30th Anniversary SOFSEM 2004        

Abstract:

World Wide Web is a highly distributed and dynamic information environment. Users access information through a variety of devices, performing a wide spectrum of user tasks. This poses many challenges for designing effective information services and applications. Addressing these challenges requires a good understanding of Web characteristics and their implications on the design and usability of services and applications.

For example, it is important to note that, in the Web environment, processing of users' requests for information and delivery of documents in response to these requests are typically two disjoint processes. Indeed, services like Web search and online directories typically provide the user with URLs of information sources rather than the documents themselves. Documents are hosted on Web site servers, out of control of the service itself. This separation places a particular importance on the design of the user's interface to the Web, the Web browser, which needs to provide the bridge between information services and document delivery.

Furthermore, authoring of the Web content and collection of that content by centralized services (e.g., through Web crawling) are typically completely disconnected processes. This implies that services create and use data representations that are out of date. Moreover, because of the sheer scale of the Web, analyses of the Web data that services and applications can apply are rather simplistic. As a consequence, information they provide to the user about the content and structure of Web sites is typically suboptimal.

Use of Web on mobile devices (e.g., Personal Digital Assistants (PDAs) and Internet enabled mobile phones) poses further challenges. Viewing Web pages of complex layout structure is incompatible with the restricted screen space. This can lead to disorientation during page viewing and difficulty with identifying relevant parts of the page content.

In this presentation, we discuss findings of user studies that point to issues which arise within Web environment. We promote the idea of generating information about the Web site and page structure at the authoring or publishing stages, and providing that information to applications and services upon request. We describe the MIDAS framework (Meta-Information Delivery and Annotation Services) which we implemented to enable creation, distribution, and utilization of metadata for describing Web structure and contents. We illustrate how MIDAS can be used to enhance the user's experience during search and navigation and highlight the way it complements the Semantic Web effort.