Joonhee Yoo
Abstract: Few designers explicitly think about their applications' interrelationships. Designers appear not have a deep enough conceptualization of their domains to identify intuitive relationships and realize the full scope and interconnections within domains. RNA (Relationship-Navigation Analysis) gives designers and developers an analysis tool to think about an information domain in terms of its interrelationships. RNA incorporates a complete taxonomy of generic relationship types that would apply to any application domain.Categories and Subject Descriptors: D.2.1.[Software Engineering]Requirements/Specifications - Elicitation methods, methodologies, H.5.4.; [Information interfaces and presentation] Hypertext/Hypermedia -Theory, ...
General Terms: Design, Theory
Additional Key Words and Phrases: Hypertext, Hypermedia, Relationship Taxonomy, Relationship Analysis, Links, Navigation
How often do you want to point to an object on the computer screen and say "This looks interesting. Tell me more about it. What is it? How can I use it? What do I need to know to use it? Can I modify it? What can I do once I have deployed it? How does it differ from other similar objects?" These are all relationships, and we should be able to access them.
Yet, few designers explicitly think about their applicationsí interrelationships and whether users should access and navigate them directly. Why not? In part, it has not occurred to many designers and developers to incorporate hypermedia functionality. (Hypermedia can be viewed as the science of relationship management - structuring and navigating information through relationships, i.e., links.) Most designers and developers do not have a hypermedia mindset; they and their users have seen few examples and do not demand this functionality yet. In part, people do not have the time to reengineer existing applications, especially when migrating them to the World Wide Web with other projects waiting in the wings. Developers also have few tools and techniques for designing and incorporating hypermedia functionality easily. Developers will not do this until it is natural to conceive and easy to implement.
Science fiction writers fantasize about people who vividly can see the relationships among people and objects. Designers should do this naturally for their information domains. But they rarely do. When the first author did a relationship analysis on an application he had used extensively for several years, in a period of two hours he discovered over twenty useful types of information, which clearly should have been directly accessible within the system [Bieber]. As further anecdotal evidence, this happened again when in the space of a few hours, an expert in decision analysis greatly expanded his prototypeís design with several extremely intuitive relationships, which he had failed to see in years of previous work.
Designers appear not have a deep enough conceptualization of their domains to identify intuitive relationships. It seems they need help in realizing the full scope and interconnections within their domains. Yet, no approaches exist for determining the relationships within an information domain or system. Such an analysis should precede the design phase, which many of the hypermedia design methodologies address [Balasubramanian 1998,Christodoulou 1998,Garzotto 1993,Isakowitz 1995,Lange 1996,Schwabe 1996].
The goal of our research is to help designers and developers develop a hypermedia mindset by giving them an analysis tool, RNA (Relationship-Navigation Analysis), to think about an information domain in terms of its interrelationships [Yoo]. Hypermedia links represent relationships in applications. Once identified, our DHymE hypermedia engine can generate links for many of these relationships automatically [Bieber]. Over time, we aim to establish new standards for the application development process and for usersí interaction with applications. Designers should conduct a relationship analysis in order to more deeply understand the relationship structure of their applications and generate a full range of links to implement. Users should be able to point to any object of interest and find out whatever they want about it through these links.
A hypermedia mindset promotes both a design philosophy and a philosophy of maximum access based on relationships. Granting users full freedom to access and explore at will helps them better understand a domain as a whole and build confidence in application results. Under this philosophy, any element of interest to a user should be a candidate for linking. RNA supports a hypermedia philosophy of design, which embodies a systematic approach to realizing a philosophy of maximum access within computer applications [Yoo 2000].
The RNA approach has five steps: stakeholder analysis, element of
interest analysis, relationship analysis, navigation analysis and
implementation analysis. Designers employ RNA as a knowledge
elicitation or brainstorming tool, using it as a framework or set of
categories with which to examine their application and its
information environment. During stakeholder analysis designers
determine everyone who will use some aspect of the application. This
can produce surprising revelations. For example, when designing a
university class web site, we realized that stakeholders included the
department chair at evaluation time, high school students considering
our university and colleagues at other universities interested in
what we are doing. During element of interest analysis we determine
what each stakeholder might want to find out more about. Table
1 shows RNAís general relationship taxonomy. During the
relationship analysis stage, each generic relationship provides a
series of questions which designers can ask themselves about each
element of interest for a given stakeholder. During navigation
analysis, designers then think of each element of interest in terms
of how the stakeholder might usefully access it, e.g., as part of an
index, guided tour or customized trail. During implementation
analysis, designers do an informal cost/benefit analysis to decide
which relationships would be useful and feasible enough to actually
include in the application design. For some the information a
relationship may point to might be impractical to implement.
Designers should feel cognitively unbounded during the first four
stages of the analysis, putting off practical constraints to the last
stage.
|
Generic Relationship |
Internal |
Generalization/ |
|
|
Self |
Characteristic Descriptive Occurrence |
||
|
Whole-part/ |
Configuration/Aggregation Membership/Grouping |
||
|
Classification/ |
|||
|
External |
Comparison |
Equivalence Similar/Dissimilar |
|
|
Association/ |
Ordering Activity Influence Intentional Socio-organizational Temporal Spatial |
Table 1: RNAís General Relationship Taxonomy
Space constraints prohibit a full discussion of RNA here, so we illustrate with just a few example questions using the RNA relationship taxonomy during the relationship analysis stage. Suppose we are building a repository of multimedia components for on-line courses. The taxonomy's descriptive relationship might prompt us to ask whether course components have a description, definition, explanation, set of instructions, documentation or illustrations available, within or external to the system. Any or all could be implemented later as links. The taxonomy's occurrence relationship would prompt us to ask "What are all the course topics that would use this component. Show me all courses that currently use it." The attribute/parametric relationship prompts questions such as "What are the componentís inputs and outputs? Which media types does it use? What software do I need to execute it? What operating system does it require? How much space and memory does it need? How long does it take to use? How do I implement it? What assumptions underlie it?" The ordering relationship would prompt us to ask "what prerequisites exist for this component? What logically follows it for a given userís (student, teacher, courseware developer) purpose?" The similar/dissimilar relationship prompts the questions "Which other components are similar to this one, and in which ways?" The developer could decide to implement any or all of these later as links, but even if he (or she) did not, he would gain a deeper understanding of how components fit in that domain.
Our goal is to ascertain a complete taxonomy of generic relationship types that would apply to any application domain. Part of proving this will be to show that any other link or relationship taxonomyñdomain independent or domain specificñwould map clearly to ours. Then we need to conduct experiments and field studies to show that RNA indeed produces richer analyses, than if designers did no relationship analysis.
But we wonder if RNA has a deeper role, serving perhaps as a generic ontology of relationships in any complex systemñcomputerized or otherwise. How often do you hear a remark, see cursory mention of something in a magazine article, spot something interesting in a catalog, see a component in a diagram or blueprint, stumble across an item in a supermarket, and want to say "Tell me more about this!"? The longer we work on RNA, the more we find ourselves questioning the relationships of the objects and systems surrounding us in the everyday world. This hypermedia vantage point makes the world a richer place for us. Would that all our computer applications were this rich.
We gratefully acknowledge funding for this line of research by the NASA JOVE faculty fellowship program, by the New Jersey Center for Multimedia Research, by the National Center for Transportation and Industrial Productivity at the New Jersey Institute of Technology (NJIT), by the New Jersey Department of Transportation, and by the New Jersey Commission of Science and Technology, and by Rutgers University.
[Balasubramanian 1998] V. Balasubramanian, M. Bieber and T. Isakowitz, Systematic Hypermedia Design, Information Systems Journal, 1998. [on-line: http://www.cis.njit.edu/~bieber/pub/ISJ98/isj98.pdf]
[Bieber] M. Bieber, Supplementing Applications with Hypermedia, in submission. [On-line: http://www.cis.njit.edu/~bieber/pub/supp/supp.html]
[Christodoulou 1998] S. Christodoulou, G. Styliaras and T. Papatheodourou, Evaluation of Hypermedia Application Development and Management Systems, ACM Hypertext'98 Conference, Pittsburgh, 1998, 1-10.
[Garzotto 1993] F. Garzotto, P. Paolini and D. Schwabe, HDM - A Model-Based Approach to Hypertext Application Design, ACM Transactions on Information Systems 11(1), 1993, 1-26.
[Isakowitz 1995] T. Isakowitz, E. Stohr and P. Balasubramanian, RMM: A Methodology for Structuring Hypermedia Design, Communications of the ACM 38(8), 1995, 34-44.
[Lange 1996] D. Lange, Object-Oriented Hypermodeling of Hypertext Supported Information Systems, Journal of Organizational Computing and Electronic Commerce 6(3), 1996, 269-293.
[Schwabe 1996] D. Schwabe, G. Rossi and S. Barbosa, Systematic Hypermedia Application Design with OOHDM, Hypertext'96 Proceedings, March 16-20, 1996, 116-128.
[Yoo 2000] Yoo, J. and Bieber, M., Towards a Relationship Navigation Analysis, Hawaii International Conference on Systems Sciences, January, 2000. [On-line: http://www.cis.njit.edu/~bieber/pub/hicss00/hicss00.html]