Formalization of ORM Revisited

Abstract

Fact-oriented modeling approaches such as Object-Role Modeling (ORM) and Natural Language Information Analysis Method (NIAM) enable conceptual information models to be expressed using graphical diagrams that may be assigned formal semantics by mapping them onto sets of logical formulae. Various formalizations for such mappings exist. This paper extends such previous work by providing a new approach to formalizing second generation ORM (ORM 2). We show that the metalevel association between semantic value type and data type must be a mapping relationship rather than a subtyping relationship, and we axiomatize a special representation relationship to support this mapping at the instance level. Our new formalization includes coverage of preferred reference schemes and additional constraints introduced in ORM 2. Other issues examined briefly include the use of finite model theory, sorted logic, and practical choices for implementing certain kinds of logical formulae as constraints or derivation rules.