Issue
Article
Vol.28 No.1, January 1996
Article
Issue
Issue |
Article |
Vol.28 No.1, January 1996 |
Article |
Issue |
We started this 11/2 day workshop at CHI'95 with the purpose of exploring issues and challenges in integrating various existing techniques that have been developed to support user-interface design processes. The goal of the workshop was to develop a conceptual map for these techniques, which will guide user-interface designers to determine which support techniques are appropriate in which design situations, and how they can be integrated. The focus during the construction of this map was on finding suitable classification schemes for the various techniques.
Motivations and Goals of the Workshop
Background
User interface design is a knowledge-intensive task. Design requirements are given as a form of specification in terms of the application domain. User-interface designers map the specification (which is often fluctuating) to an interface design by taking into account a variety of design knowledge including but not limited to: (1) recommendations or quasi-standards of the application domain, (2) corporate-specific user interface guidelines or style-guides of both clients' site and software developers' site, (3) a set of general user interface guidelines, and (4) user interface standards if previous versions are already in use. While using the design knowledge, user interface designers cope with a huge amount of information during the design process. They have to consider the dynamic behavior of the system during use, the layout, choice and combination of media, and appropriateness for different cultural backgrounds. Required knowledge differs from one application domain to another, and body of knowledge is subject to constant changes. With the types and the amount of such knowledge, user interface designers cannot be aware of the complete set of knowledge that has to be considered during the user interface design process.
In response to the problem, knowledge-based approaches have been studied to support user-interface design. Typically, the design information that will be necessary during the user-interface design process is stored in a large database or a knowledge-base. The use of the stored "knowledge" varies in different approaches:
Individual approaches have reached a stage where they are about to leave or have already left research laboratories for industrial practical usage. And yet, none of these approaches have provided a "silver bullet" for issues and challenges of user interface design. At CHI'94, a panel was held to discuss the "best" interaction technique for user interface design. At the end of the discussion, the panelists as well as the audience came to the conclusion that each of the proposed techniques has both strengths and weaknesses and that a combined approach will result in a better support for the user interface designer than taking an individual approach.
CHI'95 was appropriate for having a workshop to discuss the integration of the approaches, with respect to both the expected participants and the point of time. In the past few years, several workshops have already taken place, each of which focused on a single knowledge-based supporting technique, for example a critiquing approach or a constraint-based approach. Strengths and weaknesses of each approach have been understood well enough to make the comparison among the approaches. And yet, no formal forums have been offered to discuss integrating the approaches. This workshop was held to find a way to combine the different approaches to overcome the weaknesses of the individual approaches.
We set our goal of the workshop as to develop a conceptual map that helps user-interface designers to determine which support techniques are appropriate in which circumstances, and how they can be integrated through addressing the following questions:
It was not expected that this workshop would deliver the solutions to the above questions. Instead, the goal was to establish a community that develops a shared understanding of the problems and of the existing approaches, and would help one another in contributing to a better integrated approach towards a system to support the user interface design process.
Prior to the workshop, we asked potential participants to write position statements to describe their current interests regarding knowledge-based support for a user interface design process. Twelve participants were selected. The selected participants demonstrated a nice variety of roles and of interests: theorists and practitioners, academia and industry, and interests and experiences for different supporting techniques, though we did not have anyone particularly interested in constraint-based supporting techniques.
We started the workshop by defining its scope. We then used a brain-storming session to enumerate existing user-interface design support techniques to identify classification schemes for the techniques. We have gone through the process of identifying key questions to ask during the classification. We used the HUMANOID system [Szekely, Luo, Neches 1992, 1993] as an example technique to classify to evaluate our classification scheme. We concluded the workshop by identifying future directions.
Right from the beginning we encountered issues in defining terminology:
With this, the scope of the workshop was more concretely defined as to support stakeholders in a process of interactive system design.
Using the brainstorming technique, we tried to identify as many interactive system design support approaches as possible. Then, we tried to roughly characterize each approach. The purpose of this activity was to identify what types of questions we ask ourselves in classifying the techniques, and not to develop the "right" categorization of the approaches.
Note that this is not an exhaustive list but only a list of approaches mentioned by the participants during the workshop. Also, since it was not the goal to classify systems and techniques "correctly," some categorizations may not be most suitable for characterizing the system or technique.
Our search for a classification scheme started with collecting all the criteria we thought important to describe the differences among approaches. Approximately thirty criteria (in question forms) were raised as used. After this initial step, we came up with one prime question per criteria group. Six questions were identified to classify supporting techniques and systems:
While going through the six prime questions, one more interesting question came up:
The above list was developed on a theoretical basis for the classification scheme. We have applied the classification scheme to some of the example approaches listed in the brain-storming session to test whether the scheme appropriately characterizes systems and techniques. Below is an example of applying the questions to the HUMANOID system. We used HUMANOID because some of the developers of the system were presented at the workshop.
The developers of HUMANOID at the workshop claimed that this was a fair description of their approach.
Based on the experiences with the example classifications (each participant did one as homework overnight), we decided that this scheme was too high-level and still not satisfactory. We thought that it might be interesting to define what is actually supported by each approach. "Supportable" steps during a user-interface design process include:
We have applied this new classification scheme to the approaches raised in the brain-storming session. Table 1 illustrates a summary of classifications. The result brought us the initial version of a conceptual map that can guide us to choose interactive system design support techniques.
Table 1: Example classifications of the techniques
------------------------------------------------------------------
Generation Evaluation Guidance Shift
towards of
Recognition Saving profitable Paradigm
of Solutions Steps subareas
------------------------------------------------------------------
Style-guides X X
Examples X X X X
Reusable X X
components
Constraints X
Critics X
GOMS X
Modalities X X X
Task/service X X
mapping
Transformation X X
components
Traceability X X
support
Empirical X
------------------------------------------------------------------
Using Table 1, we easily identified three research areas that need more work. First, we need techniques that can support the designer in recognizing already existing solutions, including reuse. Second, we need more support to guide the designer in refining a design problem into profitable sub-problems. Finally, more research needs to be done in helping designers to shift the paradigm to achieve a creative solution.
We held this workshop in recognition of the need to develop a deeper understanding of what the strengths of the different approaches are and how they can be combined to make them most useful for user interface designers. We used an assumption that combining the approaches increases usefulness as a prerequisite. We were successful not having any discussions about "which technique is better."
In summary, we partially achieved our goal of the workshop to develop a conceptual map, which describes which technique supports which design aspects. Although we did not have time to develop our understanding of necessary technologies to integrate different approaches, the classification schemes, the table and other outcomes of the workshop will help us to achieve the goal.
Kumiyo Nakakoji received her B.A. degree in computer science from Osaka University, Japan, in 1986, and her Ph.D. from the Department of Computer Science and the Institute of Cognitive Science at University of Colorado, Boulder in 1993, through a scholarship provided by Software Research Associates, Inc., of Japan. Her thesis work explored the role of a specification component in domain-oriented design environments and its use for fine-tuning the critiquing component. Nakakoji has published a number of papers in the topic of knowledge-based design environments and knowledge-delivery techniques. She was a co-organizer of the "Cross-cultural Issues in Human-Computer Interaction" workshop at InterCHI'93, and served as a design tool session discussant in CHI'94. Her recent research interests include knowledge-based approach for multimedia authoring environments and cross-cultural considerations in human-computer interaction design and in software engineering. Currently, she is an Adjoint Assistant Professor at University of Colorado and a visiting Assistant Professor at Nara Advanced Institute of Science and Technology while working for Software Research Associates, Inc.
Uwe Malinowski received a doctoral degree for a work on knowledge- based user support from the University of Braunschweig, Germany. He worked in the field of intelligent tutoring systems and adaptive help systems for the last six years. Since 1991, Malinowski has been a research scientist as a member of the adaptive user interface research team at Siemens Corporate Research and Development, Munich, Germany. He has organized several international workshops including "Adaptive User Interfaces -- Results and Prospects" held in Munich, March 1992, and on "Computer-Aided Adaptation" at InterCHI'93 in Amsterdam, May 1993. He has published several papers in this field and is a co-editor of the book "Adaptive User Interfaces -- Principles and Practice". From October 1993 to September 1994. Malinowski worked as a visiting researcher with the Human Computer Collaboration Group at the University of Colorado at Boulder.
Jonas Löwgren is Assisting Professor in the Human-Computer Interaction research group at the Department of Computer and Information Science, Linköping University, Sweden. Löwgren is a head of the research group that focuses on supporting professional usability-oriented design by studying design practice and developing methods, tools and techniques. He has been interested in knowledge-based tools for design support, primarily based on critiquing approaches for more than five years. He has published several papers in the field, addressing technical issues as well as the feasibility of design-support tools in professional practice. Löwgren participated in workshops at CHI'92 and InterCHI'93 and co-organized a workshop on the teaching of HCI at CHI'94.
Kumiyo Nakakoji, CB430 Department of Computer Science, University of Colorado, Boulder, Co. 80309-0430, USA; tel: +1 (303)492-3912, fax: +1 (303)492-2844; kumiyo@cs.colorado.edu
Uwe Malinowski, Corporate Research and Development, AG, Germany. Uwe.Malinowski@zfe.siemens.de
Jonas Löwgren, Department of Computer and Information Science Linköping University; jlo@ida.liu.se.
Issue |
Article |
Vol.28 No.1, January 1996 |
Article |
Issue |