www.yourcompany.com

The World of Data

Federated Databases & Semantic Integration

There are two approaches to semantic reconciliation among structured databases: loose coupling (rule-based exchange of data between independent databases) and tight coupling (mapping two or more databases so that they operate as if they were one database).  Of the two kinds of semantic discrepancy among structured databases, structural semantic discrepancy has been dealt with quite thoroughly in the ‘data transformation’ capabilities of data warehouses and middleware. ‘ Fundamental semantic discrepancy’ occurs where tight coupling cannot be achieved without changing at least one of the structured databases.  Techniques used to integrate multiple structured databases include semantic dictionaries (OBC Topics 1 and 2), concept hierarchies (OBC Topic 3), similarity-based criteria, concept closeness evaluation, semantic equivalence in ORM/NIAM, text database principles applied to structured databases, and semantic integration mapping languages like EXPRESS   Of particular note is an integrative concept, ‘situated integrity’ that builds on shared semantic understanding among the people involved in the situation that the data is about.  This is what OBC Topic 2 ‘concept’, ‘community’ and ‘context’ addresses.

Copyright © 1998-2023 Business Semantics Ltd