On Fri, 24 Sep 1999, Max More wrote:
> At 01:36 PM 9/23/99 -0700, Robert Bradbury wrote:
> >On Thu, 23 Sep 1999, Max More wrote:
> Robert, please be careful with your attributions. I did not *write* the
> above--I *quoted* Gina's quotation of the article. So please re-direct your
> "horse puckies"!
Apologies if this was misattributed. I was mainly intending to comment on Markman's column that you referenced (@moneycentral.msn.com) as well as comments on the Nanovation pages.
[As a more general comment, I'll simply state that interpreting letters in hypertext as non-hypertext is probably an error prone process so until we all have a common set of tools that interprets both or we have a better interface in general, mistakes may get made. So I guess both care and tolerance are suggested.]
> The issue of digial gridlock is interesting from the point of view of an
> investor looking for opportunities.
True, and in that respect, high frequency (GHz) radio for the last mile seems quite interesting.
> I agree with your two points about, if we're talking about *net* gridlock
> (which is probably what the above quote means), but that's not the
> *only* kind of gridlock. I have not personally examined what Nanovation
> is up to, so I don't know whether or not it's really addressing gridlock
> internal to computer systems.
Nanovation is directly trying to address telecommunications bottlenecks and perhaps to a lesser degree computer bottlenecks. I get bent when news (based on corporate press) excessively promote a solution to a problem which (to me) seems nonexistent [i.e. net gridlock].
> Yes, there are such areas of gridlock or bottlenecks. (Which I'm glad
> to hear that Intel is boosting bus rates and not just microprocessors
> while I put off buying a new computer...) >
Putting off buying a new computer seems to make a huge amount of sense now-a-days. However, no matter how much they boost bus speeds, the processor clock speeds will be boosted faster. The Memory(bus)-to-Processor speed gap *IS* the fundamental problem in computer architectures at this point and bus speed increases or cache size increases are only stop-gap measures before the fundamental architectural problem needs to be addressed. You *either* must go to optical buses (less likely) or shift the archtecture to processor-in-memory (probably more likely).
> Right. That's one reason I own a little piece of JDS Uniphase -- one of the
> best perfomers in my portolio (or in anyone's portfolio) this year. ETEK
> and other companies are also (more narrowly) working on boosting optical
I'm unfamiliar with the company(s), but if that is their strategy, it seems very sound to me.