In part I of this paper, a mathematical model for data structures has
been presented. Based on this model and on a corresponding formalized notion of
retrieval operation, a model for storage access systems is introduced, which is called a
Q-system. Equivalence, separation, and decomposition of Q-systems are studied,
resulting in a set of basic principles for transforming Q-systems. These can be used
for a systematical description of practical storage access systems for multi-attribute retrieval.