Thin-client web framework delivers interactive maps and charts

An open-platform framework improves Internet web mapping by combining the richness of traditional desktop applications with the reach of browser-based systems.

 

Thin-client web interfaces are rapidly becoming the de facto standard for accessing information. Unlike traditional client-server, browser-based web applications that provide an intermittent user interface, thin-client web applications have the added benefit of a direct-manipulation user interface that is typical of desktop applications. These interfaces are greatly preferred by users,1 but were not possible on the web because of latency.2 Technologies that combine the functionality of a desktop application with the wide reach of a browser are called Web 2.0 (see Figure 1). In Web 2.0, JavaScript code in the browser caches and displays information that has been asynchronously downloaded from the server using extended markup language (XML). (The programming paradigm for this Asynchronous JavaScript XML is called AJAX.3) Additional JavaScript code handles panning, zooming, scaling, and data validation. Web 2.0 allows users to work with few delays and provides an interface that can be directly manipulated. In our research, we addressed some of the remaining limitations of this approach by creating a thin-client mapping framework.

 

Web 2.0 eliminates the intermittent nature of traditional web applications, provides a direct manipulation user interface, and maintains the wide reach of a web browser.

 

Dynamic web sites such as Google Gmail, Flikr, and Orkut are based on AJAX. Recently, several new web applications, such as Google Maps, Microsoft’s Virtual Earth, and Google Suggest, combine AJAX, dynamic hypertext markup language (DHTML), and vector graphics. In these applications, manipulations occur almost instantly without reloading pages. For example, in Microsoft’s Virtual Earth, the user clicks on the map and scrolls to the targeted location with the cursor. Google Suggest automatically attempts to complete a search query.

 

Google and Microsoft have added unique technical features to their products. In addition to map image tiles being cached and stored in servers, document object model (DOM) programming creates features that show locations, directions, hotels, and points of interest. Depending on the browser, vector markup language (VML) or scalable vector graphics (SVG) draws vector data, such as paths. Google offers an application program interface (API) so that third-party developers can add proprietary geospatial content.

 

Although Google’s and Microsoft’s web mapping technology is extremely clever, there are problems that must be overcome before it can be used to display arbitrary geospatial data. First, ?´per click?? commercial business models make deployments expensive. Second, proprietary map data from providers such as Navtaq and Tele Atlas make displaying arbitrary image data difficult. Although open source projects attempt to connect Google Maps to Internet web mapping servers (WMS), none has been particularly successful. Third, many applications need additional imagery that must be organized in server farms in the appropriate formats.Fourth, map tiles from different servers showing, for example, terrain, elevation, and weather cannot be overlayed.

 

To address these limitations, we created a Web 2.0 thin-client mapping framework called thinc GeoBoost (Figure 2) where browser-based geospatial applications run on top of open geospatial standards. This open-platform framework for Internet web mapping is unique because applications have the richness of traditional desktop applications and the reach of browser-based systems.

 

A thin-client web-mapping AJAX framework can provide interactive maps and charts.

 

Along with a thin-client JavaScript DHTML-AJAX web mapping client and pre-tiling/caching of WMS and web feature service (WFS) data, the framework supports Open Geospatial Consortium (OGC) standards such as geography markup language (GML), Geographically Encoded Objects for RSS feeds (GeoRSS), WMS, and WFS. Support for these standards enables systems to interoperate successfully.

 

The framework combines imagery from WMS into a single map display, supports merging image and feature data that is streamed to the client as GML, and integrates linked thin-client components such as time lines, graph displays, and simple charts. Customized applications are designed using an API, a programming model, and access to events.

 

Also, our work supports AJAX user interface frameworks (including those that will be introduced in Microsoft’s Atlas4 later this year), systems built using service-oriented architecture (SOA), and JavaServer Pages (JSP) and .NET tags. There is a wide diversity in geospatial platforms that will include both Java 2 Platform, Enterprise Edition (J2EE) and .NET systems. Custom tags provide an API to simplify programming in both environments. The next generation of web mapping frameworks will work across multiple platforms and software environments, including Microsoft’s IIS and the Apache Tomcat Web server.

 

Stephen Eick

SSS Research, Inc.
Naperville, Illinois

National Center for Data Mining,

University of Illinois at Chicago

Chicago, Illinois

 

Stephen Eick is the founder and chief scientist at SSS Research and deputy director for the National Center of Data Mining at the University of Illinois Chicago. Previously, he was a leading scientist at Bell Laboratories where he led the Visualization Research Group.

 

1. Ben Shneiderman,
Designing the User Interface, 3rd Edition,
Addison Wesley, 1998.

2. Stuart Card, Jock Mackinlay, Ben Shneiderman,
Readings in Information Visualization: Using Vision to Think,
Morgan Kaufmann, 1999.

3. L. D. Paulson, Building rich Web applications with Ajax,
IEEE Computer,
Vol: 38, no. 10, pp. 14-17, 2005.

4. http://www.atlas.asp.net/.
DOI: 10.1117/2.1200608.0332

 

(Credits SPIE)