Representing Databases in Frames

Three methods for representing data in a relational storage system with an in-core frame-based system are experimented with and reported upon. Tradeoffs among these three representational methods are sizes of databases, times for loading data, and performance of queries. Essentially, these methods differ in ways of capturing relationships among frames. The three different ways of capturing such relationships are via links (pointers), symbolic names (keys), or both. Results of the experiments shed light on efficient interfacing of databases with frame-based systems.

Authors

• Ey-Chih Chow

Keywords

No keywords are indexed for this paper.