Handling behavioral semantics in persistent meta-modeling systems

Abstract

The aim of our work is to integrate the benefits of classical MMS together with the advantages of DBMS. Our approach claims the extension of database perspective. Our idea is to define a multi-level DBMS to store metamodels, models and instances with an associated SQL-like declarative exploitation language which provides the capability to define, manipulate (query, update and delete) these different abstraction layers. The different model management operations (e.g., model transformation) should also be available in PMMSs using flexible procedural mechanisms such as external programs (e.g., Java, C++) or web services. Thus, our approach focuses on the capability to dynamically define operators that could be exploited by the PMMS exploitation language. In particular, such operators could be implemented with internal database mechanisms (e.g., triggers, stored procedures) or external programs stored outside the database (e.g., Java or C++ programs), or with web services. As a consequence of this extension, we will be able to perform advanced model management tasks such as model transformations in database, trigger web services from databases. Information exchange and data integration could be also supported in this context.

Extending Ontology-Based Databases with Behavioral Semantics

“…Our idea is that non-canonical concepts can be both computed by internal mechanisms of databases as well as external mechanisms such as reasoners, web services and remote programs. Our approach borrows concepts and technique from persistent meta-modeling systems (PMMS) [2]. It consists in extending PMMS to define OBDBs with the capability to introduce dynamically operators that can compute non-canonical concepts independently of any ontology-model.…”