Algebraists do it by symbolic manipulation.

Algebraists do it in a ring, in fields, in groups.

Analysts do it continuously and smoothly.

Applied mathematicians do it by computer simulation.

Banach spacers do it completely.

Bayesians do it with improper priors.

Catastrophe theorists do it falling off part of a sheet.

Combinatorists do it as many ways as they can.

Complex analysts do it between the sheets

Computer scientists do it depth-first.

Cosmologists do it in the first three minutes.

Decision theorists do it optimally.

Functional analysts do it with compact support.

Galois theorists do it in a field.

Game theorists do it by dominance or saddle points.

Geometers do it with involutions.

Geometers do it symmetrically.

Graph theorists do it in four colors.

Hilbert spacers do it orthogonally.

Large cardinals do it inaccessibly.

Linear programmers do it with nearest neighbors.

Logicians do it by choice, consistently and completely.

Logicians do it incompletely or inconsistently.

(Logicians do it) or [not (logicians do it)].

Number theorists do it perfectly and rationally.

Mathematical physicists understand the theory of how to do it, but have difficulty obtaining practical results.

Pure mathematicians do it rigorously.

Quantum physicists can either know how fast they do it, or where they do it, but not both.

Real analysts do it almost everywhere

Ring theorists do it non-commutatively.

Set theorists do it with cardinals.

Statisticians probably do it.

Topologists do it openly, in multiply connected domains

Variationists do it locally and globally.

Cantor did it diagonally.

Fermat tried to do it in the margin, but couldn't fit it in.

Galois did it the night before.

Mðbius always does it on the same side.

Markov does it in chains.

Newton did it standing on the shoulders of giants.

Turing did it but couldn't decide if he'd finished.