Building place reference systems that can be used as annotation standard for integrating web resources and geographic information. In particular, distinguish a variety of relevant types of place by reconstructing them from affordance-perception in the environment.
Final Report Abstract
Reference to places is a central but largely underexposed problem of information science. Place has been a major object of research in Geography and GI Science, however, geographic information systems (GIS) and gazetteers have been built on space reference systems. While reference to space only partially captures reference to place, most existing definitions of place either reduce the latter to the former or lack a formal characterization. In a spatial reference system, locations are referenced by their relative distances to other referents. In this project, we developed the idea that place reference systems can be built based on localizing things (locatums) involved in activity simulations relative to other involved referents (locators). In this way, place reference systems serve to automatically identify, classify, and localize places independently from spatial reference systems, based on potential activities that can be performed in them. In the course of this project, we published a formal theory in two variants of logical expressivity (HOL/OWL 2), which can be reused to reason on places on the Semantic Web and in HOL theorem provers, respectively. The OWL variant of the theory (http://www.geographicknowledge.de/vocab/PlaceReferenceTheory) allows automated classification of places based on activities. The theory is based on cognitive capabilities to simulate actions and involvements of referents in the perceived environment, which allow to localize places relative to involved referents. However, this principle can be extended by technical action simulations, which allows to refer to places also in the absence of human observers. We have suggested ways to implement place reference systems in a GIS, based on a number of concrete scenarios of place reference, such as market places and places of sight. The project has implications not only for place-related GI technology, but also for GIS technology in general. GI research is confronted with a paradigm shift from coordinate centered to place centered geo-reference. Our theory shows a way how the technological development of a “placial” GIS may be approached under the new paradigm. This may include, e.g., place GIS operators which automatically identify, classify, and localize places from activity simulations or observations. Furthermore, the work contributes to the problem of spatial reference of information sources in the Semantic Web.
