This has been abstracted from a recent series Ramesh Jain has written on his blog. The EventWeb, as he and I look at it, is a fundamental upgrade from the DocumentWeb that we are familiar with. Todays World Wide Web is primarily built around documents. We have also referred to it as the Reference Web. As Ramesh put it in a recent interview: Todays Web is a document Web. Everything is presented as a page. Yet, audio and video are becoming easier to store and disseminate. Figuring out how well search through events is only one small piece of the problem. For example, as mobile phones become the primary client, people will use the devices to look for information: how do you design search tools for that platform?
The EventWeb will be the foundation for the Now Web. Rameshs first post deals with the importance of Events.
Two things are distinct in emerging applications of information systems: they contain vast amount of multimedia (both live and archived) data and attention is moving away from examining isolated silos of data toward more holistic pictures of evolving situations. Multimedia systems, including text, video, images, and audio, provide both information and experience related to a dynamic situation. As information and communication technology evolved, multimedia has become increasingly ubiquitous; structured data is now a very small fraction of useful data in emerging applications.
Current information tools are very good in dealing with entities, objects, and keywords. To address the needs of information management in dynamic multimedia environments, new concepts and techniques are needed. It is clear that the current concepts and tools are good for the text oriented and structured information systems. These tools are not good for dealing with images, video, audio, and other sensory information. A very good example of their limitation is the poor results that one sees on every major search engine for images and video. Since these search engines try to apply search tools effective for text to the text associated with images and video, but not processing images and video, their results, contrasting them to text search, are surprisingly bad.
Current information tools evolved before the wave of mobile phones, digital cameras, and broadband systems changed the landscape of information systems. With all these advances, experiences are becoming an integral part of information systems. In fact, we can already see signs of the dawn of experiential computing. There is a very intimate relationship between events and experiences in experiential computing, events will play a central role.
I believe that events may be used as fundamental organizational concept in multimedia systems that are becoming ubiquitous. There are strong and deep conceptual, engineering, computational, and human centered design reasons to consider events as a primary structure for organizing and accessing dynamic multimedia systems. We are developing multiple applications using event models to validate our hypothesis that events are effective in capturing multimedia semantics and building efficient systems to deal with multimedia information.
Tomorrow: EventWeb (continued)