What is Arom ?

The Arom project is an effort undertaken by the members of the Romans action at INRIA Rhône-Alpes, in collaboration with Pierre Mendès France University , Joseph Fourier University at Grenoble and the Laboratoire d'Informatique de Marseille. This project results from the synthesis of the work undertaken in the field of knowledge representation by the members of the former SHERPA project at Inria Rhône Alpes. The aim of this project is to provide the research and academic community working on object representations and/or wishing to perform experiments in the field of knowledge modeling with a powerful, extensible and portable tool.

Arom is a knowledge representation system, a generic tool for modeling and exploiting knowledge. It is designed as a modular system that can be used as a building block for building knowledge-based systems. Arom is based an original knowledge representation model based on classes and relations (Arom is the acronym of Associating Relations and Objects for Modeling).

A second originality of Arom is the presence of an algebraic modeling language. Such a language allows one to write equations, constraints and queries involving slots of both classes and associations. Declarative and expressive, this language makes it possible to implement mechanisms for exploiting knowledge such as constraint satisfaction and discrete-time simulation, thus opening knowledge representing systems to a broader class of applications.

The Arom project is far from being over but we think that the result of our work, even if it is not polished in its current state, may be of interest. For instance it has allowed us to build a Web knowledge server in electro-myography in the context of a European project, and a Web knowledge server for numerical solver choice in collaboration with the Dassault company. For this reason, we have decided to publish Arom, while it is still evolving.

Arom is written in Java 1.2. In the long run, Arom might become an open source project, so that other people can join us. However, this project is still too much in its infancy to be worth releasing the sources. for the time being, only the byte-compiled classes are published. The present version of Arom contains the following components:
In the short run, other functions will be added to Arom. In particular, a Java Application Programming Interface will make it possible to manipulate a knowledge base programmatically.