i think we should take mr. f's suggestion and run with it. use only algebraic methods, and certain "primitive" limits like sin(x)/x as x --> 0 (which can be proved geometrically).
yes, this may not be the original poster's intent, but are we mathematicians, or mice?