Matthew McGranaghan

Formalization of families of categorical coverages

Categorical coverages are popular for the presentation of thematic data in cartographic form and have gained widespread use. Routines for their preparation are included in most GIS software. Categorical coverages are simple to implement and easy to understand. The popularity of the categorical coverage is closely related to the intuitive rules that determine its behaviour. This paper presents a formalization of these rules based on the concepts of a partition of space and the refinement of such partitions of space. A family of categorical coverages is defined as all categorical coverages that can be produced from a data set through aggregations of categories. It is shown that categorical coverages are partially ordered by refinement of partitions. This ordering is preserved in the mapping from the ordering of the partition of the attribute domain to the partition of space. This formalization is not only useful for implementors of GIS, but the list of possible operations and their properties is the base fo...

Automated Provision of Navigation Assistance to Drivers

Advancing computer technology has made possible automated navigation aids for drivers. These aids are supported by sufficient positioning and data storage technology. The most crucial component of these systems, the user interface, is also the least understood. Few cartographers have anticipated and concerned themselves with the unique design demands of information display in this environment. Several questions about the suitability of current user interfaces are raised and future research needs identified.

Matching Representations of Geographic Locations

Many references to geographic locations are made linguistically, using words to describe the location in terms of objects in the landscape and the relations among them. In order to attach coordinates to these locations, one must be able to identify them in a database which contains representations of the objects and relations as well as coordinates for these features. This article assesses the correspondence between the features used in describing locations in an herbarium collection and those represented in standard cartographic databases. An alternative representation for geographic data is suggested.

Ordering Choropleth Map Symbols: The Effect of Background

Guidelines for choropleth map design have concentrated on printed maps. This paper examines whether the conventional use of darker symbols to mean “more” is applicable on maps presented in other media. Forty-seven subjects performed map reading tasks using slides photographed from a CRT screen. Results indicate that map readers take darker symbols for “more” in these displays. Further, the tendency for map readers to take darker symbols to represent greater magnitudes is modified by the value of a maps background. Approximately one-fifth of the experimental subjects reversed symbol order because of background value. Backgrounds of intermediate value impeded symbol ordering. This evidence suggests that for at least some subjects, a maps background anchors map symbols and influences their apparent order.

Spatial-linguistic reasoning in LEI

In LEI, the Locality and Elevation Interpreter, lingusitic and geometric reasoning are joined to resolve natural language representations of locations as geodetic coordinates. The system uses the PAU (Parser and Understander) system to translate locality descriptions into semantic relations. The relations are then disambiguated by instantiation from USGS data bases. The projects goal is to replace operator intervention in locality interpretation.