According to ISO 19105 “Geographic information – Conformance and testing” standard, conformance assessment is based on abstract and executable test cases that are hierarchically organised into test modules and test suites. This standard specifies a conformance testing framework and criteria to be used when claiming conformance to the family of ISO geographic information standards.
In this paper we explain the conformance and testing challenges for Land Parcel Identification System (LPIS), which has to meet not only the technical, but also legal requirements of the Common Agricultural Policy (CAP). Recognising the benefits of formal models in terms of efficiency of design, enforcements of standards and consistency of implementation, we used Geography Markup language (GML) profile of Unified Modeling Language (UML) for developing a conformance testing framework both for implementation model and data testing.
The simultaneous development of the LPIS core design model with the testing model and the tracing relationships helped us to arrive to a compact, yet detailed model, where workflows and users’ interactions with the information system became unambiguous. This is an asset that enables the repeatability of test verdicts, transparency of procedures and thus, trustworthiness of testing. The model based approach is also efficient for deriving documentation targeted at different users, such as quality inspectors or IT developers. The UML model also allows exporting machine-readable Extensible Markup Language Schema Definition (XSD) application schemas that we used for developing a tool for schema matching between the System Under Test (SUT) and the reference schema, the LPIS core model.
With the model and standard driven approach we clarified how abstract and executable test should be applied in the LPIS testing framework and described the role of data quality measures as indicator-metrics for conformance.