Multivalent Documents: Documents as interfaces to information in networked digital repositories

Thomas A. Phelps and Robert Wilensky, Dept. of Computer Science, UC Berkeley

Seminar on People, Computers, and Design
Stanford University January 26, 1996


Digital documents thus far have have largely mirrored their pre-digital ancestors. As such, they provide little enhancement of non-digital counterparts other than searching and hyperlinks.

We believe that far greater functionality is desirable and possible. To achieve such functionality, we have been developing a new model of documents, called "multivalent documents". A multivalent document comprises any number of distinct, but intimately related, layers of content, along with small, dynamically loaded programs, called behaviors. In effect, each document becomes an interface to its particular set of contents. Multivalent documents are especially well-suited to the highly networked work, as the various layers of a document can reside on different machines, or even be created dynamically.

A prototype version of MVDs has been implemented in Java. This prototype allows access to our digital library collection of scanned page images, but with added functionality. In particular, users can search for terms in a page image, and have the resulting matches outline on the screen; they can select regions of the image and paste ASCII text created from the image by optical character recognition; some tables in the images have active properties, such as sorting themselves in response to mouse clicks. Other functionality under development includes distributed annotations and geographic information systems capabilities.


Tom Phelps graduated from the University of Illinois at Urbana-Champaign in 1990 with an A.B. Summa Cum Laude in Economics and a B.S. Highest Honors in Computer Science. He is now a graduate student in Computer Science at the University of California, Berkeley, where he works on the Digital Library Project. His thesis work concerns a new model for digital documents called Multivalent Documents; recently a paper describing this idea co-authored with Robert Wilensky was named Best Paper, Digital Documents Track at the Hawaii International Conference on System Sciences '96. He is also author of the popular graphical manual page browser TkMan. His interests include digital documents, the World Wide Web and Java, user interfaces and multimedia.

Robert Wilensky received his B.A. in mathematics from Yale College in 1972, and his Ph.D. in Computer Science from Yale University in 1978. Since that time, he has been on the faculty of the University of California at Berkeley, where he is now Professor and Chair of the Computer Science Division. Professor Wilensky founded BAIR, the Berkeley Artificial Intelligence Research Project, and the Berkeley Cognitive Science Program. He also serves as Director of the UC Berkeley/Hewlett Packard Science Center.

Professor Wilensky has published numerous articles in the area of artificial intelligence, planning, knowledge representation and natural language processing. He is the author of Planning and Understanding: A Computational Approach to Human Reasoning, BLISPCraft, and Common LISPcraft. His current interests include intelligent information access to electronic libraries.


Titles and abstracts for all years are available by year and by speaker.

For more information about HCI at Stanford see

Overview Degrees Courses Research Faculty FAQ