P vs NP and a finite number of sybmols?

From: Casey Hawthorne (caseyh_at_istar.ca)
Date: 04/25/04 

Date: Sun, 25 Apr 2004 17:52:54 GMT

Is the essence of P vs NP that NP machines use an exponential number
of symbols (Turing machines - another one for each branch in the
computation) to see if there is one accepting state?

Therefore: there would not be a proof of P != NP using a finite number
of symbols?

Regards,
Casey