14, 10, and 7 Nanometers — Oh My!

I guess the cost is the controlling factor.

If Moore's Law was first derived from historical data and then given credence, then success (a return on investment) in producing smaller chips was achieved by doubling the density every two years. The results of which reinforced Moore's law. A sort of self-prophesying law.

This was relatively easy to do in the early days but as time goes by it is getting harder to stay on track with Moore because the science in the associated industries has not progressed at the same pace.

For instance, are the results from the work of the LHC needed to establish what could be achieved.

If so, this means investment in making 'chips' that much smaller and cheaper cannot be financed without income from industry that does not yet need smaller chips because they themselves are waiting for results on which to base their own investments.

As any engineer knows, extrapolating a (straight) line on a graph any more than a minor increment, from a mass of disperse earlier data, is very risky. Very often this sort of thing follows an exponential law where a minor deviations now could soon produce unpredictable results.

Chip makers need to look at what could have been achieved had Moore not made his law. Or what can be achieved by ignoring Moore.

Which of course they probably are - it does not need me to tell them.