In this talk, I sum up some aspects of code design that has been floating
around since the early 1990s but have failed to materialize in the
CLASSIC/MAD9 debacle. In particular, I examine two things:
1) a polymorphic type for the ray leading to the computation of virtually
anything in perturbative theory.
2) the beam line viewed as a doubly-linked link list where each element is
actually the discretized position along the ring. Hanging off the
chain links are elements of the Euclidean groups as well as magnets or
pointers to magnets. Each magnet contains a single particle propagator (and
a reverse propagator).
I have applied these structures to recirculators and a pretzel type double
Talk Slides: PDF