Yesterday I came across two articles that misstated the famous Moore's Law. While not science, it has had a profound effect on computing and it is sad that many people don't seem to understand its underpinnings - even at a rough level.
A decade ago, when I was in the research division of a then large telecom, I was part of a group that was charged with a mid and long range projection of the impact of technology on telecom (I'm sure dozens of such exercises were in play largely to keep the research division feeling engaged with a business it was largely decoupled from).
It is striking to look at the projections for 2002 (we were asked to do 2, 5, 10 and 25 year projections - we only delivered 2, 5 and 10 noting that going beyond that was impossible). Many of the projections on bandwidth use and computing power were reasonably accurate, but the service projections were very incorrect.
There was one business projection - namely that the value of conventional switches in the network would diminish and it would be difficult for telecoms to sustain their profits and markets. (this was by no means an uncommon projection at the time). My director wanted that part removed from the report and it was "unduly and improperly excite people" ...
In any event the main prediction tool we used was Moore's Law. I remember writing a ten page explanation of it as an appendix for the paper. Recently Jon Stokes wrote a nice summary paper for Ars Technica. I recommend it to anyone who is unfamiliar with the underpinnings.
6:33:59 AM
|