Unambiguous Object-Role Models

How and when ORM models can be unambiguous

Victor Morgante
9 min readApr 25, 2023

I have written quite a bit about the cases where sentences of Object-Role Modeling (ORM) may be ambiguous, and without having heard a word, I know it upsets some in the ORM community who hold onto a Platonic ideal of an isomorphism of ORM models to theorems of finite model theory and/or first-order logic (as per its predecessor NIAM).

My writing has been mostly of necessity, and I will explain.

To do the things I wanted to do with ORM, I required Object-Role Modeling to be the most ambiguous of conceptual modelling languages. This by necessity.

Those who know my work, know that this is what I wanted to do with the metamodel of ORM, and where all three of these languages are within the ORM metamodel as ORM diagrams:

Variable Interpretation of Theorems of ORM from within an ORM Metamodel. Image by author.

FactEngine’s Boston software exploits an inherent weakness in NIAM/ORM which comes about via ORM’s Fact Tables, and where differential interpretation of Facts (Sample Data) within Fact Tables within ORM can lead to a desired result of deliberate misinterpretation. That is, Boston deliberately looks for and finds an ambiguous interpretation of graphical ORM diagrams, that include Fact Tables, to achieve what Boston is most well known for…morphing between various conceptual modelling languages and storing more…