PROJECT DESCRIPTION
DHymE (Dynamic Hypermedia Engine): Support for Computational Applications

Michael Bieber
Institute for Integrated Systems Research
New Jersey Institute of Technology


Project Web Site


Table of Contents
Abstract
Description
Progress
A Final Word
References


Abstract

Back to Table of Contents

This research will produce a hypermedia engine uniquely designed to provide hypermedia functionality (e.g., annotation, navigation, guided tours) to computational applications, such as decision support systems, with minimal or no alterations. People use computational applications primarily for their underlying analytic functionality, not for navigating among large amounts of display information-the focus of most hypermedia systems. Hypermedia direct access increases user comprehension and improves application effectiveness. The engine uses logical rules called bridge laws to represent the application's internal structure and effect dynamic mapping of application components to hypertext components.


Description

Back to Table of Contents

Over the past several years, Dr. Bieber has specialized in integrating hypertext functionality into nonhypertext-oriented "computational" applications [ KP90, Bi92, BK92, Bi95, BK95, WB95]. (We do not distinguish between the terms hypertext and hypermedia here.) Many scientific and business applications are computational as opposed to display-oriented. Examples include accounting applications, computer-aided design systems, computer-aided software engineering environments, decision support systems, expert systems and statistical tools. People use computational applications primarily for their underlying analytic functionality, not for reading or navigating among large amounts of display information. Despite the predominance of computation-oriented applications, the vast majority of hypertext research and systems address only display-oriented domains. Yet, hypertext functionality can enhance an application's computational power. Hypertext is the science of relationships. Managing the myriad of interrelationships among an application's knowledge (data and calculated information) and enabling users to add to this set of interrelationships can increase user comprehension and thereby improve application effectiveness. Hypertext functionality comprises navigation, annotation and view-oriented features [BK95]. Navigation features include link traversal (browsing), backtracking [BW94] and query. Annotation includes comments and marking special positions (as "bookmarks" or in a "hot list"). View-oriented features include global and local overviews of the application, component views (akin to database views), and trails and guided tours of an analysis or of noteworthy objects [Bi92, Wa96]. Augmenting an application with hypertext functionality results in new ways to view its knowledge, navigate among items of interest, and annotate comments and relationships.

Our current research goal is to develop a hypermedia engine that runs concurrently with other applications and automatically gives them hypermedia functionality. The engine makes two major contributions. First, it maps application objects and functionality to hypertext constructs upon demand at run time based on the application's internal structure. This allows the engine to serve computation applications that dynamically generate their content (e.g., calculations within a modeling system or explanations within an expert system). Second, one shall be able to integrate the engine with minimal changes to the application [Bi95].

The engine intercepts messages between an application's computational features and its interface, automatically detecting the location of interrelationships in messages and documents based on its knowledge of the application's internal structure. This may involve arbitrarily complex inference based upon the current user and task. The engine represents the application's internal structure through bridge laws. Bridge laws are logical rules matching each of an application's constructs (e.g., object classes) and functions to its hypertext equivalent [Bi92, BK94, WBWN94, Bi95, WB95, WBWN95, Wa96]. The engine uses bridge laws to infer dynamically which message contents have interrelationships, automatically making these into hypertext "buttons". When the user selects one of these buttons, the engine again uses bridge laws to infer the available "links" to related application objects, appropriate application commands and any user-specified annotations, making all these accessible to the user. Within an expert system, for example, the engine would infer the expert system rules which use a particular data variable, and which other rules a particular rule can influence during execution. Within a decision support system, for example, the engine would infer the models which use a particular data variable and the origins of calculations. Whenever a data variable or rule is to be displayed within the expert system, or a variable, model or calculation is to be displayed within the decision support system, the engine automatically would map a button to it, giving users direct access to its interrelationships [Bi92].

Several research efforts have succeeded in creating links among primarily non-hypermedia information systems and in facilitating their traversal. Systems such as Microcosm [RS92], PROXHY [KL91] and Kacmar's [Ka95] each has a hypermedia engine that executes concurrently with external systems and provides linking support. These systems, however, concentrate primarily on supporting display-oriented, as opposed to our domain, computation-oriented applications. They either support only manual linking or provide automated linking based on an object's display value only (e.g., keyword look-up) as opposed to the object itself. In analytic applications different objects often share the same numeric display values and an object's display value may vary under different conditions. Another approach would be to use hypertext "storage engines" or "link servers" called "hyperbases" such as Hyperform [WL92] and HB3 [LS94]. While very useful in many situations, these require developers to alter existing applications significantly and develop new ones with a hypertext-oriented design; this defeats our research goal of augmenting nonhypertext applications with minimal design changes.


Progress

Back to Table of Contents

This research builds upon our prior success. Dr. Bieber developed the hypertext functionality and interface for a proof-of-concept prototype hypertext-supported decision support system (without bridge laws) called Max, illustrated in [Bi92, BK92]. He has developed a preliminary architecture of a more flexible engine incorporating bridge laws, described in [Bi95]. One of his Ph. D. students is adapting a scaled-down version of this architecture for Microsoft Access [Wa96]. Last Fall Dr. Bieber spent several weeks at the University of Southampton, where Microcosm was developed, to investigate building the hypertext engine using Microcosm as a basic platform. It turns out that, although heretofore only used for display-oriented applications, Microcosm's architecture can support much of the dynamic, computational environment in which we plan to conduct our research. Because Microcosm already has an established communications architecture and connections to several front-ends (such as Microsoft Word, other text-based viewers and several multimedia viewers), we hope to save much development time. One extends Microcosm by adding new modules (called "filters") to the existing module base. We currently are developing a prototype implementing the computational hypertext engine as a set of filters written in Prolog and C for TEFA [Bh90], Max' model management analysis tool. We expect to finish a preliminary prototype by early Fall. We also are planning a suite of additional filters adapting hypertext features (navigation, annotation and view-oriented) specifically for computational applications that dynamically generate information. By year's end we expect to implement multi-window backtracking [BW94], annotation of computational objects and guided tours [Wa96]. Little research exists to date on how to design or develop such functionality for computational environments (see [Bi92, BK95]). We are developing an infrastructure to experiment with and develop hypertext support for computational applications.


A Final Word

Back to Table of Contents

We believe this research will have a major impact on the development of computational applications, making them more intuitive, approachable and comprehensible to users. Giving users direct access to the interrelationships and exact information they want should increase application effectiveness. This line of research also will affect the hypermedia research community, as well as developers of hypermedia-supported applications in industry. Along with a handful of other researchers, we actively are trying to direct this community towards the huge base of computational applications that constitute a large percentage of today's computer applications [Is93, BK95]. Hypermedia functionality indeed can serve this domain successfully.


References

Back to Table of Contents

[Bh90] H. Bhargava, A Logic Model for Model Management, Ph. D. Dissertation, The Wharton School, University of Pennsylvania, Philadelphia, PA 19104, 1990.

[Bi92] M. Bieber, "Automating Hypermedia for Decision Support," Hypermedia 4, No. 2 (1992) 83-110.

[Bi95] M. Bieber, "On Integrating Hypermedia into Decision Support and Other Information Systems" Decision Support Systems 14, 1995, 251-267.

[BK95] M. Bieber and C. Kacmar, "Designing Hypertext Support for Computational Applications," Communications of the ACM 38, No. 8 (1995).

[BK92] M. Bieber and S.O. Kimbrough, "On Generalizing the Concept of Hypertext," Management Information Systems Quarterly 16, No. 1 (1992) 77-93.

[BK94] M. Bieber and S.O. Kimbrough, "On the Logic of Generalized Hypertext," Decision Support Systems (11), 241-257.

[BW94] M. Bieber and J. Wan, "Backtracking in a Multi-window Hypertext Environment," European Conference on Hypermedia Technologies'94 Proceedings (Edinburgh, September 1994), ACM Press, 158-166.

[DK94] H. Davis, S. Knight and W. Hall, "Light Hypermedia Link Services: A Study of Third Party Application Integration," European Conference on Hypermedia Technologies'94 Proceedings (Edinburgh, September 1994), ACM Press, 41-50.

[Ka95] C. Kacmar, "A Process Approach for Providing Hypermedia Services to Existing, Non-hypermedia Applications," Journal of Electronic Publishing: Organization, Dissemination and Design (forthcoming).

[KL91] C. Kacmar and J. Leggett, "PROXHY: A Process-Oriented Extensible Hypertext Architecture," ACM Transactions on Information Systems 9, No. 4 (1991) 399-419.

[KP90] S.O. Kimbrough, C. Pritchett, M. Bieber and H. Bhargava, "The Coast Guard's KSS Project," Interfaces 20, No. 6 (1990) 5-16.

[Is93] T. Isakowitz, "Hypermedia in Information Systems and Organizations: A Research Agenda," Proceedings of Twenty-Sixth Annual Hawaii International Conference on System Science (HICSS) (Maui, Jan. 1993), IEEE Press, Washington, D.C., Volume III 370-379.

[LS94] J. Leggett and J. Schnase, "Viewing Dexter with Open Eyes," Communications of the ACM 37, No. 2 (1994) 76-86.

[RS92] A. Rizk and L. Sauter, "Multicard: An Open Hypermedia System," Proceedings of the ACM Conference on Hypertext (Milan, Nov. 1992) 4-10.

[Wa96] J. Wan, Integrating Hypertext into Information Systems through Dynamic Linking, Ph. D. dissertation, New Jersey Institute of Technology, CIS Department, Newark NJ 07102, 1996.

[WB95] J. Wan and M. Bieber, "A Logic-based Approach to Integrating Hypertext and Information Systems," submitted to Decision Support Systems.

[WBWN94] J. Wan, M. Bieber, J. Wang and P. Ng, "Document Management Through Hypertext: A Logic Modeling Approach," Proceedings of the Twenty-Seventh Annual Hawaii International Conference on System Sciences (Wailea, Maui; January 1994), IEEE Press, Washington, D.C., Volume III, 558-568.

[WBWN95] J. Wan, M. Bieber, J. Wang and P. Ng, "LHM: A Logic-based Hypertext Data Model for Integrating Hypertext and Information Systems," Proceedings of the Twenty-Eighth Annual Hawaii International Conference on System Sciences (Wailea, Maui; January 1995), IEEE Press, Washington, D.C., Volume III 350-359.

[WL92] U. Wiil and J. Leggett, "Hyperform: Using Extensibility to Develop Dynamic, Open and Distributed Hypertext Systems," Proceedings of the ACM Conference on Hypertext (Milan, Nov. 1992) 251-261.