Lessons Learned

A topology defines spatial integrity rules and GIS behavior for vector features. One example is that census blocks must be contained (nest) within and share the boundaries of the block group.

Although the results of our design projects are interesting (see "Status Report on ArcGIS Data Models"), some of the lessons are quite subtle. The major message is that there is nothing significantly different about a geodatabase design over former GIS data designs. Esri defined a geodatabase design process similar to traditional GIS design efforts. Users will find that new geodatabase designs will mirror the best parts of their existing coverage and shapefile designs.

Simplicity is key when it comes to building reusable, multipurpose data. Whenever practical, your design decisions should employ generic feature classes as opposed to custom features. Simple features are easy to prototype, support simple data migration, and adhere to open standards. Using simple features, you can implement important GIS rules and behavior using topologies and networks. Since simple feature classes are Open GIS Consortium (OGC) and International Organization for Standardization (ISO) compliant, your data sets can be easily shared.

The geodatabase designs we built using topology closely capture user notions for integrity rules and behavior. For example, "counties cannot overlap one another and must nest within states" can be expressed as rules in a topology. This is a clear advantage over coverage-based data models since users can simply define topology rules rather than build ARC Macro Language (AML) code to implement behavior and integrity requirements.

A significant lesson learned is the value of integrating feature classes that were previously managed independently. For example, we were able to integrate simple transportation layers for linear referencing, cartography, and road construction using topologies. Such integrated data management should lead to real savings by removing the need for redundant data collection efforts.

Perhaps the most important lesson we learned was how productive our design process became once we began to experiment with prototype designs and pilot implementations. Using personal geodatabases to prototype aspects of your geodatabase design, such as topology rules, is highly productive. This method enables you to quickly test and prototype any design idea to determine if it will meet your real world requirements.

After prototyping, piloting your database with the same technology planned for full implementation is critical to prove your design; if the intended use of your database is to support a multiuser editing environment, then pilot your database using ArcSDE with a significant amount of data.

Geodatabase Design Lessons

• Build on existing GIS conceptual designs.
• Generic GIS data types are open and multipurpose and result in practical, sophisticated models.
• Previously independent feature classes can be integrated resulting in more collaboration.
• Geodatabase design can be coupled with traditional DBMS design.
• Prototyping with personal geodatabases is easy, simple, and effective.
• Pilot your planned system architecture with representative data volumes.

Topology Plays a Key Design Role in GIS

Topology is implemented as a set of integrity rules that define the behavior of spatially related geographic features and feature classes. Topology rules enable GIS users to model such spatial relationships as connectivity (road lines connect at endpoints) and adjacency (no gaps or overlaps between parcel polygons). Topology is also used to manage the integrity of coincident geometry between different feature classes (e.g., coastlines and country boundaries are coincident).