Issue
Article
Vol.30 No.4, October 1998
Article
Issue
Issue |
Article |
Vol.30 No.4, October 1998 |
Article |
Issue |
The goal of this workshop was to explore emerging theoretical foundations of design, use and evaluation of interactive hypermedia systems.
With the advent of cheap and powerful multimedia computers and hypermedia authoring tools, more and more information that was traditionally disseminated in the printed form are being made available in the hypermedia format. The rate of creation of hypermedia information, both on the World Wide Web and as CD-ROMs, has far out paced developments in the theory of how to design such systems to best facilitate users' comprehension of, and navigation within, the information being presented. Despite (or perhaps, because of) the commercial success of these systems, their design has largely been guided by common sense and intuitions rather than theory. This workshop was convened to address the situation of practice in the marketplace getting ahead of the development of theoretical foundations. The proliferation of hypermedia systems indicate an urgent need for consolidating basic scientific research and developing theoretical foundations for the design, use and evaluation of such systems.
Printed books allow sequential, random, and semantic (based on the table of contents or the index) navigation. A hypermedia system can provide these three modes of navigation as well. However, unlike a printed book, hypermedia systems can present dynamic information through continuous media such as video or animations. Furthermore, through hyperlinks, the designer can build in additional ways of semantic navigation. This can provide total freedom to the user to navigate through the system in non-linear ways. Does such freedom aid or hinder comprehension? There are many studies indicating that the flexibility of navigation in electronic worlds can confuse or disorient users, hindering the system's effectiveness. On the other hand, constraining navigation to confirm to linear sequences does not seem advisable either. Do theories exist that allow a designer to find a good middle ground?
Some of the fundamental questions faced by every hypermedia designer are the following. On what basis can one choose to present information using different media (text, photographs, diagrams, animations, video, etc.)? How does one decide on the temporal and spatial synchronization of various components? How can a complex hypermedia system be structured (tree vs. network structure, broad and shallow trees vs. narrow and deep trees, etc.)? How can the different components be tied together with hyperlinks to provide both a comprehensible overall structure and flexible navigation and interaction facilities to users? One kind of answer to these, that of basing the design primarily on the printed counterparts and intuitions, is what is found in most current systems. Are there more well-founded answers? Beginning an exploration of this question by a small group of computer and cognitive scientists was the primary achievement of this workshop.
This was a one and a half day workshop. It opened with Helen Gigley of the Office of Naval Research (ONR) providing a brief overview of why ONR is funding research on hypermedia and multimedia. This agency is very interested in using multimedia and hypermedia technology to construct Interactive Electronic Training Manuals. According to her, current systems used by the Navy are mere databases which, from an HCI point of view, are no better than electronically linked index cards.
Then Alistair Sutcliffe from City University compared and contrasted two approaches -- reusable patterns and cognitive models -- to understanding and modeling hypermedia interaction. He suggested that developing hypermedia design patterns may be a quick way to get design wisdom to practitioners, but the patterns should ultimately be grounded on empirically validated cognitive models.
This was followed by three theme-oriented sessions on Design Principles, Evaluation: Tasks, Methods & Experiments, and Architecture & Navigation. Each session consisted of presentations followed by often heated discussion. The workshop concluded with an open discussion session that focused on two key questions: What did we learn? Where do we go from here?
This session began by Fabio Nemetz enumerating a set of hypermedia design problems that await the development of sound theoretical principles -- naturalness and realism, media allocation, redundancy, significant contributions of media, exploration, and quality of information representation.
This was followed by Richard Mayer and Roxana Moreno providing some of those very principles, the lack of which was bemoaned by Nemetz. This presentation distilled several years of research into educational uses of multimedia into a set of fundamental, empirically validated and very common sense principles:
Multiple representation principle: it is better to present an explanation in words and pictures than solely in words.
Contiguity principle: when giving a multimedia explanation, present corresponding words and pictures contiguously in space and time rather than separately.
Split attention principle: when giving a multimedia explanation, present words as auditory narration rather than as visual on-screen text.
Individual differences principle: the foregoing principles are more important for low knowledge than high knowledge learners, and for high spatial ability rather than low spatial ability learners.
Coherence principle: when giving a multimedia explanation, use few rather than many extraneous words and pictures.
Maria Costabile and Maristella Matera discussed their group's research addressing the problem of a lack of hypermedia-specific usability evaluation techniques. They showed how abstract and concrete tasks for usability evaluation can be developed from a framework called SUE (Systematic Usability Evaluation).
James Blustein considered the problem of translating paper documents to hypertext -- what are some of the useful link design strategies and evaluation methods? He proposed that common tasks of scanning, browsing and querying can be effectively supported by the automatic construction of structural, definitional and semantic links, and outlined an evaluation plan to verify this claim.
Mary Hegarty discussed three empirical studies that compared learning outcomes from studying how a household device (a flushing cistern) worked using a hypermedia manual and paper manuals. Interestingly enough, it appears that when descriptive and depictive information presentations are designed according to a multimodal comprehension model, whether the information is presented in a static medium (paper) or a dynamic medium (hypermedia) does not make a big difference!
Peter Pauen discussed research on HyDev, a hypermedia application development model intended for the analysis and design phases of development. It captures requirements with three models (a domain model, an instances model and a representation model), which improves upon extant models such as the Amsterdam Hypermedia Model and the Object Oriented Hypermedia Design Model.
Vivek Ram discussed his ongoing research into dynamically generating hypermedia links from underlying semantic networks of information and a blackboard architecture containing domain specific expert knowledge modules.
Paul Maglio postulated that people conceptualize navigation in hypermedia information spaces in ways similar to how they conceptualize navigation in physical space, using notions such as anchor points and familiar routes. This implies that notions of agency, embodiment and immersion may be central to the design of future interfaces to hypermedia information spaces. A cognitive linguistic analysis of the utterances of web surfers and an empirical study of their navigation patterns provided credence to this claim.
Most attendees felt that this was the most productive phase of the workshop. The discussion elicited a set of collective opinions regarding the future of hypermedia research and a picture of the building blocks needed to construct a solid theoretical foundation for hypermedia (see figure 1).
Significant open issues are the large gap that exists between cognitive models of hypermedia use (i.e., models of comprehension, navigation, interaction, etc.) and operational design support based on these models (such as design guidelines or patterns), as well as the large gap that has to be bridged in order to construct effective hypermedia systems from extant design guidelines.
First, models at multiple levels and multiple kinds of models (e.g., cognitive models, task models, software engineering models) are needed. There are several hard issues in modeling hypermedia interaction. For example, should one focus on developing general cognitive models with wide applicability (but low utility, since general models do not provide strong constraints to the designer of a task-specific application), or task-specific models with high utility (but applicable to only a specific hypermedia task)? How can individual differences among users be incorporated into such models? Second, explicit hypermedia design guidance, justified by validated models, are what practitioners urgently require. Third, design evaluation studies that empirically demonstrate the effectiveness of design guidance are also woefully lacking. Computational and cognitive models, and design guidance derived from these models are needed to help designers prune the large design space of hypermedia.
It was agreed that more research into reusable, composable and customizable user interfaces for hypermedia systems is needed as well. Individual differences among users do matter; this calls for research into easily customizable or self-adapting hypermedia systems. Hypermedia systems engineering is more than mere software engineering since HCI is such a central issue in hypermedia design. So software engineering models and methods by themselves are inadequate for the design of effective hypermedia systems. While much more work on empirically-based models of hypermedia interaction remains to be done, there is also an immediate need (particularly in the light of technology and commercial practice getting far ahead of theory and evaluation) to get current design wisdom out of the labs and into the hands of practitioners.
Figure 1: Theoretical Foundations of Hypermedia
A poster summarizing the main ideas that emerged from the workshop was exhibited at the main conference.
These papers may be downloaded from the workshop web site at http://www.eng.auburn.edu/cse/research/vi3rg/WS.html
Evaluating automatically generated hypertext versions of scholarly articles, by J. Blustein.
Abstract tasks and concrete tasks for the usability evaluation of hypermedia applications, by M. F. Costabile, F. Garzotto, M. Matera and P. Paolini.
Towards a theoretical space for hypermedia systems, by N. Dahlback.
Designing hypermedia manuals to explain how machines work: Lessons from evaluation of a theory-based design, by M. Hegarty, J. Quilici, N. H. Narayanan, S. Holmquist and R. Moreno.
A position paper by P. P. Maglio.
A cognitive theory of multimedia learning: Implications for design principles, by R. E. Mayer and R. Moreno.
Towards principled multimedia, by F. Nemetz and P. Johnson.
Modeling of hypermedia applications with HyDev, by P. Pauen, J. Voss and H-W. Six.
Negotiated links in open hypermedia systems, by V. Ram.
Theories, models and patterns: Which way forward to understand hypermedia interaction?, by A. Sutcliffe.
James Blustein, University of Western Ontario
Maria Francesca Costabile, Universita' di Bari
Helen Gigley, Office of Naval Research
Mary Hegarty, University of California at Santa Barbara
Selma Holmquist, Auburn University
Paul P. Maglio, IBM Almaden Research Center
Maristella Matera, Politecnico di Milano
Richard Mayer, University of California at Santa Barbara
Roxana Moreno, University of California at Santa Barbara
N. Hari Narayanan, Auburn University
Fabio Nemetz, Queen Mary and Westfield College
Peter Pauen, FernUniversität Hagen
Vevek Ram, University of Natal
Alistair Sutcliffe, City University
N. Hari Narayanan, Visual Information, Intelligence & Interaction Research Group, Auburn University (Chair).
Helen Gigley, Cognitive and Neural Systems Division, Office of Naval Research.
Wendy Hall, Multimedia Research Group, University of Southampton.
Mary Hegarty, Department of Psychology, University of California at Santa Barbara.
Yvonne Rogers, School of Cognitive and Computing Sciences, Sussex University.
Alistair Sutcliffe, Center for HCI Design, City University.
Yvonne Wærn, Department of Communication Studies, Linkoping University.
N. Hari Narayanan is a faculty of Computer Science & Engineering at Auburn University and an adjunct faculty of Computing at Georgia Institute of Technology. His research interests include cognitive model-based hypermedia design and evaluation, and navigation in large hypermedia information spaces. More details on his research can be found at the web site of the Visual Information, Intelligence & Interaction Research Group: http://www.eng.auburn.edu/cse/research/vi3rg/vi3rg.html
N. Hari Narayanan
107 Dunstan Hall
Auburn University
Auburn, AL 36849
USA
email: narayan@eng.auburn.edu
Tel: +1-334-844-6312
Issue |
Article |
Vol.30 No.4, October 1998 |
Article |
Issue |