John - Re: " where Intel and the marketplace are headed."
Intel will head in whatever direction the market allows it to mainitain and expand their existing business. Clearly, the "simplest" approach is to supply the market with faster and faster x86 compatible chips that maintain 100% compatibility with all existing x86 based software. This multi-billion dollar software franchise is theirs (shared with Microsoft) and should be protected at all costs.
One possible divergence may be the much talked about P7 chip that will supposedly maintain x86 compatibility along with instruction set compatibility with the PA-RISC architecture of HP. What advantage the HP link provides seems pretty small (meager?) today, since it seems to be entireley supported by HP - operating system (HPUX) and applications.
However, simply providing faster and more efficient x86 processors by multiple pipelines (superscalar), advanced process scaling (0.25 micron/0.18 micron feature size) should be sufficient to keep the existing ball rolling.
Intel seems to be focusing well on the systems side with AGP port developments for faster graphics, their link up with Lockheed on very fast 3D graphics boards, Serial Bus interfaces for peripherals, advanced PCI buses, (hopefully) faster external bus speeds for motherboards (introduced by Cyrix - 75 MHz), support of faster bit-rate technologies for internet access, MMX instructions, etc.
These efforts, when realized, will improve PCs in all aspects, providing increased value to all users.
Computers, I believe, are computers. They are not toasters or dog-feeders. Improving the existing x86 machines has brought Intel this far, surpassing nearly all RISC architectures, and everything else that has crawled out of the technological woodwork.
When Intel has tried to deviate from x86 machines, they have traveled a rocky road. Their iAPX432 machine was technically brilliant, but a market flop. Their i960 is a great machine but targeted by necessity to low cost embedded applications. Their 80860 was very fast at the time and incorporated some brilliant features in the graphics and FPU areas. When was the last time you heard of this chip?
The real R & D required is that which provides new uses for PCs, not new kinds of PCs themselves. Typically, this falls in the software arena, surrounded by support hardware. Networking is one big example, and Novell and the internet are the most obvious instances of these. In turn, routers, switches, hubs, etc have been developed to improve the software aspects of networking.
I'd say Intel is on the safe and correct track - improving the breed of x86 machines and systems. New applications will follow, which will lead to more PC sales.
Paul |
