How do you spell relief? Not RMBS

[Robert Myers]

http://www.eetimes.com/semi/news/OEG20040217S0007

o 3.6Gbits per second per pin
o No multi-drop bus
o No Rambus

This scheme, not Yellowstone, will be how Intel gets the enormous
memory bandwidth it predicts it needs by the end of the decade.

RM

 

[chrisv]

http://www.eetimes.com/semi/news/OEG20040217S0007

o 3.6Gbits per second per pin
o No multi-drop bus
o No Rambus

This scheme, not Yellowstone, will be how Intel gets the enormous
memory bandwidth it predicts it needs by the end of the decade.

What scares me, is that it's likely that many future technology
changes like this will be shaped not only by technical considerations.
Rather, a major goal will be to think-up something that some other
idiot hasn't already thought-up, patented, and looking to hold the
world hostage for.

 

[Robert Myers]

What scares me, is that it's likely that many future technology
changes like this will be shaped not only by technical considerations.
Rather, a major goal will be to think-up something that some other
idiot hasn't already thought-up, patented, and looking to hold the
world hostage for.

People have been trying to do that in the computer business
essentially forever. Interoperable standards that really work are,
IMHO, no older than the PC business. With such a large commodity
market, people who decide to do their own thing had better be prepared
really to go it alone--a very difficult thing to do in a truly
commodity market.

People have an exaggerated fear of Intel. As has been repeatedly
demonstrated, Intel cannot do whatever it damn well pleases. It
operates as part of an ecological system to which it cannot
unilaterally dictate, much as it might like to.

Witness the failure to impose Rambus and the failure to hold back a
move to a 64-bit ISA for x86. And this particular development
happened in cooperation with two memory makers (_not_ another pure IP
company--lesson learned).

It is absolutely true: Intel has played games with interconnect
standards and is headed toward a proprietary on-board interconnect.
Some of the game-playing Intel has done with interconnect standards
has gored _my_ ox and held back probably for years the development of
true commodity high-speed low-latency board to board interconnects.

It is _very_ clear that Intel would like to extend its proprietary
influence beyond the proprietary CPU interconnect as far as it
possibly can. It has even said as much

http://www.computerworld.com/networkingtopics/networking/story/0,10801,90031,00.html

Try as it might, it won't succeed. Anybody who tries and succeeds is
a niche market player and isn't going to hold the whole world hostage.

In the end, as George Macdonald has correctly pointed out, Intel's
business model is to sell zillions of whatever it is. It's really
hard to do that and hold the whole world hostage at the same time.
IBM mastered the art of charging enormous premiums for
highly-specialized boxes, and even IBM can't even come close to
getting by on that kind of business anymore. The closest anyone has
come in the PC era is Microsoft, and Microsoft's empire is crumbling,
even with the help of some of the sleaziest business tactics I've ever
seen from another player in the industry who need not be named.

RM

 

[Stacey]

What scares me, is that it's likely that many future technology
changes like this will be shaped not only by technical considerations.
Rather, a major goal will be to think-up something that some other
idiot hasn't already thought-up, patented, and looking to hold the
world hostage for.

Well this isn't the only industry this happens in. Saginaw owns the patent
for the planatary gearing used in most automatice transmissions and Honda
developed a different type of automatic to avoid these patents. The early
versions had all sorts of weird problems and always shifted "weird" but did
work around the patents.

These latest patent wars remind me of the early days of the net when people
bought up blocks of domain names and then would sell the rights to use them
to companies whose names they registered before the company did.

 

[George Macdonald]

Well this isn't the only industry this happens in. Saginaw owns the patent
for the planatary gearing used in most automatice transmissions and Honda
developed a different type of automatic to avoid these patents. The early
versions had all sorts of weird problems and always shifted "weird" but did
work around the patents.

Interesting. I'd have thought that any patents on the planetary gear
system would have expired a while back... or do they keep adding new
patents nobody can live without?

Rgds, George Macdonald

"Just because they're paranoid doesn't mean you're not psychotic" - Who, me??

 

[Stacey]

Interesting. I'd have thought that any patents on the planetary gear
system would have expired a while back... or do they keep adding new
patents nobody can live without?

This was back in the mid seventies and was told this by the guys at Honda
Tech. I guess by now Honda has gotten this pretty sorted out and just stuck
with this design. It's really pretty neat, looks like a manual tranny with
clutch packs where the syncro hubs would normally be.

 

[Robert Myers]

This was back in the mid seventies and was told this by the guys at Honda
Tech. I guess by now Honda has gotten this pretty sorted out and just stuck
with this design. It's really pretty neat, looks like a manual tranny with
clutch packs where the syncro hubs would normally be.

So now we have two design approaches instead of one. Inconvenient for
Honda in the short run, but conceivably a benefit in the long run.
You never know unless you try.

What I really don't understand, and it would drag the group way off
topic, is how the drug manufacturers have finessed this problem so
handily to the benefit of all (more drugs, faster pace of discovery),
and the electronics industry just seems to get buried under it.

Somebody comes out with a molecule that acts through a particular
metabolic pathway or attaches to a particular receptor. Next thing
you know, you've got a half dozen molecules, all obviously using the
same basic mechanism, all patented, and all competing with each other.
One absorbs faster or slower. One shows a slight difference in the
effectiveness on particular sets of symptoms or has a slightly
different set of side effects. No problem with the prinicpal of
equivalence, apparently. Patent granted, enforceable, the drug is
marketed, and life goes on. Why doesn't the electronics industry work
that way?

RM

 

[Stacey]

Somebody comes out with a molecule that acts through a particular
metabolic pathway or attaches to a particular receptor. Next thing
you know, you've got a half dozen molecules, all obviously using the
same basic mechanism, all patented, and all competing with each other.
One absorbs faster or slower. One shows a slight difference in the
effectiveness on particular sets of symptoms or has a slightly
different set of side effects. No problem with the prinicpal of
equivalence, apparently. Patent granted, enforceable, the drug is
marketed, and life goes on. Why doesn't the electronics industry work
that way?

Lawyers?