The World Wide Web has changed our expectations of information access forever. We now take it for granted that we can search vast repositories of information with ease and at extremely low cost. Whereas we used to travel to the information repository to manually seek out what we sought we are now able to use the services of digital robots and vast digital indeces to seek out that information. But as with most activities on the Web the searching activity takes place in isolation. We are unable to communicate with the information provider about our needs. If we know what we are looking for and generally where it is we have a reasonable chance of finding it. If we don't know what we are looking for there is no one to turn to.

A new technology, XML, provides a possible mechanism for addressing this problem. XML, which is an evolutionary branch of SGML, can be used as a means of providing the "electronic reference librarian". XML's impact on Web information delivery is already being felt in "push" technologies which are based on "CDF", an implementation of XML.

This paper will explore how XML can be used:

• to move beyond turning the Web into a thousand speciality TV channels
• by institutions to inform Web surfers what the institution knows, how it organizes that knowledge in a meaningful way, and how it wants the user to access that knowledge base
• to create interactive digital dialogs that will guide a user through a knowledge base
• as a feedback mechanism that will help providers understand how users interact with their presentations.