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.