
and there was an attempt to shove them into a niche they didn't
fit...but that failed, partially due to rather illegal cartel action by
the DRAM vendors.

The important thing is that they didn't get the "niche" they wanted. Your
"cartel" allegations I addressed elsewhere and I won't repeat.

Similarly, I suspect a fair number of folks are interested in their
expertise (and the IP generated from said expertise) for I/O.


OK, why don't you start then by telling us which of their patents are
bogus? I have no doubt that a portion (say 5-15%) might be sketchy;
however, EVERYONE gets sketchy patents today.

Well they started with ~50 violation claims against the memory mfrs - at
last count they were down to four. IOW there's a much greater authority
than I who has declared "bogus" and it seems like it would be easier to
start from the opposite direction: which ones were actually good? IOW your
5-15% is umm, inaccurate. As for "sketchy", as I said, the patent system
is broken... thus we ended up with approvals for a crock of shit like a
programmable count-down register.

 

[Del Cecchi]

I figured you'd know I was referring to their patents on DLLs and
count-down registers, etc... to do with memory interfaces. Just
because
the court threw most of them out does not exonerate them. I'd also be
interested to know what you think of their declared intention to
globally
chase down DDR signalling as their IP? Then there was RSL - I couldn't
figure out if they just wanted to patent the voltage or the
oscillographs?


Yeah well you'd have gotten the same retort if you'd tried that on with
me
before your "credentials" were "established". E.g. I know that
Daytripper used to post under his own name and figure that he came
across
some good reason to adopt an alias. OTOH the previous poster is
apparently
just an anonymous sniper... without evidence to the contrary another of
those Usenet dilletantes.

I haven't looked at their patents but rather at some of their circuit ip.
From reading web site it seems like they at least know what the problems
are in doing high speed interfaces (serdes etc). I haven't looked much
at DDR. But some of the rambus folks go back to the 80's or early 90's,
as in the days of DASH at Stanford. They were publishing while we were
working on SCI, before some of this stuff became common knowledge. It is
an easy mistake to say "that isn't patentable, everyone knows that"
without looking at the date of filing. I am not saying that is the
situation but it is easy to do. If you have a particular patent you are
incensed about I will provide my opinion for free, if you post the
number. And you will get your money's worth. I am a circuit
designer not a lawyer. But I have been involved in a few patents and can
provide my personal opinion as "one skilled in the art".

del

 

[Yousuf Khan]

Funny thing is: if all the DRAM mfrs/vendors had switched to DRDRAM they
might just as easily have been accused of (prosecuted for) price gouging
for adopting an unnecessarily high-priced technology for no effective gain
to consumers.

It wouldn't have mattered if it was DDR or RDR that emerged as the
victorious standard. The DRAM manufacturers would've still conspired
with each other to fix prices.

Yousuf Khan

 

[David Kanter]

I don't think so, they could simply claim it was better technology...

Define better. There was no evidence to back up that claim and in some
instances it was worse... and Intel knew it.

There are any number of definitions, better WRT bandwidth or latency
scaling. Higher internal clock speed...it doesn't really matter.
Ultimately, if it came to court (which it likely wouldn't), the DRAM
vendors could simply say it was a better solution...not everyone might
agree, but in general courts tend NOT to challenge/overrule technical
decisions. Simply put, most courts trust in companies to be better
technical experts than lawyers and judges...

Of course, if evidence to the contrary came out...then things might be
different.

But I am not talking about any of the evidence you are thinking of; I
mean stuff like internal memos stating that DRDRAM is a bad
technically, but represents a great opportunity to gouge customers.

David

 

[David Kanter]

What the hell does that mean? PCI-e is easy stuff and you don't need the
standards defn to design IP for it?... or you find other ways to get the
defn without paying and participating in the standards process?

Well, Rambus certainly has PCIe products. I'd guess that they
principally concentrate on the high speed interconnects, rather than
PCIe controllers. I know for sure that a major graphics company is
using their PCIe technology/products...

"made"?? They don't "make" anything. The word out of Intel was that they
were good at designing "negative margins".

Way to address 1/3 points I made (and the least technical). Moreover,
you're wrong. Their EPS has been at about $0.20 for the last few
years...not that it matters much.

The important thing is that they didn't get the "niche" they wanted. Your
"cartel" allegations I addressed elsewhere and I won't repeat.

Well, your addressing left a good bit to be desired.

Well they started with ~50 violation claims against the memory mfrs - at
last count they were down to four. IOW there's a much greater authority
than I who has declared "bogus" and it seems like it would be easier to
start from the opposite direction: which ones were actually good?

It seems to me that you are probably rather unfamiliar with the
american legal system. Rambus claimed 50 violations, and now they are
claiming 4 violations. That has nothing to do with whether the patents
are valid (since they weren't struck down, and you certainly have not
indicated that), it has to do with the DRAM vendors using techniques in
those patents.

Moreover, in patent, lawyers aim to maximize the number of
infractions/violation that stick. Since the violations that stick are
a subset of those that are made, lawyers will in effect maximize the
claims they made and include rather...spurious ones. That's simply
part of their job in representing their clients, and has nothing to do
with technology. IOW, if that interesting anecdote had any meaning, it
is more of a commentary on legal proceedings than anything else. Now
if they had patents struck down, that's a different issue (but you
never said they were struck down, rather that there were found to be
fewer violations).

IOW your
5-15% is umm, inaccurate. As for "sketchy", as I said, the patent system
is broken... thus we ended up with approvals for a crock of shit like a
programmable count-down register.

Well, that's the government's faults...and everyone takes advantage of
it. Rambus is equally guilty along with IBM, AMD, Intel, HP,
Microsoft, Eolas and every other company that has patented something.

David

 

[David Kanter]

I doubt it's very hard...I mean

What the hell does that mean? PCI-e is easy stuff and you don't need the
standards defn to design IP for it?... or you find other ways to get the
defn without paying and participating in the standards process?

Rambus has PCIe products, and I know for a fact that there is at least
one major graphics company using their products:
http://www.rambus.com/products/pciexpress/index.aspx

"made"?? They don't "make" anything. The word out of Intel was that they
were good at designing "negative margins".

You've managed to address 1 point out of the 3 I made...congrats. It's
also the least technical, and you're wrong. Their EPS has been around
20-30 cents for the last two years. I could look farther back, but I'm
too lazy.

The important thing is that they didn't get the "niche" they wanted. Your
"cartel" allegations I addressed elsewhere and I won't repeat.

I don't really think you did address them satisfactorily. Not only do
the DRAM manufacturers have a prior history of antitrust violations,
but Samsung already plead guilty to artificially raising the price of
RDRAM. If you'll recall the main complaint of most users was that
RDRAM was too expensive...funny that.
http://www.informationweek.com/showArticle.jhtml?articleId=174402774

Well they started with ~50 violation claims against the memory mfrs - at
last count they were down to four. IOW there's a much greater authority
than I who has declared "bogus" and it seems like it would be easier to
start from the opposite direction: which ones were actually good? I

Actually, that has nothing to do with patents being valid or not, that
has to do with patent violation claims. Believe it or not, those are
two different things.

Besides, this is a lawsuit which has nothing to do with technical
stuff. The way lawsuits work is that you stick in as many claims as
you can make, and that way if some get thrown out it's ok. IOW, claims
get thrown out all the time, and the job of a good lawyer is to
maximize the number of claims that stick; this means maximizing the #
of claims.

IOW your
5-15% is umm, inaccurate. As for "sketchy", as I said, the patent system
is broken... thus we ended up with approvals for a crock of shit like a
programmable count-down register.

You have provided no evidence that relates to 'bad patents', only
patent violations, which as I have said is distinct from a misbegotten
patent.

David

 

[David Kanter]

Later of course, Intel relented on DDR and adopted it. Rambus got

If you mean resisting being held up for ransom, yes the memory mfrs, apart
from a coupla wimps -- and the biggest one has recently reneged on their
"friendship" with RMBS -- wanted nothing to do with Rambus or their "IP".

Right, Sony had to search long and hard to actually find someone to
make all those DRDRAMs for the playstation...be serious. Anyway, they
did indeed resist, but apparently in a potentially illegal manner. Two
wrongs sure as hell don't make a right, and the fact of the matter is
that most of this stuff has yet to be decided. We do know that Samsung
plead guilty to price fixing on RDRAM, and several other manufacturers
have antitrust violations against them.

Now, is this totally damning evidence? No. Does it mean that Micron,
Infineon etc. violated antitrust laws? No. Does it mean that Rambus
violated antitrust laws? No.

However, it does mean that there is a very significant possibility that
all the parties involved violated the law. We won't really know until
the whole matter is done with (assuming the various matters aren't
sealed).

The timelines are quite clear for anybody who followed it, as well as the
web of deceit sewn by Rambus through a contorted mess of abandonments,
continuations & divisions of patent applications.
They *were* censured for
document shredding and even the judge who upheld their appeal against
Infineon remarked on their (lack of) business ethics.

That's just looking at one isolated part of the case. I could just as
easily say that Samsung is a bunch of scum sucking antitrust violators
b/c they plead guilty.

I think most outsiders would say that there are several lawsuits all
really inter-related and that until the last ones have been settled, we
won't know the whole truth. Rambus did shred some documents, but what
would you be saying if a memo came out from the executives of all the
DRAM vendors that basically said:

"We hereby agree to act together as a cartel to keep Rambus out of the
DRAM market, inflate the prices of their products to minimize traction
and gouge consumers as much as possible."

You're really jumping to conclusions without all the information, and I
really doubt any of us has a full understanding of what went on. I am
pretty confident that your perspective is very skewed, based on your
posts. However, I think the entire matter is somewhat unclear. To me,
it seems like both parties did some pretty bad stuff and we're just
going to have to wait and see what falls out.

The DRAM *manufacturers*/vendors fell over for the "guilty pleas" because
it was cheaper; Christ at the time they were accused of price fixing, they
were all losing money hand over fist.

Oh, and that's an excuse? Bullshit.

What the hell did the legal system
expect them to do?... just declare banruptcy?... keep making and selling a
product which had negative returns? IIRC there's one guy in prison because
he had erased notes on competitors' pricing. Now tell me that sales reps,
in any industry, don't have notes on what competitors are charging.<shrug>

There are four from infineon and one guy from Micron that I know of
now, and they appear to be executives:
http://www.techworld.com/storage/news/index.cfm?NewsID=2749

Rambus were scum; whether they've "reformed" we'll see - watch the lawyer
count. They *knew* they couldn't kill the DRAM vendors -- just look at the
investment required for a fab vs. RMBS capitalisation... not even in the
same ballpark. They just wanted to suck blood... at extortionate rates on
a per chip royalty basis.

I don't know what the rates were, and whether they were high or not.
Obviously, that's what the DRAM vendors say, but...taking their word on
that, is like trusting wolves to herd sheep.

There have been things that have happened and deals which have been made
which we'll probably never know about but here's a coupla little teasers
fer ya: how much did IBM pay to Rambus in licensing fees for the XDR IP
for Cell?... and why did nobody just buy RMBS up when it was at $4. or
so?

Well, I have no clue about either. I can't find any public info on
IBM's deal, but they wouldn't even need to charge to make money. They
might make all the money they need on the DRAM, rather than the
interface...

Also, I don't bother trying to explain investor behavior. It's far too
complicated and poorly understood.

DK

 

[David Kanter]

Later of course, Intel relented on DDR and adopted it. Rambus got

If you mean resisting being held up for ransom, yes the memory mfrs, apart
from a coupla wimps -- and the biggest one has recently reneged on their
"friendship" with RMBS -- wanted nothing to do with Rambus or their "IP".

Right, Sony had to search long and hard to actually find someone to
make all those DRDRAMs for the playstation...be serious. Anyway, they
did indeed resist, but apparently in a potentially illegal manner. Two
wrongs sure as hell don't make a right, and the fact of the matter is
that most of this stuff has yet to be decided. We do know that Samsung
plead guilty to price fixing on RDRAM, and several other manufacturers
have antitrust violations against them.

Now, is this totally damning evidence? No. Does it mean that Micron,
Infineon etc. violated antitrust laws? No. Does it mean that Rambus
violated antitrust laws? No.

However, it does mean that there is a very significant possibility that
all the parties involved violated the law. We won't really know until
the whole matter is done with (assuming the various matters aren't
sealed).

The timelines are quite clear for anybody who followed it, as well as the
web of deceit sewn by Rambus through a contorted mess of abandonments,
continuations & divisions of patent applications.
They *were* censured for
document shredding and even the judge who upheld their appeal against
Infineon remarked on their (lack of) business ethics.

That's just looking at one isolated part of the case. I could just as
easily say that Samsung is a bunch of scum sucking antitrust violators
b/c they plead guilty.

I think most outsiders would say that there are several lawsuits all
really inter-related and that until the last ones have been settled, we
won't know the whole truth. Rambus did shred some documents, but what
would you be saying if a memo came out from the executives of all the
DRAM vendors that basically said:

"We hereby agree to act together as a cartel to keep Rambus out of the
DRAM market, inflate the prices of their products to minimize traction
and gouge consumers as much as possible."

You're really jumping to conclusions without all the information, and I
really doubt any of us has a full understanding of what went on. I am
pretty confident that your perspective is very skewed, based on your
posts. However, I think the entire matter is somewhat unclear. To me,
it seems like both parties did some pretty bad stuff and we're just
going to have to wait and see what falls out.

The DRAM *manufacturers*/vendors fell over for the "guilty pleas" because
it was cheaper; Christ at the time they were accused of price fixing, they
were all losing money hand over fist.

Oh, and that's an excuse? Bullshit.

What the hell did the legal system
expect them to do?... just declare banruptcy?... keep making and selling a
product which had negative returns? IIRC there's one guy in prison because
he had erased notes on competitors' pricing. Now tell me that sales reps,
in any industry, don't have notes on what competitors are charging.<shrug>

There are four from infineon and one guy from Micron that I know of
now, and they appear to be executives:
http://www.techworld.com/storage/news/index.cfm?NewsID=2749

Rambus were scum; whether they've "reformed" we'll see - watch the lawyer
count. They *knew* they couldn't kill the DRAM vendors -- just look at the
investment required for a fab vs. RMBS capitalisation... not even in the
same ballpark. They just wanted to suck blood... at extortionate rates on
a per chip royalty basis.

I don't know what the rates were, and whether they were high or not.
Obviously, that's what the DRAM vendors say, but...taking their word on
that, is like trusting wolves to herd sheep.

There have been things that have happened and deals which have been made
which we'll probably never know about but here's a coupla little teasers
fer ya: how much did IBM pay to Rambus in licensing fees for the XDR IP
for Cell?... and why did nobody just buy RMBS up when it was at $4. or
so?

Well, I have no clue about either. I can't find any public info on
IBM's deal, but they wouldn't even need to charge to make money. They
might make all the money they need on the DRAM, rather than the
interface...

Also, I don't bother trying to explain investor behavior. It's far too
complicated and poorly understood.

DK

 

[David Kanter]

Later of course, Intel relented on DDR and adopted it. Rambus got

If you mean resisting being held up for ransom, yes the memory mfrs, apart
from a coupla wimps -- and the biggest one has recently reneged on their
"friendship" with RMBS -- wanted nothing to do with Rambus or their "IP".

Right, Sony had to search long and hard to actually find someone to
make all those DRDRAMs for the playstation...be serious. Anyway, they
did indeed resist, but apparently in a potentially illegal manner. Two
wrongs sure as hell don't make a right, and the fact of the matter is
that most of this stuff has yet to be decided. We do know that Samsung
plead guilty to price fixing on RDRAM, and several other manufacturers
have antitrust violations against them.

Now, is this totally damning evidence? No. Does it mean that Micron,
Infineon etc. violated antitrust laws? No. Does it mean that Rambus
violated antitrust laws? No.

However, it does mean that there is a very significant possibility that
all the parties involved violated the law. We won't really know until
the whole matter is done with (assuming the various matters aren't
sealed).

The timelines are quite clear for anybody who followed it, as well as the
web of deceit sewn by Rambus through a contorted mess of abandonments,
continuations & divisions of patent applications.
They *were* censured for
document shredding and even the judge who upheld their appeal against
Infineon remarked on their (lack of) business ethics.

That's just looking at one isolated part of the case. I could just as
easily say that Samsung is a bunch of scum sucking antitrust violators
b/c they plead guilty.

I think most outsiders would say that there are several lawsuits all
really inter-related and that until the last ones have been settled, we
won't know the whole truth. Rambus did shred some documents, but what
would you be saying if a memo came out from the executives of all the
DRAM vendors that basically said:

"We hereby agree to act together as a cartel to keep Rambus out of the
DRAM market, inflate the prices of their products to minimize traction
and gouge consumers as much as possible."

You're really jumping to conclusions without all the information, and I
really doubt any of us has a full understanding of what went on. I am
pretty confident that your perspective is very skewed, based on your
posts. However, I think the entire matter is somewhat unclear. To me,
it seems like both parties did some pretty bad stuff and we're just
going to have to wait and see what falls out.

The DRAM *manufacturers*/vendors fell over for the "guilty pleas" because
it was cheaper; Christ at the time they were accused of price fixing, they
were all losing money hand over fist.

Oh, and that's an excuse? Bullshit.

What the hell did the legal system
expect them to do?... just declare banruptcy?... keep making and selling a
product which had negative returns? IIRC there's one guy in prison because
he had erased notes on competitors' pricing. Now tell me that sales reps,
in any industry, don't have notes on what competitors are charging.<shrug>

There are four from infineon and one guy from Micron that I know of
now, and they appear to be executives:
http://www.techworld.com/storage/news/index.cfm?NewsID=2749

Rambus were scum; whether they've "reformed" we'll see - watch the lawyer
count. They *knew* they couldn't kill the DRAM vendors -- just look at the
investment required for a fab vs. RMBS capitalisation... not even in the
same ballpark. They just wanted to suck blood... at extortionate rates on
a per chip royalty basis.

I don't know what the rates were, and whether they were high or not.
Obviously, that's what the DRAM vendors say, but...taking their word on
that, is like trusting wolves to herd sheep.

There have been things that have happened and deals which have been made
which we'll probably never know about but here's a coupla little teasers
fer ya: how much did IBM pay to Rambus in licensing fees for the XDR IP
for Cell?... and why did nobody just buy RMBS up when it was at $4. or
so?

Well, I have no clue about either. I can't find any public info on
IBM's deal, but they wouldn't even need to charge to make money. They
might make all the money they need on the DRAM, rather than the
interface...

Also, I don't bother trying to explain investor behavior. It's far too
complicated and poorly understood.

DK

 

[George Macdonald]

I haven't looked at their patents but rather at some of their circuit ip.
From reading web site it seems like they at least know what the problems
are in doing high speed interfaces (serdes etc).

Well they bought a SerDes company a coupla years back and have picked up
related assets from other companies - not sure how much they acquired in
IP & expertise in those transactions and how much they knew before.

I haven't looked much
at DDR. But some of the rambus folks go back to the 80's or early 90's,
as in the days of DASH at Stanford. They were publishing while we were
working on SCI, before some of this stuff became common knowledge. It is
an easy mistake to say "that isn't patentable, everyone knows that"
without looking at the date of filing. I am not saying that is the
situation but it is easy to do. If you have a particular patent you are
incensed about I will provide my opinion for free, if you post the
number. And you will get your money's worth. I am a circuit
designer not a lawyer. But I have been involved in a few patents and can
provide my personal opinion as "one skilled in the art".

I don't think it's necessary to go digging through the Rambus patent
porfolio - sorting through all the mess of abandonments, continuations,
divisions etc. To take a simple case, as a circuit designer, how would you
feel if they came after you to tell you "you cannot use a programmable
count-down register in that application because it's our IP". That is
precisely what they tried with the memory mfrs. Correct me if I'm wrong
but I'd have thought that a count-down register was a pretty standard
engineering device.

Similarly for DDR, it seems like a simple question: does Rambus own the
concept of DDR signalling or not? They did announce at one point that they
would go after every such implementation for their err, pound of flesh.
This would mean even things like the ATA disk interface.

In general I feel uncomfortable with the situation where people keep one
foot in academia and the other in a highly IP-oriented business and then
play the patent game for millions of $$ in personal gain. I've seen the
grant system in action and the scope for abuse is umm, bothersome to me.
E.g. if the "man from NSF" doles out research $$ to a prof. who uses it to
develop IP, should he then be allowed to go hold up entire industries for
ransom?... or should that IP be automatically in the publc domain?... or
belong uniquely to the academic organization? It would certainly be
interesting to know who/what was the original sponsor of Prof. H's
research.

 

[George Macdonald]

It wouldn't have mattered if it was DDR or RDR that emerged as the
victorious standard. The DRAM manufacturers would've still conspired
with each other to fix prices.

Yeah and if that was conspiring, what *were* they doing when I paid ~$250.
for 32MB back in 1996/7.

To return to the original subject, I wonder how AMD feels about paying out
largish IP licensing fees to a company whose chairman is a former employee.
16/17 years is a long time but all the same.....

 

[Del Cecchi]

Well they bought a SerDes company a coupla years back and have picked up
related assets from other companies - not sure how much they acquired in
IP & expertise in those transactions and how much they knew before.




I don't think it's necessary to go digging through the Rambus patent
porfolio - sorting through all the mess of abandonments, continuations,
divisions etc. To take a simple case, as a circuit designer, how would you
feel if they came after you to tell you "you cannot use a programmable
count-down register in that application because it's our IP". That is
precisely what they tried with the memory mfrs. Correct me if I'm wrong
but I'd have thought that a count-down register was a pretty standard
engineering device.

A count down register is a standard device and is not patentable as
such. But say someone you for example, invents a new way to do
something by using a programmable counter. Say it is a way to
automatically parallelize scalar code at run time, for argument's sake.
And one of the ways to implement your invention employs a count-down
register. Should your invention be not patentable because it uses
something that is common knowledge? Or is the fact that a piece of
common circuitry is used be irrelevant? To carry what seems to be many
folks problem to an extreme, these circuits are all built out of
transistors, and transistors have been around for years so circuits
aren't patentable? No rational person would say that. So why would a
rational person like yourself say "it uses a counter. counters are well
known. therefore it is not patentable"?

The issue in DDR I would venture is knowing when to sample the incoming
data. This information was formerly provided by a clock edge. If it
isn't then how do you figure it out? Is the diclosed apparatus or
method novel and not obvious to one "skilled in the art" (this does not
mean an expert by the way)?

Similarly for DDR, it seems like a simple question: does Rambus own the
concept of DDR signalling or not? They did announce at one point that they
would go after every such implementation for their err, pound of flesh.
This would mean even things like the ATA disk interface.

Do they "own" DDR signaling on all interfaces? Memory Interfaces? No
Interfaces? Do they own a particular style of signaling on some devices?

In general I feel uncomfortable with the situation where people keep one
foot in academia and the other in a highly IP-oriented business and then
play the patent game for millions of $$ in personal gain. I've seen the
grant system in action and the scope for abuse is umm, bothersome to me.
E.g. if the "man from NSF" doles out research $$ to a prof. who uses it to
develop IP, should he then be allowed to go hold up entire industries for
ransom?... or should that IP be automatically in the publc domain?... or
belong uniquely to the academic organization? It would certainly be
interesting to know who/what was the original sponsor of Prof. H's
research.

The ownership of IP developed under government or academic funding is a
subject of negotiation. The existence of such IP is obvious. Through
your elected representatives NSF or DARPA or University policies towards
ownership of IP can be changed, if there is a concensus that they should
be.

You really shouldn't use such loaded terms as "hold industry for
ransom". If someone invents something they, or someone like their
employer, are legally entitled to the benefits of that invention. The
fact that it is inconvenient for someone else is of limited relevance.
CDMA is patented and it is inconvenient to some folks in the Cell Phone
equipment industry but you don't seem to be upset over that.

 

[Del Cecchi]

What the hell does that mean? PCI-e is easy stuff and you don't need the
standards defn to design IP for it?... or you find other ways to get the
defn without paying and participating in the standards process?

go to http://www.pcisig.com/specifications/order_form and pay your
money. A cd with everything is 1500 USD. Individual docs are 100 USD.

"made"?? They don't "make" anything. The word out of Intel was that they
were good at designing "negative margins".




The important thing is that they didn't get the "niche" they wanted. Your
"cartel" allegations I addressed elsewhere and I won't repeat.




Well they started with ~50 violation claims against the memory mfrs - at
last count they were down to four. IOW there's a much greater authority
than I who has declared "bogus" and it seems like it would be easier to
start from the opposite direction: which ones were actually good? IOW your
5-15% is umm, inaccurate. As for "sketchy", as I said, the patent system
is broken... thus we ended up with approvals for a crock of shit like a
programmable count-down register.

See my other post.

 

[Christy_z1]

George Macdonald wrote:
"The timelines are quite clear for anybody who followed it, as well as
the
web of deceit sewn by Rambus through a contorted mess of abandonments,
continuations & divisions of patent applications. They *were* censured
for
document shredding and even the judge who upheld their appeal against
Infineon remarked on their (lack of) business ethics. "

Regarding the document shredding, the courts have found that there is
no spoilation. The courts have also found no unclean hands. If you have
an open mind read the following, but based on branding Rambus as scum,
I am pretty sure you will not look at any of the links. In your eyes
Rambus is a worse company than the DRAm manufactureres, even though the
cartel has been convicted in courts of price fixing, and some of the
employees are going to jail.

http://investor.rambus.com/downloads/Uncleanhandspdf.pdf

Regarding JEDEC, the courts have found that JEDEC rules were not clear
cut, and indeed Rambus did reveal the 703 patent on time.
http://media.corporate-ir.net/media_files/NSD/RMBS/msoA7__.pdf


If anybody is really open and would like to learn about all this
history, read the following. I am pretty sure most of you will not.
This is the decision by the Judge in the FTC case. It covers all the
topics in every detail regarding this whole Rambus saga.
http://www.ftc.gov/os/adjpro/d9302/040223initialdecision.pdf


Before you start spewing bunch of false statements, know all the facts.

 

[Robert Redelmeier]

Regarding the document shredding, the courts have found
that there is no spoilation.

"In the decision Rambus Inc. v.Infineon Technologies AG, 2004
WL 547 (E.D. Va. Mar 14, 2004), the United States Court of
Appeals for the Federal District ruled that Rambus was guilty
of spoliation of documents because it carried out their records
retention policy without regard to anticipated litigation. This
decision significantly broadened the standard for spoliation. In
the past, bad faith was the key determinant for spoliation. The
implication of the Rambus decision is that even if a company
destroys records in good faith (as part of their standard records
retention policy), they can still be sanctioned for spoliation
if they destroy records relevant to anticipated litigation."

Perhaps RMBS didn't expect and is aggrieved by this ruling,
but there it is.

In your eyes Rambus is a worse company than the DRAm
manufactureres, even though the cartel has been convicted in
courts of price fixing, and some of the employees are going
to jail.

Ad-hominem. I don't much like DRAM mfrs since some fixing must
have been done to keep DRAM at $40/MByte for almost a decade.
But that doesn't stop me from disliking RMBS for submarine patents.
Marketing crimes versus technical crimes. The former are easier
to fix. The latter get buried in silicon.

http://investor.rambus.com/downloads/Uncleanhandpdf.pdf

Regarding JEDEC, the courts have found that
JEDEC rules were not clear cut, and indeed
Rambus did reveal the 703 patent on time.
http://media.corporate-ir.net/media_files/NSD/RMBS/msA7__.pdf

I'm afraid I find this far less biased and more credible:
http://www.abanet.org/antitrust/committees/intell_property/sep24report.html

-- Robert

 

[George Macdonald]

There are any number of definitions, better WRT bandwidth or latency
scaling.

It offered marginally better bandwidth in *some* situations and even with a
CPU designed around that *and* two memory channels, the observed gain was
just not there. As for "latency scaling", depends on how far you want to
scale but from the POV of what actually happened and was adopted, nope you
have it wrong there.

Higher internal clock speed...it doesn't really matter.
Ultimately, if it came to court (which it likely wouldn't), the DRAM
vendors could simply say it was a better solution...not everyone might
agree, but in general courts tend NOT to challenge/overrule technical
decisions. Simply put, most courts trust in companies to be better
technical experts than lawyers and judges...

But you would disallow them the benefit of that :technical" judgement
against a parasitic blackmailer? At that point Rambus had no credibility
in the industry and thought they could piggyback their tax on the thousands
of existing memory-based patents.

 

[George Macdonald]

Rambus has PCIe products, and I know for a fact that there is at least
one major graphics company using their products:
http://www.rambus.com/products/pciexpress/index.aspx

There's no mention of any "major graphics" supplier there. From what I see
they seem to be present as an embedded(?) PHY supplier and if it works and
the price is reasonable, good luck to them. Let them compete with the
others in that market.

You've managed to address 1 point out of the 3 I made...congrats. It's
also the least technical, and you're wrong. Their EPS has been around
20-30 cents for the last two years. I could look farther back, but I'm
too lazy.

What do you not understand about info from inside Intel to do with
"designing" How could you possibly take that to mean anything to do with
stock value? "Negative margins" was a reference to the timings of the
DRDRAM channel as recounted by a former Intel employee. I can give a
reference but if you're too lazy.........

You know you need to get a grip David, and quit firing off knee-jerk junk
posts -- two to one of mine is hardly reasonable. I'm not going to bother
even reading the first one in this series.

As for your 3 points, why should I "sit back" on command by you and swallow
your personal anecdotal, technical evaluations.

I don't really think you did address them satisfactorily. Not only do
the DRAM manufacturers have a prior history of antitrust violations,
but Samsung already plead guilty to artificially raising the price of
RDRAM. If you'll recall the main complaint of most users was that
RDRAM was too expensive...funny that.
http://www.informationweek.com/showArticle.jhtml?articleId=174402774

I thought we lived in a captalist system - mfrs have the right to pick &
choose the technology they adopt. If I'm the only guy selling something
I'd figure I have a right to charge whatever the market will bear; if the
market is short of something, like e.g. refined petroleum, you pay more.
Rambus' whinging on this issue is pathetic; one assumes they must
fab-outsource for prototype parts - let them do similarly for production...
or build a fab. They have been documented as behaving unethically in
business - people would rather not have done business with them... that's
how it works. So Samsung copped a plea to end the angst... BFD... but
another enemy for Rambus... the only "friend" they had in the memory
business.

Actually, that has nothing to do with patents being valid or not, that
has to do with patent violation claims. Believe it or not, those are
two different things.

Besides, this is a lawsuit which has nothing to do with technical
stuff. The way lawsuits work is that you stick in as many claims as
you can make, and that way if some get thrown out it's ok. IOW, claims
get thrown out all the time, and the job of a good lawyer is to
maximize the number of claims that stick; this means maximizing the #
of claims.

50:4 is a poor ratio by any standards - at that level it reeks of a
plundering exercise gone wrong. They should have been censured for wasting
the court's time with frivolous claims - in many cases a court would have
been correct in rejecting the whole 50, on principle.

You have provided no evidence that relates to 'bad patents', only
patent violations, which as I have said is distinct from a misbegotten
patent.

Facts are facts. The courts are there to hopefully correct the
inadequacies/incomptence of the patent system. If a court will not stand
behind a claim the validity of the patent is certainly in doubt. D'oh, it
cannot be enforced.

 

[George Macdonald]

Right, Sony had to search long and hard to actually find someone to
make all those DRDRAMs for the playstation...be serious. Anyway, they
did indeed resist, but apparently in a potentially illegal manner. Two
wrongs sure as hell don't make a right, and the fact of the matter is
that most of this stuff has yet to be decided. We do know that Samsung
plead guilty to price fixing on RDRAM, and several other manufacturers
have antitrust violations against them.

Boutique products tend to cost more. A manufacturer of any ilk has the
right to say to someone who wants to sell them a "new idea": "no thanks, we
don't like it - go away". As a sole supplier Samsung had the right to
charge what it wanted -- that's how supply/demand works -- and, despite its
reputation for semiconductor R&D excellence, it *did* have trouble making
the stuff as evidenced by the situation Dell and its customers found
themselves in.

Now, is this totally damning evidence? No. Does it mean that Micron,
Infineon etc. violated antitrust laws? No. Does it mean that Rambus
violated antitrust laws? No.

However, it does mean that there is a very significant possibility that
all the parties involved violated the law. We won't really know until
the whole matter is done with (assuming the various matters aren't
sealed).

People take plea bargains all the time... avoids the uncertainty and all
the reversals & appeals process of court time.

That's just looking at one isolated part of the case. I could just as
easily say that Samsung is a bunch of scum sucking antitrust violators
b/c they plead guilty.

I think most outsiders would say that there are several lawsuits all
really inter-related and that until the last ones have been settled, we
won't know the whole truth. Rambus did shred some documents, but what
would you be saying if a memo came out from the executives of all the
DRAM vendors that basically said:

"We hereby agree to act together as a cartel to keep Rambus out of the
DRAM market, inflate the prices of their products to minimize traction
and gouge consumers as much as possible."

Rambus was *not* *in* "the DRAM market". If Rambus really had wanted to
enter it, they were free to do as any number of other fabless companies do
with a "better idea"... or find a better class of VC to fund a fab.
Instead they were intent on hijacking it with a per chip tax. Now tell me
you wouldn't say the same to someone who wanted to elbow his way into your
product market.

You're really jumping to conclusions without all the information, and I
really doubt any of us has a full understanding of what went on. I am
pretty confident that your perspective is very skewed, based on your
posts. However, I think the entire matter is somewhat unclear. To me,
it seems like both parties did some pretty bad stuff and we're just
going to have to wait and see what falls out.

I'm err, pretty confident I have it right.

Oh, and that's an excuse? Bullshit.

Umm.... HOSS-SHIT!

There are four from infineon and one guy from Micron that I know of
now, and they appear to be executives:
http://www.techworld.com/storage/news/index.cfm?NewsID=2749

Yeah puir wee things - they should have done as Rambus and left the
chicanery, document shredding and email deletions to the legal dept.

I don't know what the rates were, and whether they were high or not.
Obviously, that's what the DRAM vendors say, but...taking their word on
that, is like trusting wolves to herd sheep.

A per chip tax is/was absurd on a commodity part. Would Rambus have
consented to a negative tax when DRAMs sell below cost?... which happens
pretty regularly.

Well, I have no clue about either. I can't find any public info on
IBM's deal, but they wouldn't even need to charge to make money.

Well that's precisely my point!! Sounds a bit fishy to me: give it away to
one guy... charge his neighbor an arm & a leg. Is that legal?... ethical?

They
might make all the money they need on the DRAM, rather than the
interface...

Also, I don't bother trying to explain investor behavior. It's far too
complicated and poorly understood.

C'mon, at $4. a share it was a no brainer - the beast was only wounded.

 

[Christy_z1]

Robert Redelmeier wrote:
"Ad-hominem. I don't much like DRAM mfrs since some fixing must
have been done to keep DRAM at $40/MByte for almost a decade.
But that doesn't stop me from disliking RMBS for submarine patents.
Marketing crimes versus technical crimes. The former are easier
to fix. The latter get buried in silicon. "

It doesn't change the fact that, one group of companies have committed
a crime, unlike Rambus. Also they are not submarine patents, but I am
not going to get into that. You didn't even bother to read one of the
links I posted.


"I'm afraid I find this far less biased and more credible"

You didn't even check out the links did you Which Judge are you
calling biased and not credible? Or are you calling the Appeals court
decision not credible and biased? Please check the links again, they
are not Rambus propaganda, they are actual court documents. I know you
won't even look at them ....

"

"In the decision Rambus Inc. v.Infineon Technologies AG, 2004
WL 547 (E.D. Va. Mar 14, 2004), the United States Court of
Appeals for the Federal District ruled that Rambus was guilty
of spoliation of documents because it carried out their records
retention policy without regard to anticipated litigation"

This is wrong, where did you find this? the writ of mandamus was
denied. The court of appeals never looked at the spoliation issue
directly on an appeal. Get your facts straight.

http://www.ftc.gov/os/adjpro/d9302/040818applsctorddenymandamus.pdf

There is only one court decision regarding spoliation and it is the
first link I posted in my previous post.

 

[Robert Redelmeier]

It doesn't change the fact that, one group of companies
have committed a crime, unlike Rambus.

For which they paid the price. I'm not even sure
Infineon was ever charged or plead out.

Also they are not submarine patents, but I am not going to
get into that.

I'm glad you conceed.

"In the decision Rambus Inc. v.Infineon Technologies AG,
2004 WL 547 (E.D. Va. Mar 14, 2004), the United States
Court of Appeals for the Federal District ruled that
Rambus was guilty of spoliation of documents because it
carried out their records retention policy without regard
to anticipated litigation"

This is wrong, where did you find this? the writ of mandamus
was denied. The court of appeals never looked at the spoliation
issue directly on an appeal. Get your facts straight.

http://www.ftc.gov/os/adjpro/d9302/040818applsctorddenymandamus.pdf

Since you like the FTC, try this later link:

http://www.ftc.gov/os/adjpro/d9302/050513commodrgrntpartccmocompel.pdf

You sound like a Rambus shill. I admit they may have had
some interesting technology. But they greedily overplayed
their hand.

-- Robert