Mark - Re: " It will be embedded --
not an Intel strength --..."
As a point of fact, Intel is extremely active and successful in the embedded processor market.
Their i960 risc family seems to be the processor of choice for use in the data communications industry, printer controller industry, and other related industries. The great majority of HP laser printers manufactured over the past 4 years contain i960 processors.
Intel's 8051/2/3, 151, 251 family of controllers has ammassed over 100,000,000 unit shipments over the past 15 years, and is widely second sourced by Atmel, Philips/Signetics, Dallas Semi, Matra Harris, etc.
Intel's 80196 family was adopted many years ago by Ford and has been included in the engine controls of all Fords for about the past 10 or 12 years. (I think Ford has gone back to Moto for their next generation, however).
Of more interest, the i960 was chosen by Sun River Systems (they have a new name, something like Bubbling Technologies) as the basis of a family of three NC computers just announced this week.
In view of the NC market, Intel I think is showing great restraint in not chasing the latest bottle rocket fired off by someone who wants to replace the Wintel monopoly with his own monopoly (Larry Ellison/Oracle). They chased the PDA marketplace and found out how fleeting these new-fangled contraptions can really be.
There are now MANY different kinds of NC computers announced, based on ARM processors, PowerPC processors, i960, x86, WebTV, etc. Someone has to write software for all these even if that software only has to display Internet/WWW scrren shots. This mutiplicity of platforms and architectures has divided the market. A divided market is like the tower of babel, no one architecture will reach a critical mass. No critical mass results in few software developers supporting any one platform.
And it is SOFTWARE that sells hardware.
My guess is that time will dull the enthusiasm of NC afficionados. Too little horsepower, too few offerings, etc., too limited capability.
Check out IBM's NC announcement of yesterday. A PowerPC based NC for $700 - no display and no disk storage. I won't delve into the extra costs required to make this useful. You can, for your own exercise, find out what a COMPLETE FULL BLOWN x86 PC you can get for $1000. ($700 plus a display cost plus disk drive & modem). This will get a REAL computer (486 or low end Pentium) with real software that you can buy and run today and tomorrow.
Regarding 30% vs 90% market penetration for Intel. Over the past 14 years, the price of a PC has dropped in absolute dollars, the performance has skyrocketed, while wages of the basic family have shown perhaps 40 - 50% absolute increases. If Intel/PC markets continue their price decline, performance increases while family incomes continue to rise, the PC (or "connected PC") will grow and penetrate in market accceptance.
But keep in mind - people need a reason to buy a computer. My guess is that 70% of households don't want, need, understand or care about computers. Some are barely surviving trying to make the rent, some are content to watch OPRAH on TV, and some are satisfied with reading good books (well, a few, maybe). Many have no college education.
How mnay of your friends have difficulty setting the time display on their VCRs? And you expect these people to saddle up to an NC and log onto the net? Oh yeah!
A good relative question is this. In 1920, about 20 years after the automobile was first produced in the US, did 30% of all families have automobiles? In other words, is the PC penetration into homes any different than automobile penetration in its ramp up cycle?
Paul |
