INSPIRE-compliant and easy-to-use GeoModel Editor
J. Schulze Althoff+, C. Giger*, Lorenz Hurni+
+Swiss Federal Institute of Technology, Zurich, Switzerland
*Giger GeoIT, Embrach, Switzerland
Conceptual Geospatial modelling plays a prominent role in the INSPIRE data specification, as well as in other implementations of spatial data infrastructures. The produced conceptual models enhance the communication within the domain as well as with technical experts and influence the interface specifications of spatial services. They are also used as basis for the development of end products like database schemas, format specifications or various forms of documentation.
There are various standards, languages and approaches on conceptual modelling and it is necessary to refine and restrict modelling languages like UML with according spatial constructs, stereotypes and rules of usage. Such standardisation and instructions, like contained in the INSPIRE Generic Conceptual Model (http://inspire.jrc.ec.europa.eu/documents/Data_Specifications/D2.5_v3.3_vs_3.2.pdf) support the creation of homogenous and similarly structured models. Compared to the possibilities that UML, ISO/TC-211 and the INSPIRE rules offer, the actual created INSPIRE data specifications are even more restrictive; no nesting of packages is used, the associations are very limited and only a few simple geometry types are used.
Figure 1: Metamodel of the used Conceptual Schema Language
To simplify modelling for experts, the idea of adopted and easy to use modelling editors for geospatial, conceptual models arises. For the working field of INSPIRE data specifications such model editors should offer UML Class modelling, but incorporate the rules of ISO/TC-211 and the INSPIRE Generic Conceptual Model as well as exposing only the data types of ISO-19103 Schema Language, ISO-19107 Spatial Schema, ISO-19108 Temporal Schema, etc. that are practically useful respectively used in the actual data specifications.
The presented modelling tool for INSPIRE data specification is an adaption of the HUMBOLDT Model Editor that was presented in the INSPIRE conference 2008. The HUMBOLDT model editor was developed within the HUMBOLDT project (http://esdi-humboldt.eu/home.html) on data harmonisation. It was built according to the project’s requirements and therefore did not exactly match the needs of INSPIRE data specification. The adapted model editor for the INSPIRE community shall overcome those problems of interoperability. The creation of conceptual models is simplified, structural consistency and a set of additional quality constraints is guaranteed. The created models can be stored locally in a specific data format with the graphical layout and arrangement, but also exported to be viewed and edited in the standard INSPIRE modelling environment.
Technically the adopted INSPIRE model editor is built on the open source frameworks of the Eclipse Modeling Project (http://www.eclipse.org/modeling/emf) and the Eclipse Graphical Modeling Project (http://www.eclipse.org/modeling/gmp). The core of the development contains a reflection of the requirements towards INSPIRE conceptual models in an abstract language syntax of a newly created specialised conceptual schema language (CSL), which preserves key concepts and the notation of UML.
To offer interoperability with other modelling environments, transformations from the internal CSL towards profiled UML must be possible. The chosen transformation approach is the Xtend model-to-model transformation language from the former OpenArchitectureWare Framework.
Figure 2: Screenshot of the prototype for an INSPIRE compliant Conceptual Model Editor
The presented software approach includes as much as possible prior knowledge of the intended content and structure for specific conceptual spatial models in the modelling software itself. By doing so, the domain experts are exposed to significantly smaller and more easy to use modelling software that also ensures basic model quality with structural and configurable validators. By using sophisticated language based transformation frameworks, the interoperability to existing models and approaches is guaranteed.