Mark Stefik

WYSIWIS revised: early experiences with multiuser interfaces

WYSIWIS (What You See Is What I See) is a foundational abstraction for multiuser interfaces that expresses many of the characteristics of a chalkboard in face-to-face meetings. In its strictest interpretation, it means that everyone can also see the same written information and also see where anyone else is pointing. In our attempts to build software support for collaboration in meetings, we have discovered that WYSIWIS is crucial, yet too inflexible when strictly enforced. This paper is about the design issues and choices that arose in our first generation of meeting tools based on WYSIWIS. Several examples of multiuser interfaces that start from this abstraction are presented. These tools illustrate that there are inherent conflicts between the needs of a group and the needs of individuals, since user interfaces compete for the same display space and meeting time. To help minimize the effect of these conflicts, constraints were relaxed along four key dimensions of WYSIWIS: display space, time of display, subgroup population, and congruence of view. Meeting tools must be designed to support the changing needs of information sharing during process transitions, as subgroups are formed and dissolved, as individuals shift their focus of activity, and as the group shifts from multiple parallel activities to a single focused activity and back again.

The cost structure of sensemaking

Making sense of a body of data is a common activity in any kind of analysis. Sensemaking is the process of searching for a representation and encoding data in that representation to answer task-specific questions. Different operations during sensemaking require different cognitive and external resources. Representations are chosen and changed to reduce the cost of operations in an information processing task. The power of these representational shifts is generally under-appreciated as is the relation between sensemaking and information retrieval. We analyze sensemaking tasks and develop a model of the cost structure of sensemaking. We discuss implications for the integrated design of user interfaces, representational tools, and information retrieval systems.

Expert systems: perils and promise

Based on a review of some actual expert-system projects, guidelines are proposed for choosing appropriate applications and managing the development process.

The organization of expert systems, a tutorial☆

Abstract This is a tutorial about the organization of expert problem-solving programs. We begin with a restricted class of problems that admits a very simple organization. To make this organization feasible it is required that the input data be static and reliable and that the solution space be small enough to search exhaustively. These assumptions are then relaxed, one at a time, in case study of ten more sophisticated organizational prescriptions. The first cases give techniques for dealing with unreliable data and time-varying data. Other cases show techniques for creating and reasoning with abstract solution spaces and using multiple lines of reasoning. The prescriptions are compared for their coverage and illustrated by examples from recent expert systems.

WYSIWIS revised: early experiences with multi-user interfaces

WYSIWIS (What You See Is What I See) is a foundational abstraction for multi-user interfaces that expresses many of the characteristics of a chalkboard in face-to-face meetings. In its strictest interpretation, it means that everyone can see the same written information and also where anyone else is pointing. We present several examples of multi-user interfaces that start from the WYSIWIS abstraction. In our attempts to build software support for collaboration in meetings, we have discovered that WYSIWIS is at once crucial and too inflexible in its strictest sense. WYSIWIS must be relaxed for all our software tools to better accommodate important interactions in meetings. Relaxations to WYSIWIS are characterized in terms of constraints on its four key dimensions: display space, time of display, subgroup population, and congruence of view.

Integrating Access-Oriented Programming into a Multiparadigm Environment

The Loops knowledge programming system integrates function-oriented, system object-oriented, rule-oriented, and—something notfound in most other systems—access-oriented programming.

City lights: contextual views in minimal space

City Lights are space-efficient fisheye techniques that provide contextual views along the borders of windows and subwindows that describe unseen objects in all directions. We present a family of techniques that use a range of graphical dimensions to depict varied information about unseen objects. City Lights can be used alone or in conjunction with scrollbars, 2D overview+detail, and interaction techniques such as zoomable user interfaces.

Cognoter: theory and practice of a colab-orative tool

Cognoter is a program helps a cooperating group of people to organizing their thoughts for a presentation, e.g., a paper or talk. It is designed for use in the Colab, an experimental laboratory created at Xerox PARC to study computer support of cooperative real-time group problem-solving. Cognoter provides a multi-user interface and a structured meeting process. An annotated graph of ideas is built up by the group in three stages: brainstorming for idea generation, ordering for idea organization, and evaluation for choosing what will be finally be presented. Interesting aspects of Cognoter include direct spatial manipulation of ideas and their order relationships, support of parallel activity, and incremental progress toward a total ordering of ideas.

Perspectives on artificial intelligence programming

Programs are judged not only by whether they faithfully carry out the intended processing but also by whether they are understandable and easily changed. Programming systems for artificial intelligence applications use specialized languages, environments, and knowledge-based tools to reduce the complexity of the programming task. Language styles based on procedures, objects, logic, rules, and constraints reflect different models for organizing programs and facilitate program evolution and understandability. To make programming easier, multiple styles can be integrated as sublanguages in a programming environment. Programming environments provide tools that analyze programs and create informative displays of their structure. Programs can be modified by direct interaction with these displays. These tools and languages are helping computer scientists to regain a sense of control over systems that have become increasingly complex.

Strategic computing at DARPA: overview and assessment

Strategic Computing, a 10-year initiative to build faster and more intelligent systems, is ambitious, flawed by overscheduling perhaps and problems of definition, but basically sound.