Abstract
The FRIEND21 Project, which ended on March 31, 1994, is briefly described and
its final results are presented. The FRIEND21 Project conducted research into
human interfaces having two-tiered architecture consisting of "metaware" and
"agency model." This approach attempts to deal with human-interface
issues from the very top layer of a symbolic and cognitive
environment constructed between the user and the machine to the
bottom layer of machine functions organized into the agency model.
Keywords:
Symbolic environment, symbolic context, two-tiered HI architecture, metaware,
agency model.
Introduction
The Future Personalized Information Environment Development (known as
the "FRIEND21" Project and referred to below as "F21"), a Japanese
government project aimed at conducting research into human interfaces
for the 21st century, came to a close on March 31, 1994. With the aim
of making it possible for everyone in the 21st century to use
computers in their daily lives, F21 Project research focused on the
development of basic software technologies that, five years after the
completion of the project, would be commercially available. The
Computer Human Interaction (CHI) community has already been introduced
to the F21 Project in the CHI'91 report [1] and CHI'92 video [2].
Moreover, the results of the F21 Project have been collected in the
form of HI software design guidelines in a book published in October
1994. Work on the English translation is now progressing rapidly and
is scheduled to be completed in early 1995.
RESULTS: CONCEPTS
[c1]HI design
A symbolic environment, or in other words, the
information environment, is created between computers and users.
Human interface design is based on the idea that the information
environment should be dynamically constructed according to the user's
everyday changing activities.
[c2] Symbolic context---HI context
For the information environment, it is essential that computer
functions be given a symbolic and modal expression on the monitor
screen, and be provided in a combined context of space and time,
and/or in a symbolic context.
[c3] Changing HI architecture
Each individual user formulates his/her own interpretations and
understanding of specific contexts based on his/her own experiences.
One issue that HI design must deal with is how to make the interfaces
themselves change in response to different user interpretations, and
by so doing, assist users using computers to execute their tasks
appropriately.
[c4] HI configurations
HI design architecture takes on a two-tiered form that is derived from
its symbolic nature. One is "metaware," which is cognitive
architecture capable of building a symbolic human environment. The
other is the "agency model," the HI's specific structural and
mechanical software architecture. Based on the preceding concepts,
the project get out to give a concrete form to metaware and the
agency model, i.e., to the cognitive architecture and the mechanical
architecture.
RESULTS: METAWARE
[m0] "Metaware" refers to interface design techniques that employ appropriate
representations (i.e. metaphorical figures) to dynamically drive tasks using
specific cognitive principles.
[m1] Depending on the user and the context, the relationship between
representations and tasks is polysemous and/or synonymous, the system
must therefore be able to create polysemy and/or synonymity
appropriate to the context.
[m2] There must be agreement on the general syntactical manner in
which diverse functions and information are to be arranged and
sequenced, as well as on how they are to be displayed.
[m3] The third strategy for metaware is the use of context dependence.
1. Combining and presenting the most appropriate relationships between
representations and tasks according to context is called
"context-sensitive selection."
2. In contrast, appropriately
positioning these figures spatially or in a
time sequence is referred to as "context-dependent display."
[m4] The degree to which the displayed interface metaphor is effective
for the user depends upon the understanding of the interface
metaphor. The following three metaware methodologies are used as
context-dependent selection techniques that will lead to a dynamic
understanding of the metaphors on the part of users.
1. Presentation of similarity: High similarity must be provided to the task domain intended by the user.
2. Arrangement of polysemy: One symbol must ordinarily play different roles
within multiple contexts. In other words, a symbol should facilitate the
execution of a different task in a different environment.
3. Arrangement of synonymity: Metaphors appropriate to users' task
goals should naturally be context-dependent. In other words, when
given a task, the preparation of an environment in which appropriate
(synonymous) metaphors can be selected in accordance with each
context is equivalent to arranging synonymity.
RESULTS: AGENCY MODEL
[a0] The agency model executes the metaware and
is a form of architecture that aims to achieve an integrated
operational environment that will allow independent modules, referred
to as "agents," to communicate with each other via one
or more "studios."
[a1] The system consists of three types of agents: function agents,
which achieve tasks, presentation agents, which express
representations, and system agents, which provide an integrated
environment in conjunction with the other two types of agents.
[a2] The presentation and function agents exist autonomously and can
both be distributed independently. Expendability and other field
upgrades are made possible by this distribution.
[a3] Each of the first two agents has its own native protocol,
referred to as presentation and function protocols. Conversion of
these protocols via studio(s) enables information exchanges.
[a4] System agents also include I/O managers, which directly handle
the user interface, a dialog manager, which supports dialog between a
user and the system, and the user information manager, which
administers various kinds of information about a user.
[a5] The user information manager has a role to provide database
profiles of comparatively long-term activities of a user. User's
short-term characteristics i.e. operational histories, are
administered by studio(s). These are used for the context dependent
operations [3].
DISCUSSION
This national project was unique in its being a software project
linking the information society, the trend toward
internationalization, and different industries (i.e. the convergence
of media industries). However, the project's final implementation was
targeted almost solely at human interfaces, and as a result,
internationalization became a dominant factor for the duration of the
project.
Acknowledgments
We would first of all like to extend our appreciation
to everyone who participated in this project. We would also like to
thank the following people who contributed much of their time to
summarize the results of the project in document form: T. C.
Watanabe, M. Hirose (Fuji Xerox Co. Ltd.), M. Uyama (Fujitsu
Laboratories Ltd.), N. Asahi (Mitsubishi Electric Corporation), and
Y. Takashima (NEC Corp.).Special thanks are also extended to K.
Tamura (Electrotechnical Laboratory ), M. Ishikawa (Kyushu Institute
of Technology) and H. Miyauchi (IWANAMI SHOTEN PUBLISHERS), who
provided the writers of the FRIEND21 book with invaluable
suggestions, assistance and encouragement.
References
- [1] Nonogaki, H. Ueda, H.: FRIEND21 Project: A Construction of 21st
Century Human Interface, ACM, CHI'91 Proceedings, pp. 407-414, 1991.
- [2] Ueda, H: FRIEND21 (Aiming for the Personalized Information
Environment), ACM SIGGRAPH VIDEO REVIEW, ISSUE 79, CHI'92 Special
Video Program, No.10, 1992.
- [3] Uyama, M.: A Blackboard-based Architecture for Filtering New
Software Features, ACM COOCS'93 Proceedings, pp. 210-215, 1993.
