Preface

Allgeographic information systems (GIS) are built using formal models thatdescribe how things are located in space. A formal model is an abstract andwell-defined system of concepts. It defines the vocabulary that we can use todescribe and reason about things. A geographic data model defines thevocabulary for describing and reasoning about the things that are located onthe earth. Geographic data models serve as the foundation on which allgeographic information systems are built.

We are all familiar with one model for geographic information—themap. A map is a scale model of reality that we build, using a set ofconventions and rules (for example, map projections, line symbols, text). Oncewe construct a map, we can use it to answer questions about the reality itrepresents. For example, how far is it from Los Angeles to San

Diego?Or, what cities lie along the Mississippi River? The map model also serves as atool for communicating facts about geography visually: Is the terrain rough?Which way is north? In fact, when we seea map, we often understand things that might not even occur to us as specificquestions.

Mapswork because we know the “rules” of conventional map reading: blue lines arerivers, North is toward the top of the page, and so on. In a similar way,geographic data models define their own set of concepts and relationships,which must be understood before you can expect to create or interpret your owndata model. These concepts relate to how you can represent geographicinformation in a computer system, rather than, as in the map example, on paper.

InModeling Our World, Michael Zeilerhas written an excellent primer for understanding the various models used torepresent geographic information in ArcInfo™ 8 software. He presents, usingstraightforward text and excellent illustrations, the concepts and vocabularyemployed in the design, implementation, and use of the ArcInfo 8 geographicdatabase. In addition to explaining the ArcInfo data model (objects, features,surfaces, networks, images, and so forth) in detail, Michael also provides goodinsight into how to use this framework to design useful information models thatfit your particular

needs.

Thisbook serves a variety of different purposes. For the geographer or scientist,it defines a conceptual context for representing geographic information.  For the GIS specialist, it serves as aguidebook in designing and using geographic databases. Finally, it introducesdatabase concepts to a geographic audience, and geographic concepts to the databasespecialist.

ArcInfo8 defines a unified framework for representing geographic information in adatabase. Several different generic data models are supported within thisframework:

•     cell-based orraster representation

•     object-based orfeature-based representation

•     network orgraph-element representation

•     finite-elementor TIN representation

Eachof these generic models has its own vocabulary used to define and reason aboutgeographic information. When we decide to represent roads, rivers, terrain, orany sort of phenomena in a GIS, we need to decide exactly how we defineinformation in terms of these generic models. As chapter 1 points out, thereare many ways that information can be modeled in a GIS. The representation youchoose for the data model will affect how you sample and measure geographicinformation, how you display it visually, and which relationships betweenelements can be represented, as well as query and analysis operations that canbe applied to the information.

Some have asserted that we should hiderepresentational models for geographic information (features, geometry,rasters, surfaces, and so on) from the users of geographic information systems.Somehow, these representational concepts are considered “implementationdetails.” In this view, a single real-world thing, such as the MississippiRiver, should be modeled as a single thing within the GIS. Perhaps, behind thescenes, the system could automatically use multiple representations for thesereal-world things. If you ask “What is upstream?” it could use a networkrepresentation of the river. If you ask “What is the surface area of thewater?” it could use a polygon feature representation. If you ask “What areadoes it drain?” it could use a surface or terrain representation, and so on. Whileit may be desirable to hide these concepts from some consumers of geographicinformation, I believe that a strong understanding of geographic data modelsand representations is crucial to the correct design and use of geographicinformation systems. Geographic data models act as the lens or filter throughwhich we perceive and interpret the infinite complexity of the real world. Itis only in the context of representations of the Mississippi River, forexample, that we can define specific properties, behavior, or even its identityas a “thing of interest.” Understanding geographic data model concepts iscentral to knowing how to define and collect geographic information. It is alsocrucial for correctly interpreting the results derived from the analysis ofgeographic information. This is similar to the role that statistics andsampling theory play in the natural sciences.

Forthe GIS specialist, this book serves as an introduction to a newobject-relational model for representing features, spatial relationshipsbetween features, and other thematic relationships. This new model issignificantly richer in its ability to represent features with associatedbehavior, relationships, and properties than the current coverage or shapefilemodel. If you are already familiar with coverages, shapefiles, and databasetables, the new model is a dramatic extension of concepts and capabilities withwhich you are already familiar. Our goal in building the new feature data modelhas been to move as much specialized application logic (for example,maintaining connectivity or relational integrity between objects) as possibleinto the scope of the data model itself. This allows more of the GISapplication to be defined using rules in the data model, rather than customapplication logic written for each application. For other aspects of the datamodel, which may already be familiar to the reader, the specific jargon andconcepts used in ArcInfo 8 (for topics like image data, as an example) areclearly introduced and defined.

Thisbook also connects the specialized world of geographic information systems andthe broader world of object-relational databases. ArcInfo now supports thedirect use of standard relational database technology as an integral part ofthe GIS. This introduces some new concepts to the GIS community. Topics such astransaction models for simultaneous editing of a shared, seamless database aredescribed in detail. For the GIS specialist, this provides a good introductionto standard database concepts. For the database specialist, this book serves asa good answer to the question “what is so special about spatial?”

Workingwith geographic information systems is fun for me because it serves tointegrate concepts and ideas from a variety of different disciplines— geometryand networks from applied mathematics, sampling and measurement theory fromremote sensing and physics, information modeling and multiuser database issuesfrom information technology. In working with GIS, we get to integrate all ofthis in a single, useful framework for building real systems. This bookpresents that synthesis, based on our work with ArcInfo 8. I hope you find thisbook useful and stimulating as a basis for your own work in geographicinformation systems.

ScottMorehouse

Directorof Software Development

EnvironmentalSystems Research Institute, Inc.

Redlands,California