In computer science, one way of providing a semantics for programs is by using infinite trees to model do-loops in finite flow-graphs. Dana Scott and Marshall Hall are the earliest I recall. Arbib and Manes later on.
A lot of knot theory can be tackled by a similar tack.