Center for Applied Large-scale Computing (CALC)
Polytechnic University, Brooklyn, NY 11201
Phone: (718) 260 3305, Fax: (718) 260 3930
E-mail: amehta@quasar.poly.edu
Using object modeling techniques we have developed a unified metadata model and a search framework for metadata stored in various standards. Our approach has the following advantages: 1) Current metadata standards are modeled using the entity-relationship modeling technique. We have applied object modeling techniques to various metadata standards; 2) The unified metadata model is a repository for all the geo-spatial data elements. Once metadata are stored using the unified metadata model, it will be complete and will provide universal metadata access. 3) Due to inheritance properties of the object model additional standards can be easily added to the unified metadata model. It is a very difficult task using the current metadata standards.
First of all we have combined all the existing data elements and concepts into a unified data model. This unified metadata model has access to a search framework which supports metadata conversion from one standard to another. The Search Framework has two components: Multimodel Schema, and Conversion Framework. The Multimodel Schema contains searchable fields (data elements/attributes) and equivalent fields which require conversion. These attributes are arranged in a tree hierarchy. Classes (of the Unified Metadata Model) which contain search attributes and equivalent attributes are added into the Conversion Framework. Each class inherits conversion properties from the built-in conversion classes.