We introduce a framework that generalizes algebraic specifications by equipping algebras with descriptions of evaluation strategies. The resulting abstract mathematical description allows one to model the implementation of algebras on various platforms in a way that is independent of the function-oriented specifications. We study algebras with associated data dependencies. The latter provide separate means for modelling computational aspects apart from the functional specifications captured by an algebra. The formalization of evaluation strategies (1) introduces increased portability among different hardware platforms, and (2) allows a potential increase in execution efficiency, since a chosen evaluation strategy may be tailored to a particular platform. We present the development process where algebraic specifications are equipped with data dependencies, the latter are refined, and, finally, mapped to actual hardware architectures.