I wrote:
> The actual effective limit on NUMERIC is presently 10^128K, which is
> probably enough to count the femtoseconds remaining until the heat death
> of the universe, and then multiply that by the number of elementary
> particles therein ;-).
Should have done my research first. A little googling says that
* The total number of particles in the universe has been
variously estimated at numbers from 10^72 up to 10^87.
* The time to the heat death of the universe has been estimated at
10^200 years (and if there's one significant digit in that exponent
I'd be surprised...)
So the product I fancifully mentioned would weigh in somewhere around
10^300, and thus be *well* within the capability of even the proposed
restricted numeric format.
regards, tom lane