ABSTRACT
This chapter proposes an object-oriented multi-scale data model based upon the Digital Line Graph (DLG-E) data model developed at the US Geological Survey. It extends the DLG-E model to incorporate multiple representation schemes for a single object record within a single database, thus accomplishing a link from a single deep meaning to multiple surface meanings. In addition to supporting 'surface' generalization, the data model must support operations that precede map display, that is, generalization at the deep or modeling level. Deep or modeling operations will include formation of new data or data layers, for example spatial buffering, feature intersection, or may involve results of analysis and measurement, e.g. distance computation for tolerance value selection and modification. Ideally, a cartographic data model linking multiple representations for all scales at which generalization might occur should embed knowledge about how to effect variations in the cartographic (surface) meaning to preserve associations with geographic (deep) meaning.
