Building Intelligent Multimedia Interfaces

Hypermedia provides current authors with a new way to combine different media into single artefacts. These can be developed and marketed in the style of conventional media publications. Multimedia technologies not only bring the presentation of data stored as images, sound and video but also allow the incorporation of media specific input modes (e.g. natural language, speech, pen gesture), and the generation of images, sound and video from more abstract formats. Systems combining multiple input and output modes through an abstract representation are still a long way from the market, but parts of them (such as rules for automatically creating business graphics in spreadsheets) are being incorporated into existing products now. An example multimodal system (MMI2) is described to illustrate the technologies which could be brought to the market independently. A more direct approach to introducing intelligence into multimedia systems is to incorporate it into conventional hypermedia tools. The MIPS multimedia presentation system which combines the presentation of an open pre-authored hypermedia network stored in the HyTime standard format with dynamically created web nodes containing answers to conventional database queries is described. This illustrates how conventional hypermedia tools can be extended to include intelligent automatic generation of multimedia presentations from retrieved data.