If there is enough adoption 21 million coins is not enough to go around. ... If 21 million people were interested in bitcoin the price would easily head north of 4 figures.
I strongly disagree, for this simple fact: one satoshi carries exactly the same transactional capability and benefit to consumers/merchants as does one full Bitcoin.
People need to get over the mindset of $1/
BTC1, i.e., value being assigned in integer increments.
The current Bitcoin protocol provides for a maximum of 21M BTC x 100M sat = 2.1 QUADRILLION units of fungible supply.
BenThe argument Transidium puts forth has been made before but it does not hold up to scrutiny [1]. One could attempt the same argument with gold that he is making with bitcoin, and the argument falls apart in both instances, and for the same reasons. The (invalid) argument goes like this: the smallest unit of gold is one gold atom; the number of gold atoms that have been mined is ONE CADJILLABILLION (actually I don't know, but it's more than the number of Satoshis that will ever exist); therefore, the supply of gold is effectively limitless (in terms of number of atoms); and this places some sort of ceiling on the value of gold.
One needs to realize that there is a difference between these two things:
1) the total supply of something
and
2) the number of indivisible units of that something.
The error in Transidium's logic is that he does not appreciate the difference between 1) and 2)
Thought experiment: suppose the US Mint decided tomorrow to start minting coins worth one-tenth of a penny. Will this cause the purchasing power of the dollar to fall, because there are more smallest-units in the world? (Answer: no.) Likewise: if the core developers and the community at large decided tomorrow that one bitcoin could be divided into a few extra decimal places, would that cause the value of one bitcoin to fall? (Answer: no)
~ btcT
[1] unless he is making the argument in the context of some specific implementation of the concept of colored coins, where 1 satoshi could in principle be used to label something of an arbitrarily high value; but I don't think that is the argument he is putting forth.