USER INTERFACES

Coordinator: Carl-Gustaf Jansson (calle@dsv.su.se) ________________________________________________________________________________

This program together with drafts of concrete project descriptions have been formulated by a group of researchers in the research groups: SICS/UACM, SICS/NLP, SICS/DCE, DSV/K2LAB, IT/Mediagroup and SISU/Interaction and communication group. A listing of projects with explicit links to the project descriptions are given at the end.

THE CHALLENGES

USER INTERFACES IN THE C&C VISION

This is a strategic research program in the area of user interfaces designed to fit within the Center for Computing and Communication (C&C). The vision of C&C is to contribute to the development of a future generation of telecommunication systems, that is artifacts that combines and generalizes the properties of todays telecommunication facilities, computers, computer communication networks and mediatechnology. In this vision we will no longer perceive entities as terminals, personal computers, telephones, fax machines, TV monitors, cameras etc. A variety of new devices will take the form of machine intelligent real world artifacts like tracking badges, toasters, pens and cars etc. The computation and communication devices will be embedded in the everyday environment and literally dissappear into the walls. We will see much more instinctively appealing ways of interacting that are similar to the procedures humans use when communicating with each other or when manipulating artifacts in the real world. All this should be contrasted with the user interfaces found in today's computer systems which are complex and often confusing. After some thirty years of development and evolution the designers of interfaces still seems to be catering more towards the demands of the computer itself than the goal of providing the user with a reasonable way of getting work done. Presently, there is always a significant cognitive load associated with accessing computer based applications.

THE PRIMARY CLASS OF APPLICATIONS

Future telecommunication systems will have a wide applicability, but the Center for Computing and Communication will focus on applications where life-long learning are integrated in everyday work. This should include: Methodological knowledge on how to build such applications can be found in more established application areas such as: It is for these kinds of applications and combinations of these that we primarily want to study how simple, adequate, robust and flexible interfaces can be designed, developed and maintained.

THE SCENARIO OF USE

The scenario of use that we want to focus on is the collaboration of many users distributed both in time and space. The users are typically a very heterogeneous group with respect to their age, cultural and social background as well as their competence with respect to specific tasks. Collaboration should be interpreted in a wide sense including both communication, collaborative learning, group decision making and the collaborative carrying out of concrete tasks. Considering the leaps in all these respects, going from a single user, using a single tool with a traditional user-interface for a specific task to a group of users collaborating under the generalized circumstances described will drastically change the demands on methods and tools for specification, design, development and evaluation of the user interface. The failings of todays interfaces become even more apparent in this scenario.

In the context of the initial technological vision we must also seriously consider the overall usability of the new artifacts (like the weight and duration of batteries), where the concept of user interface is strongly widened.

METHODOLOGY

FULL SCALE EMPIRICAL STUDIES

As a fundamental methodological approach we believe that true progress in field of user interactions can only be gained by realistic field tests, where specific groups of users with individual characteristics are studied in their normal working situation under realistic time and other resource constraints using a specific technological configuration of tools. There exist too many idealized studies where singular persons using singular tools under idealized circumstances. It is very unlikely that results from such studies can scale up to realistic situations. Useful results in the field of user interfaces are to a high degree «situated«, that is very much dependent on contextual factors. Only if such systematic studies can be carried out we will able to validate good guesses that are the result of more explorative work. In such studies also organizational and social changes caused by the introduction of the new tool must be acknowledged. This aproach is very much independent on the specific design choices that is possible and discussed below.

THE NEED FOR INTERDISCIPLINARY COMPETENCE

In a longer perspective persistent success in the field of user interfaces need a better integration of traditional fields like systems development, software engineering, artificial intelligence and human-computer-interaction. It is important that we also invite and accept a number of nontechnical subjects from the humanities, cognitive sciences, social sciences and medical sciences as equal or associated partners in this endeavour. Subjects like psychology, philosophy, linguistics, pedagogics and graphical design as well as comparative literature and studies in cinema and art can in different degrees contribute to the common goal.

METHODS AND TOOLS

The supposed drastically widened use of computing and communication devices will raise a demand not only for adequate knowledge on how to design and evaluate user interfaces but also for efficient methods for developing the great amount and wide variety of applications that will be requested. Methods of systems development in general have been specialized for the purpose and are used fragmentarically at present. What we need is a systematic use of methods and tools that take into acount all richness of a user interface envisioned above but also utilizes the restrictions to which we commit ourselves. Research is still needed to be able to develop such methods, within C&C in particular for for the choosen target applications.

ASPECTS OF DESIGN

METAPHORS FOR INTERACTION

After having discussed the framework in which we want to study user interfaces we now turn to the more specific design issues involved, and we start with what we prefer to call the metaphors of interaction, that is the overall perspectives on the objects and participants involved in the communicative process. A familiar metaphor extensively used today is the desktop metaphor where interaction is viewed as the movement and manipulation of objects on a desktop with some auxiliarys like the wastebasket. This metaphor has effectively competed with a more traditional metaphor like the file metaphor or the the set and function theoretic model.

We want to focus on the following metaphors:

Many user interfaces uses a mixture of metaphors and the chartering of the effective combinations of metaphors is an important research issue. Inn particular we want to focus on combinations of instantiations of the «society of agents and the «spatial« metaphor.

MODALITIES OF INTERACTION

Another important aspect of the design is the appropriate combination of communication media and their modes of use. A multimedia interface is the combination of heterogeneous media in a coordinated discourse including presentation as well perception. The individual media may be structured text, static or dynamic graphs or pictures, video, sound including voice and music as well as more exotic variants like tactile(touch) or olphatellic(smell). Hypermedia is the organization of fragments of multimedia information in a network structure. Use of more structured forms of organizations for multimedia, such as object oriented schemes, are gradually increasing. By mode of communication we mean a scheme of discourse typically using one single medium but possibly multimedia. Examples of modes of communication are textual electronic mail, one way video, interactive video or 3D workspaces. By a multimodal interface we mean a computer supported discourse which is carried out by the combination of several modes of communication. To work properly all expressions in a multimodal communication, regardless of the modality they were originally created in, must assigned the same type of interpretations. Natural language technology has power to contribute to the handling of multimodal interfaces as it does not only treat the formal qualities of language (syntactic and semantic) but also the general processes of communication and dialogue in which language use is embedded. For the novel integrated telecommunication services which C&C targets the design of appropriate multimodal interfaces is of particular importance. Examples of research question related to multimodal interfaces are:

ADAPTIVITY

To handle the problem of heterogeneous user groups and context dependent work, adaptive interfaces are needed including both the manual tailoring of interfaces to users needs during the design and maintenance phases, and the automated and dynamic adaption of interfaces during use. Research topics for adaptive interfaces include:

TRANSPARENCY versus FILTERING TECHNIQUES

At present the paradigm of direct manipulation dominates user interface design. This urge for transparency and total user control has succeeded the earlier generation of opaque user interfaces of a more traditional discourse character. The direct manipulation paradigm has many advantages, but for the kind of applications C&C conceives, the overwhelming amount of information is a powerful threat against the usability of the interface. Many users, especially novices, need filtering mechanism that can provide them with an extract of information that fullfills their current need. Such a filtering mechanism can of course be composed of many simpler transformations like:

HELP,EXPLANATION AND POWERFUL REPRESENTATIONS

When designing interfaces directed at heterogenous groups of users, we must also be able to provide help, explanations and in some cases even tutoring in learning and handling the tools we provide. Those functionalities are a necessity when the average user is a novice. To be able to adapt the help and the explanations to particular users or user groups is also important. A prerequisite for designing adequate helpfunctions, but also for designing filtering functions and adaptive mechanisms in general is to have powerful representations for both domain knowledge and knowledge about the user and the discourse process. To be more concrete we need:

MEDIATECHNOLOGY

To close the circle, research on user interfaces in C&C should also include research on media technology to contribute to the development of the new artifacts envisioned in the first section. Advanced media technology research should include work on

SUMMARY

Throughout each section of this research program a number of commitments with respect to what should be focused, have been made. The summary of these given below will provide a clear focus for user interface research within C&C:
  1. The vision of a new generation of artifacts for computing and communication that will be naturally embedded in the everyday environment.
  2. The choice of lifelong learning integrated in normal work situations as the primary application area.
  3. A primary scenario of use characterized by collaborative activities distributed in time and space and with dynamic groups of user with very heterogeneous background.
  4. The importance of long term and full scale field studies under realistic conditions.
  5. The importance of an interdisciplinary competence in the projects
  6. The need not only for exploratory research but for efficient methods and tools, stressing the importance for the «situatedness« of the methods.
  7. The systematic combination of the ¬society of agents¬, the «network« and the «spatial« metaphors for interface design.
  8. The focusing of design of appropriate multimodal interfaces.
  9. The need for filtering functions in the userinterfaces to cope with the information overflow.
  10. The need for adaptive techniques to handle the problem of heterogeneous user groups and context dependent activities.
  11. The need for powerful help and explanatory functions for the same reasons.
  12. The need for powerful representations of the domain, the discourse and the users in order to enable reasoning power that can facilitate the realisation of both filtering, adaptivity and help functions.
  13. The combinations of user interface competence expressed functinally in terms of the bahaviour of ¬personal assistants¬.
  14. The ulilization of underlying computational agent architectures for convenient implementations of such personal assistants.
  15. The need for enhancement of mediatechnology such as videoconferencing facilities, interactive walls, high resolution video and graphics as well as eye glass terminals to make progress towards the initial technological vision.

LIST OF CONCRETE PROJECT PROPOSALS

The contributing research groups are : SICS/UACM, SICS/NLP, SICS/DCE, SICS/KTM, DSV/K2LAB, DSV/SYSLAB, the IT/Mediagroup and the SISU/Interaction and Communication group. On the last row for each of the projects a set of numbers in parentheses indicates to which of the foci above that the projects contributes. Most of the projects below are potentially good projects for cooperation among several of the groups. Some of the projects are more clearly tailored to the competence of a specific group.

Knowledge Extraction in Hypermedia edited by Annika Waern, SICS/UACM and Carl Gustaf Jansson DSV/K2LAB. (a very similar proposal has been edited by Ake Malmberg, DSV/K2LAB) (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)

Computer Supported Business Intelligence edited by Peter Rosengren, SISU/Interaction and Communication

Interface Metaphores for Agent Architectures edited by Annika Waern, SICS/UACM and Lennart Fahlen SICS/DCE.

A Speech and Gesture-based Interface to a Personal Assistant edited by Mats Wiren, SICS/NLP.

Interface Design for Mobile, Heterogeneous Computing Equipment edited by Mats Wiren, SICS/NLP.

User modeling as a means for creating explanations, help and tutoring edited by Kristina Hook, DSV/UACM and Carl Gustaf Jansson, DSV/K2LAB

Constructing and evaluating a prototype for managing distributed education edited by Peter Idestam Almquist and Harald Kjellin, DSV/K2LAB.

Design of CSCW systems edited by Kia Hook, SICS/NLP and Jacob Palme, DSV/K2LAB.

Cooperative Design edited by Mats Ericsson, IT/Media group

Creating an Environment for Decentralised, Co-Operative Work in Information Systems Engineering. edited by Janis Bubenko and Bengt Wangler, DSV/SYSLAB and SISU

Advanced Media Technology Research edited by Mats Ericsson, IT/Media group