Software size is a fundamental software product measure. Many existing software size measures do not adequately address the underlying problem complexity of systems. We propose a vector measure that incorporates both functionality and problem complexity in a balanced and harmonious manner. Algebraic specification is used as a vehicle to derive foundation formulae for measuring software size in terms of its attributes and as a vector. Functionality is measured as the number of atomic units in the signature section of an algebraic specification. Problem complexity is measured as the number of atomic units in the semantic section of the specification. A vector representation, which has both magnitude and gradient, is used to demonstrate that the approach can be used to objectively compare and classify different classes of systems. Finally, we verify that our proposal satisfies the theoretical validation framework of Kitchenham, Pfleeger and Fenton.