This annotated map of Fort Scott was developed by Map Bureau and the Institute for the Future, in support of a participant exercise performed at the fall 2004 IFTF technology horizons conference. The conference, with the theme New Geography: Strategies for Deep Place was held on October 6th and 7th at the Presidio of San Francisco. From the IFTF document describing the conference:
Imagine this: As you move about the physical world, you can access not only what you can see with a naked eye but also multiple layers of previously invisible information - annotations left by friends, colleagues, and complete strangers; data about those who live close by, including their demographic characteristics and political affiliations; crime statistics for the area; traffic accidents and extreme weather incidents that have occurred in the vicinity; information about businesses, their products, their reputations; and much, much more.
This concept is illustrated by the annotated map of the Fort Scott area of the Presidio. Map layers depict features of the area derived from GIS data provided by the Presidio, and the annotations (red dots) were produced by the staff of the Institute for the Future. The menu allows selection of the topic of annotations, and shows a clickable list of the titles of currenly selected annotations.
The map was built using a new generation of technology from Map Bureau, and represents the most complete example yet of the application of Semantic Web technology to cartography. As background, here is a quick summary of the ideas underlying the Semantic Web.
The Semantic Web project, advocated most prominently by Tim Berners-Lee, the original architect of the World Wide Web, seeks to apply the same kind of global unification to data as the original web has accomplished for documents. All useful data has semantics (that is, meaning), but traditional data representations, such as those underlying relational data bases, or data languages built directly in XML, lack uniformity in their encoding of meaning. The practical result of this fragmentation is that the current web of data is composed of a multitude of separate islands of information, one or more per application area, whose contents cannot be interrelated with any uniformity. This resembles the state of documents on the internet before the advent of the WWW
The technology needed to build the Semantic Web is here now, in the form of the RDF and OWL standards for representing data and types, and supporting infrastructure. The application of this technology will connect and unify the global realm of data, by allowing rich semantic linking among domains of information.
Since most things and events that concern humans have locations and often geographical extent on the earth, cartography connects directly to nearly every other variety of information. Consequently, cartography can benefit more than most enterprises from the integration that the Semantic Web brings. There is an opportunity to improve radically on the current state of affairs, where specialized tools and representations are the rule rather than the exception. Cartography, when carried out within the Semantic Web, weaves the geographic properties of things into a fabric of information based on a uniform scheme for representation, and lacking borders where disciplines meet; it provides tools for visualzing geographic aspects of things in this fully open context.
At Map Bureau, we have been working for several years towards realization of this idea of cartography. Despite the general and abstract nature of the goals, the results are easy to describe in concrete fashion.
(1) The data representation underlying the Semantic Web, RDF, is used throughout, with the sole exception of the raster imagery. In particular, the annotations (red dots), and their content, the layers (the streets, building polygons, sidewalks, etc), and their visual characteristics, the map projections (UTM10 in this case), and the geographic coverage of data sets - all are represented in RDF. Beyond this, the programs that perform the work of manipulating this data are represented in RDF as well, via use of the Fabl language. The result is that every single item is open to semantic annotation and semantic connectivity; concretely RDF properties of any kind can be asserted of every datum, at whatever granularity.
(2) Of course, only a small minority of geographic data is presently stored in RDF. Map Bureau has developed a converter to RDF from the most commonly used format for sharing geographic data, the ESRI Shapefile format. This converter makes a vast amount of cartographic data accessible. For example, the TIGER data set from the US Census, which includes the fundamental geographic information (streets, rivers, political boundaries) for the United State, is available at no cost in Shapefile format. Shapefiles representing the layers depicted on the Fort Scott map were provided by the Presidio of San Francisco.
(3) As in all of our work at Map Bureau, people view and interact with maps using a web browser (any modern web browser will do); the Macromedia Flash plugin provides client-side rendering and interaction with the map itself, while auxilliary content is rendered as HTML. The multi-resolution photo imagery in the Fort Scott map is managed by Zoomify. Map Bureau software running at the server translates RDF for the geographic map layers and annotations into the SWF file format utilized by Flash, and generates HTML for the auxilliary content.
Notes and references:
XML provides syntactic, but not semantic, integration. See this comparison between the RDF-based approach, and the dominant XML formalism for geography (GML).
Selected links related work
RDFIG Geo vocab workspace: http://www.w3.org/2003/01/geo/