Programming in Prime

"whole number" does not exclude negative numbers and so you would have an infinite number of solution pairs.
Is this the intention or was Terry right when he assume that 0 would be acceptable but negative numbers are not?

In case that negative numbers are OK, too, you would have to provide a limit  - either a lower limit for a (runs from 2 down) or, as in the screenshot below, an upper limit for b (runs from 3 upwards):

Or are you looking for some kind of brute force attack?

Just a different take on it.  Solutions for a and b must lie on a line for any value of 't'.  So we can use y=m*x+b.  b>a only for t<2.5.