Is it a lot of hype? Hypermedia approach to document processing.

The automation of document processing systems has been a central issue in information processing systems-especially since the proliferation of image technology and its use in the design of systems. We describe the effort of a Fortune 500 company to automate document processing systems consisting of both text and graphics. A prototype was designed using hypermedia features to enhance the capabilities and features of the system. We'll also discuss the differences in the systems development life cycle when designing image-based systems versus database systems.

Automation of business information processing began when computers were first available for use. With database management systems designed to support an organization's business activities, even more manual business systems became automated. These systems included payroll, inventory control, accounts receivable, etc. Normally database systems supporting a specific business activity are designed by incorporating into the system selected data from their manual domain.

Although such a design process provides the user with important and relevant data pertaining to that business activity, it is likely to leave behind other pertinent information. On occasions where such data are not available, the user is compelled to refer to original source documents.

This issue can be resolved by designing image-based systems which keep original documents as images and as natural base entities of the system.

In designing an image-based system the method proposed here is to capture the entire image of any business document (like a letter, invoice, report etc.). By capturing the entire document into the system we remove the potential for data loss. Thus, the entire business document is available for use "as is" by the user. This approach is especially helpful for businesses that have documents that are either complex or large, making it difficult to decide what information should be incorporated into the system and which part could be left behind.

However, image-based systems may require some form of a knowledge-base to support intelligent retrieval. In document processing systems where identifying documents requires semantic property definitions (i.e., definitions of the role and life of documents and images), the retrieval system should be similar to an expert system (Celentano, Fugini, and Pozzi, 1995).

Image-based information systems are systems that have a number of images stored together in an orderly manner. Helms defines an image as a collection of picture elements that transforms the original picture into an electronic image (Helms, 1990).