The purpose of this paper is to provide a construction to model
shared-variable systems using higher-dimensional automata which is
compositional in the sense that the parallel composition of completely
independent systems is modeled by the standard tensor product of HDAs and
nondeterministic choice is represented by the coproduct.