|
A Route to Societal GIS ? - Geospatial Web Services
Societal Information Systems - The Promise of GIS Web Services Geography and GIS technology have long been recognised as an integrating sciences based on the ability to bring unrelated layers of information together through common, shared location and to analyse and visualise connections and relationships between them. However in the past GIS has been hampered as, the complex nature of spatial modelling has tended to be addressed in ways that alienate or separate the technology from the community at large. Unique data structures and computation complexity kept GIS a rather exclusive science, separated even from other information system technology. Data volumes and network bandwidth made transfer of GIS data and results hard. Such barriers have slowly been removed with the development of new storage techniques and increased conformity to general IT practices and standards. In the last two or three years increasing interest has been focused on the development of GIS Web services as a way of furthering this progress. Is Web services technology ready to meet the needs of Societal GIS? Web services model – robustness and flexible evolution A Web Service is simply a software component that can be accessed across the World Wide Web (WWW) for use in other applications. Web services are therefore another form of distributed computing architecture, of which there are plenty (Common Object Request Broker (COBRA), Distributed Component Object Model (DCOM), Electronic Data Interchange (EDI)). All these architectures have the same basic aim – to improve the flow of information across networks, to enable programs in one environment to communicate and share data and functionality with programs in another. Earlier approaches, though technically elegant solutions, suffered either because they were supported by a relatively small section of the IT industry and failed to garner widespread acceptance, or because they were complex and required a level of tight integration that was impractical for linking applications other than in large internal corporate Intranets. This limited their applicability to Societal Information systems. The Web Services architecture differs from these earlier approaches by adopting the ubiquitous World Wide Web (WWW) as the common network backbone. Web services are both discoverable across the Web and deliverable over the Web – key steps towards making them openly accessible and encouraging wide and universal participation that is demanded by Societal GIS. Anyone with a web connection (which now not only includes those with PCs, but increasingly those using web enabled wireless devices – PDAs, mobiles, etc.) can access and utilise Web services enabled applications. The ubiquity of the Web bridges technological differences and enabled Web services as a general approach to garner wide and sustained support. But the Web, at least initially, was never envisioned as a distributed computing environment across which applications can be joined and run – it was primarily designed as a client server environment in which known client software could establish dialogue with servers and retrieve and browse information from the server through fixed HTML Web pages. Web services work in a very different manner. Fundamental to Web services design are the concepts of Publish, Find and Bind that can be used to describe the basic relationship between service provider and client. A service provider can Publish a service that they wish to release in proscribed format to a Internet portal. Publishing the service identifies its existence to potentially entirely unknown and, at the time of publishing, undefined clients and provides details of how a client can communicate with the particular service.  Figure 2: Web Services relationship - Publish, Find, Bind Clients can search and Find published services in a similar way as popular Internet search engines are used to find Web pages. At present searches are more often than not undertaken manually, however the Web services architecture permits them to be undertaken automatically, so an application realising that it needs a particular function or dataset could search for and select remotely published services that meet its requirements. Published information for the selected service provides the client with all the necessary detail required to Bind to, or connect and use the service including where and how to invoke it and parameter definitions. The service may be information presented on a screen (as with traditional Web pages), raw data, a function, a complete application or any combination of these. This Web services model has significant consequences for Societal GIS. It provides the flexibility necessary both to enable services to be provided for users and tasks that are not known or clearly defined, and, to permit evolutionary development of applications to meet changing circumstances. Loosely bound and often highly componentized, Web services greatly facilitate rapid development and deployment of technology. The architecture permits such components to be served independently – users that may be entirely unknown to the originator of a service can find it independently and snap services together to develop solutions that meet their own needs (also unknown to the service provider). The end result may in turn be served as an entire new Web service. In addition, this design model provides a degree of robustness necessary for Societal GIS. Components and data making up applications can be searched across the Internet – if one service fails, it can, at least in theory, be easily replaced with another. Web services build on the dispersed nature of the Web to provide robust backup and disaster facilities that would be punitively expense to establish with traditional technology. |