"Every generation has its way." In fact, each paradigm change in information technology has also caused the emergence of new modeling methods. Relational technology had made Entity Relationship Diagrams popular, and object-oriented technology has lead to UML and ORM. Now, we are facing a new paradigm shift. The Enterprise Information Model gives way to an Open Network Information Model, and with XML we have now a powerful technology to describe complex information structures based on a document model. Within these new scenarios, existing classical modeling methods often expose problems: not expressive enough for the new data structures, not flexible enough for network structures, too slow for Internet speed.
AOM was born out of these requirements. At the moment, AOM is basically a pencil-and-paper method, but tools are beginning too emerge.
By representing relationships in their reified form as assets, AOM allows to define, describe and visualize higher order relationship, i.e. relationships between relationships. Higher order relationships are an important concept in modern information systems but cannot be appropriately describe in classical Entity Relationship Diagrams and UML. AOM adopts higher order relationships from Bernhard Thalheim's HERM (Higher Order Entity Relationship Model). However, it does not denote relationships as such, but reifies them as assets, a concept borrowed from RDF. This solves the problem that in many cases it is unclear to the designer if an information item is to be modeled as an entity or as a relationship. In AOM we are only dealing with assets.
AOM Level 2 introduces the concept of business objects which are large scale objects aggregating several lower level assets. These business objects are closely related to objects of the "real-world" and allow an intuitive and natural model of the business domain. AOM Level 2 business objects are also an important guideline when the information structures are implemented in XML. In general, every business object corresponds to an XML document type.
Due to the fact that AOM allows the definition of properties with complex structures, AOM models are usually much smaller than equivalent UML or ERD models, and of course much smaller than equivalent RDF graphs. The regular expression syntax allows the definition of structures of arbitrary complexity. The possibility to define complex types (which are assets, too) and to use these type definitions within property definitions adds to the the compactness of models and allows to reuse definitions.
By attaching globally unique namespace identifiers to models and/or assets, AOM is able to merge models both horizontally and vertically. This allows the distributed development of models and also allows to develop within a particular view of a model, similar to subject-oriented developing in OO.
AOM was developed with simplicity as one of the main design objectives. We invested more work in dropping features than in introducing new ones. The result is a metamodel of extreme simplicity, which makes AOM easy to learn and to use.
Contact: support 'at' aoModeling.org