Information Storage and Retrieval of Composite Documents: A User Oriented Model of a Computer Message System

Fox, Dr. Edward A. (1985) Information Storage and Retrieval of Composite Documents: A User Oriented Model of a Computer Message System. Technical Report TR-85-34, Computer Science, Virginia Polytechnic Institute and State University.

Abstract

This paper outlines a new model of computer message systems based on a useroriented rather than a communications perspective. Most network users spend their time working with various types of 'composite documents' such as mail messages, news, reports, or entries in directories of names. They perform operations that are aided by a logically centralized but physically distributed world-wide information storage and retrieval system. Their concerns should be with creating, filing, submitting, searching, and retrieving documents, not with routing or transmission matters. Computer networks are rapidly proliferating. Their number and size are increasing, and interconnection is commonplace. High vol~ime applications like mail and news transmission are well established in offices, large corporations, and governmentally supported internets. Computer conferencing and directory assistance services are continuing to evoive. Standards for message formats, transmission protocols, and name/address directories have been recommended by DARPA, NBS, and CCI'IT. Research in document modeling, distributing databases, and retrieving text objects is being adapted to improve the functionality of computer message systems. Advanced techniques for text analysis, automatic indexing, query construction, description of comprehensive interest profiles, and retrospective or current awareness retrieval can simplify and improve the effectiveness of user agent software. Initial experimentation has shown the value of recognizing the structure of composite documents. Current work with the CSNET name server database, and with a collection of messages extracted from AIList digests distributed over the DARPA Internet, indicates that the proposed model accomodates mail, news, and directory assistance. Other work with book passage retrieval suggests that large files can be processed too. It is hoped that integration of these services, along with conferencing and database handling, will occur, as a new network for the Virginia Center for Innovative Technology is proposed and eventually implemented.