The development of our technologies is proceeding at multiple levels in parallel, developing the underlying technological capabilities as we explore different application domains. These application domains are not intended to span all the uses of the workspace technology, but rather, to explore diverse points in the space of possibilities, while leveraging the interests and projects found among our colleagues at Stanford.
For the last two years we have been working with Prof. Martin Fischer of Center for Integrated Facilities Engineering (CIFE) to develop visualization tools to support collaborative planning of large facilities and construction projects. Today all these planning activities take place in conference rooms with architects and subcontractors surrounded by paper documents stacks of blueprints, schedules, resource histograms, organization charts, task lists, and reports on the wall. These paper documents present alternatives and evaluations of the current plan by experts. Such activity is representative of mission and logistical planning, and is an excellent driving application for our room environment. In a seed project funded by CIFE, we are helping this group develop 2D and 3D visualizations based on the visual metaphors and information contained in traditional paper documents.
Working with Prof. Mendel Rosenblum of the Computer Systems Laboratory, we have developed tools for visualizing computer systems and networks. For example, T. Munzner has developed a technique for visualizing the topology of the MBONE, and software for visualizing large graphs such as program call graphs, web link structure, and large file systems. D. Tang, working with M. Baker, has also developed software for visualizing all the Metricom network traffic in the S.F. Bay Area. We have also developed software for visualizing large log-files containing millions of events. Such files can be generated from program traces, computer system simulators (such as the SimOS system), operating system services, etc.
The RiboWeb project [Chen et al., 1997] is based at the Stanford Section on Medical Informatics, and has the goal of creating computational technologies to support scientific collaboration, particularly in the context of computational modeling of biological structures. The ribosome is an organelle within the cell where the genetic code of DNA is translated into the protein molecules that carry out most important functions within the cell. Its structure is the focus of research internationally, and there are many opportunities for sharing data, in order to improve the quality of current models. The RiboWeb system has three parts: (1) a knowledge base of the objects and relationships within ribosomal biology, including a structured representation of the contents of 100 key published articles about the ribosome, (2) a suite of computational modules that do useful things with the knowledge base (e.g., guide literature retrieval, compare two articles for consistency, build a new model based on contents of subset of articles, look for discrepancies between data sets), and (3) a session manager/user interface generator that acts as the mediator between the KB and computational modules and the user.
The work on RiboWeb has lead to the realization that the visualizations most appropriate as interfaces to systems for biologists must be based on familiar biological graphics, or (more generally) "domain graphics." Part of the work in Dr. Altman's laboratory (e.g., [Felciano et al., 1997]) has focused on methods for bringing familiar graphics into the RiboWeb system as interfaces. As part of this project, we propose to use the RiboWeb project and the biologists and computer scientists working on it as a testbed for the software we develop for the Information Mural. In particular, there are basic graphics from ribosomal science (the "RNA secondary structure" graphics discussed in the reference), that act as a focal point for group interactions. These include mentor/student (2-way meetings) meetings, research group (N-way meetings) and also meetings between biology research groups (2-way meetings between groups).
The Stanford Learning Lab has initiated a series of design experiments of new kinds of learning spaces that integrate technology and conventional learning. These spaces attempt to integrate site and long-distance learning, hands-on and virtual experimentation, and group discussion with formal instruction. Several versions of such spaces have been implemented featuring flexible reformatting of space, furniture, and representation formats. A working hypothesis of the SLL is that there is a need for systems that allow for precise control of images, including overlaying, distorting, comparing, marking and labeling dynamically, while being discussed by a class to create an effective group learning environment. These functionalities coupled with the ability to interact with visualizations of simulated experiments, systems or processes would allow the creation of virtual teaching laboratories. In the long run, a collection of Information Murals augmented with appropriate software could become standard equipment in the future university classroom and laboratory.
In our prototype application we are interested in seeing what design possibilities (and requirements) are suggested by working with areas such as theater. From a cognitive viewpoint, subject matter such as theatrical performance is difficult to understand because the event is multilayered and swiftly passing in time. In the theater, performance consists of multiple layers of expressive modes, such as movement, light and color, style and gesture, language, psychological rhythms, which interconnect in constantly changing and unpredictable ways. Study of performance involves noting and analyzing each of the elements of the ensemble to comprehend their interactions as well as to interrogate the basis for their presence: in other words students must ask what choices on part of the performers are represented. Why does performer pause here, or speak the line with this inflection, or alter the pace or emotional intensity of the action? Describing and decoding the event demands ability to separate out layers, to disentangle the constituent parts of this ensemble of events, labeling and charting them, and then playing with variations and choices
With the Information Mural, we will be able to view multiple performances simultaneously, or highlight individual elements, such as lighting or stage movement throughout single performances and between different versions. It will also be important to view the text of the screenplay, and to coordinate that with explanatory material that documents or amplifies the background information needed to understand the basis of choices.