Richard is executive director of the Center for Civic Networking in Washington, DC. Renée is a PhD candidate in urban planning at Rutgers University and can be contacted at sieber@zodiac.rutgers.edu.
Civic networking is the use of telecommunications by the general public for community and economic development, nonprofit service delivery, and citizen participation in government. Computerized telecommunications such as e-mail and the Internet expand the possibilities for civic networking.
Rapid development of geographic information systems (GISs) is creating new ways for the general public to participate in local planning decisions affecting land and water use, roads, and development. This new level of maturity makes civic networking with GIS a possibility, characterized by hardware and software integration, object-oriented programming, internetworking of distributed spatial data sets, and federal efforts that require government agencies to deliver spatial data to the public at marginal cost. These factors will generate demand for new GIS applications that enhance community-based planning, and can empower small organizations and the general public in new ways.
As applications mature, GIS will be perceived as a front end to a user who, unconcerned with the underlying technology, can visually navigate through his or her community. Increasingly, users will be private citizens--not just city officials and staff members of planning agencies.
Public users will navigate their communities using PCs in libraries, schools, and even at home. Today, municipal GIS applications are typically based on a single database located on a central host computer. If networked, such applications are connected through Ethernet-based, local-area networks to terminals or workstations in the same building. In the near future, such applications will become increasingly distributed and will be networked through TCP/IP protocols that can exploit the Internet. Coaxial cable used for television delivery is simply the physical layer of what can be configured as a two-way broadband network. It is technically feasible to build internetworked GIS applications that operate through cable-television systems and thus become available anywhere in town.
This trend towards linking a common GIS front end to a back end of internetworked, physically distributed spatial-data sets parallels efforts of the Federal Geographic Data Committee (FGDC). With the clout of an Executive Order from the President, this committee, chaired by Secretary of the Interior Bruce Babbitt, is developing standards and procedures to make the government's large investment in spatial data widely available to the public over the Internet beginning in 1995.
A GIS is a computer system that assembles, stores, manipulates, and displays data identified according to geographic location. The total system also includes critical success factors such as trained personnel, support from key stakeholders, and adequate operating budgets.
A GIS is often built from a set of topological data layers describing land forms, infrastructure, and property lines. Such maps are often created by combining vector-based Census TIGER files and USGS digital line graph (DLG) files with more-detailed local data, collected with the aid of handheld transmitters linked to geo-positioning satellites. Vector data may then be integrated with Landsat raster images. The process of transforming maps conforming to the curvature of the earth to a two-dimensional plane is very difficult. Also, local topography must often be hand-drawn on digitizing tablets.
Once work is done for a region, however, it becomes possible to use geographic coordinates to link tabular data from external relational databases. This could include U.S. Census Bureau tract data, the National Wildlife Inventory, effluent-discharge permits from the National Pollutant Discharge Elimination System (NPDES), county soil maps from the Soil Conservation Service, zoning districts, aquifers, rail lines, sewage, roads and other infrastructure, archeological and historical sites, wildlife and heritage trusts, locations of threatened and endangered species, and so on.
Geographic information systems are widely used. Businesses want to find good locations for retail outlets. Watchdog groups want to know how public money is distributed into different neighborhoods in a city. Politicians want to know where their constituencies are. Public agencies use GIS to manage physical infrastructures such as roads. Oil companies use it to improve exploration. GIS has become a powerful tool in managing environmental emergencies, helping to calculate responses to oil spills and predict impacts of toxic-waste sites. Public health and safety officials use GIS to map the spread of a disease or shifting patterns of urban crime. Maintaining accurate, quickly accessible real-property inventories that include maps and photographs is important to both tax authorities and real-estate agencies. And, of course, GIS is increasingly recognized as a tool to provide urban and regional planners with rapid access to spatial data. The list is endless--but the general public is not often regarded as a class of user for whom to develop such applications.
Mapping software is rapidly following trends familiar from word processing, spreadsheets, desktop publishing, and database-development systems. These trends are hardware and software integration, improvements in application-development tools, networking capability and, in the case of GIS, new federal policies towards public access to spatial data.
In the past, GIS applications tended to be designed for a particular purpose--that is, they were likely to be built to serve the needs of one particular agency. The database was built, managed, and located at that agency as well. This is inefficient, considering the effort required to build the base layers. Now, GIS is beginning to provide integrated services across agencies. Ultimately, a GIS designed for a specific region, such as a county, should be equally useful to an emergency-response planner, a real-estate agent, or a neighborhood volunteer crime-watch group.
Because development of the base topological layers is so difficult and labor intensive, the same system should be accessible simultaneously to the county planner's office, the Red Cross, and the public library. The Red Cross and the county planner both may maintain tabular datasets that are of interest to each other. Both datasets must be visible to a user looking at a county base map on a computer terminal in another part of town. In this configuration, a general system that leverages resources across agencies and community organizations needs to integrate physically distributed datasets across a network infrastructure.
Powerful GIS packages such as ArcInfo from Environmental Systems Research Institute (Redlands, CA) and MapInfo (Troy, NY) are available on both Windows and Macintosh platforms. Both Microsoft and Lotus are developing mapping engines that will integrate spatial datasets developed on proprietary systems with their respective Excel and 1-2-3 spreadsheets.
In May, 1994, the Department of Housing and Urban Development (HUD) began to distribute a mapping application to all of its field offices to assist communities in planning proposals for the agency. A suite of Windows and Macintosh applications built with FoxPro and MapInfo greatly automates a complicated proposal-application process that requires the generation of detailed demographic maps using Census data.
This toolkit is made available with the intent to provide citizens with a copy of the planning proposal that can be read on a standard PC, for public review. While these applications are not designed to be networked with this initial release from HUD, the Foxbase database-development system and the MapInfo GIS development system are both designed to work in networked environments. In the future, such applications may become increasingly networked, becoming available to the general public at libraries, schools, and even at home--via broadband Internet services through cable-television systems.
Object-oriented programming is making its way into GIS applications. In a spatial dataset, an object can be defined as a map feature--a bridge, for example--along with nongeographic attributes such as load stress or road traffic collected from on-site sensors. Data and instructions associated with map features have "intelligence" and can act with knowledge under certain conditions without the programmer writing additional code for each case. The object can also understand that it needs to link to a remote dataset to access data under certain conditions, an important requirement in a networked application. For example, the bridge object may understand that it needs to update traffic data periodically from a remote sensor embedded in the road.
Objects can be arranged hierarchically into classes, where subclasses inherit attributes associated with their parent class. Software for GIS that takes advantage of this important characteristic could further reduce application- and database-development time. The Argus mapping package from Munro Garrett International (Calgary, AB) is a Windows-based client/server package that employs an object-oriented approach which can link clients simultaneously to multiple databases. This approach reduces the need for a proprietary, separate spatial database--an ongoing concern by developers and public-access advocates for years.
A GIS sufficiently comprehensive for civic networking and policy analysis cannot simply portray the spatial dimension of printed reports. This is because public information exists in a variety of formats, from full-text documents, hand-drawn elevation maps, and 35-mm slides, to full-motion video and recorded sound. Additionally, related data exists in other applications such as statistics, computer-aided design (CAD), 3-D visualization, and modeling software. Increasingly GISs are being designed to handle nonconventional data. Nodes on data layers once referred only to tabular data; they can now be dynamically linked to bitmapped images of historic homes or analog oral histories of neighborhoods. Moreover, as software converges many applications that are geographically referenced will be available in a single front-end GUI of multiple representations. Also as data-compression techniques for these large datasets improve, multimedia becomes increasingly effective for LANs.
As an example, various information about proposed waterfront developments can be displayed for citizen comment. Simultaneously, multiple windows can show an aerial photograph of the site, an architectural rendering of the buildings, a 3-D simulation of the buildings in context, the numerical results of a model of generated traffic patterns, the sound generated by the traffic, textual histories of the waterfront, and names and addresses of contact people. (See "Planning with Hypermedia," by Lynn L. Wiggins and Michael J. Shiffer in Journal of the American Planning Association, Spring 1990.)
Hypermedia would also extend associative data structures to GIS. It would allow one to relate not only concepts, images and digital video, but to relate them geographically. Pointers could be to filenames or latitude/longitude coordinates, for example. Authoring tools, which long have been graphically based, are ideally suited to script this integration.
Recent experiments using the World Wide Web and Mosaic suggest that the Internet and a growing base of "resource-discovery" products may be able to work together as a suite of tools developers can use to build front ends into an internetworked GIS application. Indeed, a "map server" developed at Xerox PARC (Palo Alto Research Center) displays a map of the world to a Mosaic user, who can then select regions with a mouse, successively calling up more-detailed maps.
The Wide Area Information Server (WAIS) is a query tool that works in concert with an emerging international data-retrieval protocol known as Z39.50. Both are in the public domain. Z39.50 has been upgraded to permit queries using geographic coordinates, and enables transactional processing for payment, if needed. The World Wide Web and client-based browsing tools such as Mosaic provide gateways to WAIS servers and allow graphic- or forms-based queries for spatial data.
Structured query language (SQL) is increasingly being used to link client GIS applications to remote databases in real-time across networks. Previous versions of SQL were limited in their ability to work with spatial data; however, recent implementations have resolved some of the earlier technical drawbacks. Similar to the development of macro languages for word processing and spreadsheet software, high-level scripting languages are becoming available to facilitate the coding of such links across networks to distributed datasets. One example is Atlas GIS for Windows from Strategic Mapping (Santa Clara, CA), which employs Visual Basic as a scripting language. MapInfo also supports SQL and includes a version of Basic to develop vertical applications as well.
Eliminating a requirement for centrality through networking reduces risk of becoming locked into proprietary standards. At the same time, improving data integrity using metadata standards increases the confidence level for using data that is not necessarily under centralized control.
Database development has long been the most arduous component of a robust GIS application, and not a few developers have concerns over networked applications where data integrity is an issue. Here, the Federal government may be providing some solutions in a policy calling for a metadata standard for spatial datasets. The metadata standard would require "data about the data" to be included in a dataset, in a uniform manner. In this way, developers would have important details about the source and structure of remote data they might wish to link into. The metadata standard will also provide a coherent way for the government to mount huge spatial datasets on the Internet. These databases could be queried and downloaded for local applications, reducing development times.
The National Spatial Data Infrastructure is a federal initiative to bring together the technology, policies, standards, and human resources necessary to acquire, process, store, distribute, and improve utilization of spatial data. On April 11, 1994, President Clinton signed Executive Order 12906, "Coordinating Geographic Data Acquisition and Access: The National Spatial Data Infrastructure," which instructs Federal agencies to document spatial data beginning in 1995 and to provide this metadata to the public through an electronic clearinghouse within one year. The Internet is the most cost-efficient way to deliver such data to the public.
According to the FGDC,
The National Geospatial Data Clearinghouse is a distributed, electronically-connected network of geospatial data producers, managers, and users. The clearinghouse will allow its users to determine what geospatial data exist, find the data they need, evaluate the usefulness of the data for their applications, and obtain or order the data as economically as possible.
This Internet-based clearinghouse would require providers to describe data about their metadata. The provider may also provide access to the geospatial data; in the case of a federal agency, they may be required to do so. Metadata available through the clearinghouse will be distributed and physically maintained by providers all over the country.
Thus, the general public will be able both to locate government-collected spatial data over the Internet and to query information servers directly and acquire the data at marginal cost. This could dramatically reduce the entry costs to small entrepreneurs, community-based organizations, and private individuals.
In September, 1994 the FGDC announced awards for a new program called the National Spatial Data Infrastructure (NSDI) Competitive Cooperative Agreements Program, available to state and local governments, universities, and the private sector. A group of innovative projects awarded through the $250,000 fund will help develop the Clearinghouse itself, and the use of FGDC-endorsed standards in data collection, documentation, transfer, and search and query over the Internet. The FGDC plans to continue this cooperative program in fiscal year 1995.
Local-government planning agencies have begun development of internetworked, distributed GIS applications. For example, InfoWorld recently reported on a project of the Metropolitan Water District in Southern California. A distributed GIS front-end application linked to a central relational database helps planners monitor water-usage patterns among 16 million customers across six counties (see "Water District Fights Drought with Data Technology," by David Baum, InfoWorld, July 25, 1994). The system is expected to reliably predict water demand through the year 2000. Previously, large census files had to be processed manually to gain an understanding of where customers were located and their present and future water needs.
ArcInfo is used to process census tract data and provide demographic maps to users connected by Sun SPARCstations over an Ethernet network. However, ArcInfo is inefficient at maintaining the huge set of demographic records itself. Analysts using Sun workstations running a local GIS application link to a central database for subsets of needed census records, process this data and generate reports--usually maps--or pass results down the network to technicians using less powerful personal computers for spreadsheet analysis. The tabular census data, containing attributes such as housing and population, is maintained on a central Oracle database and can be linked, many-to-one, to the local GIS application. The Oracle database contains about 3 million records and resides on a SPARCstation 10.
Future plans for this system involve development of an object-oriented database system independent of any particular relational-database product. Using a metadata approach similar to that advocated by the federal government, the system will store the logical structure of many individual datasets without being tied to any particular physical database. This will make it possible to develop front-end GIS applications independent of the relational databases they may be linked to. In this way, such applications can link among any number of dissimilar relational databases containing tabular data without changing the application-program code.
Using an object-oriented design with global metadata structures, a local government will be able to develop a general-use, front-end GIS that can establish near-ad hoc linkages with many physically distributed relational databases. For example, place-based data maintained by a neighborhood crime-watch group could be linked into general GIS base maps maintained by a local government and made available through public libraries to citizens. The crime-watch database could be a Foxbase application, maintained on a 386, in a volunteer's home office.
How can a community-information infrastructure make a distributed GIS application useful to any neighborhood group or small organization in town? Glenview, a Chicago suburb of 35,000 was the first community in the United States--and perhaps the world--to establish broadband Internet connectivity to its civic institutions through a cable-television system. In Cambridge, Massachusetts, Continental Cablevision and Performance Systems, a commercial Internet provider, announced a joint venture in September of 1993, shortly after the system in Glenview, Illinois became fully operational. By March, 1994 commercial Internet service was introduced in Cambridge using many of the same components proven in Glenview. While Cambridge had the distinction of hosting the first commercial offering and receiving much press notice in the bargain, the two systems are nearly identical in technology and capability. Similar ventures are proposed or underway in a number of other cities around the country.
In Glenview, implementation of broadband Internet service over the cable system began in the late fall of 1992. By the summer of 1993 schools, libraries, and government agencies were connecting to a 4-megabit TCP/IP service to the Internet. This fall, according to John Mundt, Director of Administrative Computing for Glenview School #34, "3500 students will have broadband Internet access in class." Moreover, the school has installed a Cisco terminal server to enable SLIP/PPP connections for students at home.
Unlike the commercial service in Cambridge, the Glenview model connects to the Internet over an I-Net ("institutional network"). An I-Net is a common provision in many local cable franchise agreements that require noncommercial services in exchange for using the public right-of-way.
A cable-television plant is typically configured as a passive star. For example, a cable operations facility may receive satellite broadcast television signals at a head end, then retransmit them down a set of branches, where each branch serves a number of residential drops. Signal amplifiers are installed at fixed distances, but these amplifiers are generally passive and can only transmit in one direction.
However, in building an I-Net, a cable operator agrees to connect civic institutions for what amounts to close-circuit television as part of the public, education, and government (PEG) requirements often included in the franchise. Thus, the I-Net had already been installed as a bidirectional network in Glenview.
In Cambridge, broadband Internet service is priced at what the market will bear. In Glenview, access is jointly procured by civic institutions that share costs with a contracted Internet service provider. The joint procurement contract can be reviewed periodically, and a competitive bid process can leverage price. On the other hand, the commercial partnership in Cambridge may be in a better position to invest in upgrading the entire cable system to provide residential Internet access beyond what could be limited to a noncommercial I-Net in Glenview.
According to Glenview's Mundt, the equipment needed to build a community network over broadband cable is straightforward. "Each site has two pieces, a router and a broadband converter. In addition, a single repeater is required at the cable company head end."
The implementation project was undertaken by TCI, Zenith, Compatible Systems, and netIllinois. The Zenith ChannelMiser device bidirectionally converts Ethernet to broadband using two frequencies for send and receive. The ChannelMiser is capable of processing TCP/IP, Localtalk, or IPX packets. The Glenview system uses TCP/IP. Under certain situations, Localtalk packets are embedded within TCP/IP packets. A head-end repeater receives data on one channel and retransmits to another, with each channel separated by about 150 MHz. At several locations, a RISCRouter 3000E, manufactured by Compatible Systems (Boulder, CO) was installed. The router bridges between Ethernet-based TCP/IP networks and AppleTalk networks. Finally, an Internet connection was established by netILLINOIS, who provided a router and associated equipment.
The project began nearly ten years after the original I-Net had been installed. Much of the plant was old and untested, with many of the amplifiers inoperative. Neither TCI nor Zenith had outdoor broadband-equipment experience in a real-life cable-television environment. After some experience was gained, the most common failure point appeared to be the broadband amplifiers installed throughout the community. They are sensitive to temperature, and some have incurred damage from snow plows and automobile accidents during the winter. As the copper-based infrastructure gives way to fiber, reliability is expected to improve. Despite the technical flaws, the network is said to be operational 98 percent of the time.
TCI has indicated interest in further developing the system to provide Internet connectivity to the home through the cable plant. Zenith is developing the "Homework card," a lower-speed device card for PCs that is expected to provide throughput of about 200K. In Cambridge, Continental and PSI have offered a future residential service that will use either the Homework card or a similar device.
What critical success factors in Glenview made this community network possible? How can community groups in Cambridge take advantage of broadband Internet access through the cable system? What kind of public access GIS application could emerge in towns such as these?
In Glenview, a number of civic committees and organizations were aware of the I-Net capability. In exploring ways to utilize it, they decided that Internet connectivity was essential. Zenith, which manufactures bidirectional CATV amplifiers, is headquartered in Glenview. A new Internet service provider, netILLINOIS, was seeking public schools to connect to the Internet. These were all critical factors for success that came together, and the system has grown to the extent that over 3000 school children will have broadband Internet access in their classrooms in the 1994--95 school year.
Cambridge, with its universities and high-tech knowledge industries seems a perfect place to launch a flagship commercial Internet service. Yet, without community-based content, such a service may wind up being underutilized. Continental Cablevision provided a one-year grant of access to the Cambridge Public Library. In July, 1994, the library, in association with a number of community-based organizations, installed Macintosh public terminals linked to the Cambridge Civic Network, an Internet-based community network.
A community-planning initiative in Cambridge called the "Civic Forum" has held open public meetings for over a year addressing quality-of-life issues for a sustainable future. Recognizing that planning for a sustainable future is an information-intensive activity, organizers of the Civic Forum have begun to consider applications for the Cambridge Civic Network. Will a GIS be such an application? The factors for success are present. The city has plans for developing a GIS capability, and many businesses in the area have core capabilities. It is too early to tell what could emerge from a test-bed such as this, but as this article makes clear, GIS will soon no longer be an exclusive tool for planning agencies, direct-marketing firms, political consultants, or oil companies. It will be used by the general public to keep a better pulse of the lifeblood of their community.
What could a neighborhood crime-watch group do if a volunteer could connect a PC-based Foxbase application containing locally collected incident data, to a MapInfo-based community GIS application accessible over the cable system using a several-hundred-dollar Zenith Homework card? Will anyone change City Hall by sitting in the public library studying electronic maps to compare how the police were doing in different parts of town--on an hour-by-hour basis? In a few short years, we may know the answer.
Copyright © 1994, Dr. Dobb's Journal