The expanding enterprise is to GIS what the expanding universe is to astronomy. Both have only recently (in their relative histories) been recognized as realities. The concepts for both have been around for some time, and the dialog at times has been impassioned, but the proof, well, that's recent history.

GIS has clearly transcended its roots, which became established in planning, mapping, and environmental organizations decades ago. The enterprise at that time was focused, typically, within a department. And the technology of minicomputers was ideally suited. Of course, the pioneers (or perhaps, missionaries) of that era saw the benefits more broadly. But in the harsh light of reality, most GIS deployments (e.g., the Maryland Department of State Planning, 1972, and the Minnesota State Planning Department, 1973) were focused on their own programmatic requirements. That's not to say that the products and benefits did not create value beyond the deploying agency, but the benefactor and beneficiaries were, most often, single agencies and sometimes only single departments within those larger agencies.

The technology of the time (circa 1975 to 1980) was well matched for this form of deployment. A central database, a few interactive display terminals, and some display screens met most common needs. Hard-copy plots were generated on a number of devices, and data and the results of various applications were typically circulated in analog form. The communities of developers and users were frequently the same individuals.

In the mid-1980s, the enterprise became a recognized entity. There was no standard definition, but it made for great marketing press. Systems were being designed as enterprise systems, and depending on where you stood in the organizational chain of command, radically different views of the enterprise coexisted, even within the same enterprise. This obviously was good for the technology vendors, both hardware and software. Microcomputers became more powerful, and workstation technology followed. Internal data communication networks began to carry the bandwidth that was useful for graphics. It was also good for GIS.

EMOLS pedestrian and vehicular access restrictions.

The GIS enterprise implementation of that era became, first, multidepartmental within the same organization (e.g., Pittsburgh, Pennsylvania; Los Angeles, California), followed by multiorganizational implementations (e.g., Louisville-Jefferson County [LOJIC], Kentucky; the Washington, D.C., Suburban Sanitary Commission; and the San Diego Data Processing Corp. in California). At this stage, the number of GIS end users grew significantly larger than the GIS administrative organizations that had grown even larger than the GIS developer community. In many instances, the beneficiaries of the results had little or no contact with GIS technology or databases. In other instances, one organization hosted the technology and databases, and the users of the technology were spread more broadly among, and perhaps as a legacy to, not only specific departments but spread into other separate organizations. In some instances, all of the organizations picked up a piece of the financial tab and continue to do so today (e.g., LOJIC and the Indianapolis, Indiana, IMAGIS Project).

Flash forward to the new millennium. The Internet is now a fixture providing access to many types of users and organizations. In such an environment, the definition of an enterprise changes dramatically.

Take, for instance, the enterprise of the New York City GIS Utility. The utility was designed in the mid '90s as a virtual data repository for spatial information sharing among 20 or more city agencies (the '90s enterprise). The lead agency was not an operational department (the traditional location for nonenterprise GIS) but was the New York City Department of Information Technology and Telecommunications (DoITT).

In supporting the Office of Emergency Management, DoITT and PlanGraphics implemented an ArcIMS application--the Emergency Online Locator System (EMOLS). This application provides information on shelters and emergency evacuation routes during storm and extreme heat events based on a user's street address. The application is available to all city agencies for responding to information requests and to the citizenry at large via the Internet. The enterprise now accommodates the servers, as well as those served--the city and its constituents--not indirectly but through a direct connection.

In an even broader definition of an enterprise, the Information Resource Management Division of the Oregon Department of Administrative Services (DAS) commissioned a Proof of Concept for Digitally Integrated Mining of Oregon Networked Data (DIMOND), a virtual data warehouse initiative.

DAS desired to develop a means of consolidating state data resources for general broad use across state and local government. Traditionally, this would involve a massive and complex data warehousing initiative. However, PlanGraphics and Esri proposed a more dynamic means to achieve data integration and availability across the extended enterprise including two state agencies and the state's GIS clearinghouse, Multnomah County, and the Tri-Metropolitan Transit Authority (Portland, Oregon).

The Proof of Concept utilized standard Java development approaches. Many components were prebuilt classes. Other portions were custom-coded to integrate various elements. The core of the Proof of Concept application is a custom-written Java applet that is designed to be downloaded via a Web page and executed on the client's desktop machine. This applet interacts with a set of specialized Java servlets, which reside on several server computers. Each servlet receives requests from the applet and returns the results of its actions such as a map showing county boundaries. Each of these interactions was coded by and managed using standard Java techniques.

The Proof of Concept was implemented in two parts. PlanGraphics developed a related pair of demonstration applications using Java clients interacting with middleware, and data translation servlets to remotely accessed data stored in Oracle, Oracle Spatial, coverage, shapefile, and other formats. The end user need only have a Java-enabled Web browser such as Netscape or Internet Explorer. Esri staff also built an application using ArcView 8.1 as a client, which interacted with data stored in the Oracle Spatial instance managed by the middleware.

The Proof of Concept demonstration was successful. Remote data connections were established to data providers, and the results of real-time queries were merged and presented to the user via Java-based client technology without regard to data source or underlying technology. Data sharing was performed with minimal impact on data provider systems and demonstrated adequate security through the use of built-in Java security features and a user login function based on Oracle database tables and JavaServer Pages.

The Internet has forced (and still facilitates) the redefinition of the enterprise once again to include multiple organizations within multiple jurisdictions and concomitantly their individual and overlapping constituencies.

For information regarding EMOLS, contact Al Leidner, assistant commissioner, Citywide Geographic Information Systems, New York City Department of Information Technology and Telecommunications (e-mail: aleidner@doitt.nyc.gov). For information regarding DIMOND, contact Cy Smith, Oregon Statewide GIS coordinator (e-mail: cy.smith@state.or.us). For more information about PlanGraphics, contact John C. Antenucci (e-mail: jantenucci@plangraphics.com).