Keywords
declarative modelling, ontologies, code generation, platform independent models
Start Date
1-7-2006 12:00 AM
Abstract
The need for integrating dynamic models with independently developed datasets and other models has long been recognized. Only recently, advances in modelling technologies, knowledge representation and protocols for remote communication of structured content have made this goal practical. The same advances make it possible to decouple the representation of a model from its executable implementation in ways that allow unprecedented levels of architecture independence and explicit, transportable model declaration. These developments are crucial to the creation of repositories of models where the models' lifetime is not tied to that of specific modelling paradigms, execution architectures, or storage technology. In this contribution, we describe a case study involving the declarative representation of model structure in a web-based knowledge base, its extraction through standard URLs as XML content, and the automated translation of the XML via dedicated components into source code models to be compiled for three different target architectures. We present the minimal declarative specification of the model interface and its key components, and discuss how it can be translated into executable form. We also describe the advantages of linking of each model component to explicit, external ontologies for matching model inputs and outputs to datasets and to other models. The advantages of the declarative specification are discussed in the light of ongoing, largescale projects in agricultural and biodiversity modelling.
Declarative modelling for architecture independence and data/model integration: a case study
The need for integrating dynamic models with independently developed datasets and other models has long been recognized. Only recently, advances in modelling technologies, knowledge representation and protocols for remote communication of structured content have made this goal practical. The same advances make it possible to decouple the representation of a model from its executable implementation in ways that allow unprecedented levels of architecture independence and explicit, transportable model declaration. These developments are crucial to the creation of repositories of models where the models' lifetime is not tied to that of specific modelling paradigms, execution architectures, or storage technology. In this contribution, we describe a case study involving the declarative representation of model structure in a web-based knowledge base, its extraction through standard URLs as XML content, and the automated translation of the XML via dedicated components into source code models to be compiled for three different target architectures. We present the minimal declarative specification of the model interface and its key components, and discuss how it can be translated into executable form. We also describe the advantages of linking of each model component to explicit, external ontologies for matching model inputs and outputs to datasets and to other models. The advantages of the declarative specification are discussed in the light of ongoing, largescale projects in agricultural and biodiversity modelling.
