January 2002 | Google has indexed over 1.6 billion Web pages, and that number excludes "dark pages" inaccessible behind firewalls containing at least 10 times as much information. It also excludes information that is dynamically presented via databases. Search engines are getting increasingly sophisticated, but nobody has the patience to inspect each of thousands of results in one-dimensional lists of search results. As the amount of Web-based information grows exponentially, we need to develop new ways for presenting this information. Although the Web is inherently visual, its data presentation techniques today are woefully superficial: lists of results with rows of icons to indicate relevance or eye-candy animations.

A couple years ago, Furniture.com took steps to provide visual assistance for customers of its products with its Flash-based furniture layout capability. Visitors described the dimensions of a room and moved furniture stick figures around to see how the furniture might fit. The idea was good, but before placing an order, the least buyers wanted was to see fabric swatches on the furniture, look at pieces from side to side, and know if the options were in stock. Furniture's Flash-based facility was hopelessly detached from databases of furniture models, fabrics, and business inventory systems.

Perhaps learning from earlier mistakes, innovative solutions to the visual Web delivery of data-based information are emerging. I spoke recently with CEOs of two companies, Tim Ambrosino of Mariah Vision 3, based in Virginia Beach, and Tim Bray, of Antarcti.ca in Vancouver about two such solutions.

Mariah Vision's 3Scape product is a powerful, interactive authoring and delivery system enabling designers and developers to create engaging, informative Web scenes in a single content-rich experience. The development tool is similar to Flash with motion-paths and object events defined across a time-based scene. 3Scape comes with a library of over 100 3D objects and 800 textures. Perhaps taking a hint from Furniture.com's experience, the library includes furniture. However, the system can also import 3D models from external libraries. One-click publishing of content to the Web automatically generates an XML stream for communications between a firm's database and the end-user. Use of XML and server-side databases gives 3Scape an enterprise appeal to both internal groups, such as training departments and business-to-consumer applications. Since Mariah Vision builds enterprise solutions for its clients, delivering the needed plug-in is part of the service. However, Ambrosino says that the company is moving towards use of the XML-based Scaleable Vector Graphics standard. When support for XML Scaleable Vector Graphics is bundled in all major browsers, it will provide 3Scape an even broader reach. Mariah Vision's clients include DeWalt Tools and Scholastic Inc. Scholastic developed a Web-based classroom activity teaching children about immigration and used 3Scape as part of an interactive "fly-around" tour of Ellis Island allowing students to "handle" and explore objects and artifacts.

Bray's Visual Net is reminiscent of concepts promoted by Edward Tufte in works like "Envisioning Information." Visual Net uses cartographic techniques of communicating information in map-like views, using shapes, sizes, and colors to maximize the amount of information communicated. Instead of seeing lists of topics, you see two (or three)-dimensional views that you can browse to zero in on what you need. All collections and databases can be unified onto one mapping system. Databases, such as OPAC, periodicals, audio, video, and special collections, are all accessible and searchable from one visually-oriented Web site. It really doesn't matter what kind of content you are mapping; if metadata catalog information about the data set is available, Visual Net can draw browsable maps. The content itself can be traditional paper books, research databases, ebooks, images, sound, or video.

Visual Net applies patent-pending data visualization techniques to any hierarchical database. But don't think that this software stops at merely providing a visual overview of data-based content. Antarcti.ca's Visual Net mapping software has been applied to the Universal Description, Discovery, and Integration (UDDI) Registry, a sort of business yellow pages. After extensive testing with Microsoft's UDDI test database, Visual Net is now fully implemented on the production database. Visual Net can enhance users' abilities to navigate through and find desired information in the online directory of business services, such as location and relevancy to the user.

The October 29 release of Visual Net adds visual assists to searching, overcoming the limitations of one-dimensional hit lists in traditional search systems. Visual Net provides a map-like picture of categories of information and now shows the number of hits occurring in each region of the map. What's more, Bray maintains that Visual Net could use the predefined folder structure set up with Documentum as a map, for example. Visual Net could also use the growing lists of taxonomy efforts in various vertical industries. For example, a bio-pharmaceutical company could have Visual Net build its visual map from MeSH biomedical metadata or SNOMED out of the National Library of Medicine. Lastly, Visual Net could use taxonomies built by the likes of Semio's Autonomy. Since Bray is co-inventor of XML, Visual Net also supports SVG. He wants to avoid the plug-in problem though, and will use SVG when leading browsers support it natively. I guess you could say that the Web's prospects for this new visual paradigm are "looking up."