i don`t think he is - he`s just saying that if you were given the question: examine whether the function (x^2-9)/(x-3) is continuous - you would firstly examine it AS IS, then what you normally do in maths is see if the function can be simplified, expressed in a different form so the true nature of it can be examined. in this case, because the simplification is continuous, but the denominator of the original form isn`t,at x=3,, there`s a discontinuity as x->3 (aremovable one because we can redefine the function and domain)
(same reasoning as tan x=sinx/cosx - cos x = 0 for +-multiples of 90 degrees+2nPi (n=0,1,2,....), so tan x - the simplified function - is undefined where cosx=0 - these points are not removable discontinuities, because the simplified function is NOT DEFINED FOR THEM - original problem is STILL DEFINED for x=3