An evaluation framework for algebraic object-oriented query models

An evaluation framework consisting of five categories of criteria is developed for evaluating the relative merits of objects algebras, namely, object-orientedness, expressiveness, formalness, performance and database issues. Four recently proposed object algebras are evaluated against these criteria. It is shown that there exists no object algebra that satisfies all the criteria. It is argued that, since some of the criteria may not be compatible, a feasible object algebra has to make some tradeoffs to suit domain-specific needs. It is possible to identify a minimal subset of the criteria. The criterion that an object algebra should support encapsulation seems to be the most important. If an object algebra fails to support this criterion, its semantics is inconsistent with the concept of 'data abstraction', which makes a language object-oriented.<
>