Evolution of Environmental Information Models

Reusability of environmental data is essential for environmental research and control; standardized data models are being created by various organizations to facilitate this process. Due to the evolving nature of environmental science, these data models must be continuously extended for the support of new concepts, thus rapidly breaking the level of standardization achieved. The definition of reusable properties would allow for standardization of this extension process. In this paper, we first analyze the requirements to reusable properties, and explain the rational for the decision that reusable properties tightly bound to a URI would be the most apt solution; the following list of requirements was defined in order to compare the viability of the options proposed: URI Coupling, DataType Coupling, Semantics Coupling and Persistence. We then go on to explore possible avenues for implementation of reusable URI-Properties, whereby the following approaches where analysed for applicability: Data Types, Interfaces, MOF level adjustment of UML and a solution utilizing stereotypes for the definition and use of reusable URI-Properties. Of these approaches, all were deemed feasible except for the MOF level adjustment of UML; MOF level adjustment is not possible due to cardinality constraints within the MOF definition. Examples were created for the other 3 possibilities, including serialization options towards XML Schema. These examples were then compared with the requirements defined for URI-Properties; based on this analysis, the UML Stereotype based solution for the specification and use of reusable URI-Properties was deemed as most viable and is described in further detail.