Contrast the previous example, which used zero as the
boundary method, with this one, which uses wrap. Toggle the data
pane again; then select an output cell near the edge. This implementation has a quite different
behaviour: instead of some of the filter elements becoming irrelevant, we can see visually
that the inputImage behaves as though opposite sides were connected. These examples show
how just making the input-output relationships fine-grained rather than monolithic already reveals quite
a lot about what a computation actually does.