Search and inference with diagrams

We developed a process for presenting diagrammatic information on the World Wide Web such that the information is searchable and accessible to inference engines. We implemented this process for chart diagrams. Generally, information that is presented diagrammatically is stored in an image file. Typical search engines have access only to image tags and surrounding text. Hence, the information presented in such image files is not accessible to search engines. The process we developed stores information in XML documents conforming to an XML Schema. A server-side rendering engine is called to present the information to the user graphically. The software system we developed enables us to describe, store, and present bar, line, and pie charts. We feel that our process enables us to get the best of both worlds: people like certain information to be presented diagrammatically and simultaneously want to locate such information with search engines and use inference engines to extract information which is implicitly contained in the data.