Flexible Querying of Imperfect Temporal Metadata in Spatial Data Infrastructures

Spatial Data Infrastructures (SDI) allow users connected to the Internet to share and access remote and distributed heterogeneous geodata that are managed by their providers at their own Web sites. In SDIs, available geodata can be found via standard discovery geo-services that makes available query facilities of a metadata catalog. By expressing precise selection conditions on the values of the metadata collected in the catalog, the user can discover interesting and relevant geodata and then access them by means of the services of the SDI. An important dimension of geodata that often concerns such users' requests is the temporal information that can have multiple semantics. Current practice to perform geodata discovery in SDIs is inadequate for several reasons. First of all, with respect to the temporal characterization, available recommendations for metadata specification, for example, the INSPIRE Directive of the European community do not consider the multiple semantics of the temporal metadata. To this aim, this chapter proposes to enrich the current temporal metadata with the possibility to indicate temporal metadata related to both the observations, i.e., the geodata, the observed event, i.e., the objects in the geodata, and the temporal resolution of observations, i.e., their timestamps. The chapter introduces also a proposal to manage temporal series of geodata observed at different dates. Moreover, in order to represent the uncertain and incomplete knowledge of the time information on the available geodata, the chapter proposes a representation for imperfect temporal metadata within the fuzzy set framework. Another issue that is faced in this chapter is the inadequacy of current discovery service query facilities: in order to obtain a list of geodata results, corresponding values of metadata must exactly match the query conditions. To allow more flexibility, the chapter proposes to adopt the framework of fuzzy databases to allow expressing soft selection conditions, i.e., tolerant to under-satisfaction, so as to retrieve geodata in decreasing order of relevance to the user needs. The chapter illustrates this proposal by an example.