|
|
Graphics, Visualization & Usability Center,
College Of Computing, Georgia Tech, Atlanta, GA 30332
iansmith@cc.gatech.edu
Recently, the research community has begun investigating the problems of how to provide technological support for informal communication and awareness. Systems such as the PARC Media Space [1] and CaveCat [2] provide their users with a virtual " space " in which they can conduct their activities. In these systems the virtual space consists of an audio/video network which allows the participants to see and hear each other. This network allows many of the informal communication processes that occur normally in physical space to be moved into virtual space. This breakdown of the barriers of physical space is important because the communication process does not require the participants to be co-located, they may be distributed in physical space.
Let us consider an example problem in the area of awareness across physical space. Alice is an administrative assistant for Betty, who telecommutes from her home. Alice needs to judge Betty's level of " busyness " to determine whether or not to forward incoming phone calls to Betty. For a tool to help Alice solve her problem of assessing whether or not to allow callers access to Betty it must meet the following requirements:
One can imagine a software tool that allows Alice to perform this task. Such a tool would give her some type of display on her computer- perhaps a graph of some sort- of the amount of activity in Betty's office. The " activity " could be measured by tracking mouse movements, counting keyboard clicks, by taking digital video images and processing them or combinations of these or other techniques. Such a tool might also provide some type of " ambient " audio information to Alice [3] to help keep her aware of the level of activity in Betty's home-office. This software must also provide Betty with an interface to allow her to control how and when Alice (or others) may access the information the tool is gathering and, thus, insure Betty's privacy when she desires it [4].
The example just presented is a specific problem for Alice and Betty; I feel that there are many problems which are caused by geographic separation and the inability to informally communicate which can remedied with software tools like the one outlined above. As another example of this type of problem, consider the informal information which allows people to decide when to make a transition into a more formal mode of communication. This type of transition can be difficult to make when participants are not physically co-located. A tool which solved this problem would allow users to have enough information to determine when it would be appropriate to initiate a two-way video conference, for example.
There are several significant technical challenges that must be addressed. First and foremost is the question of what reusable components, which in this framework are called operators, should a toolkit present so that an application writer or HCI researcher can quickly develop their ideas into a testable prototype? Second, over what domain should these operators function? For example, in the example with Alice above, the tool she used might operate over the domains of keyboard entry, mouse activity, and digital video. Finally, what tools should be built in an effort to allow the application writer or HCI researcher to interact with the toolkit and framework in a natural way and minimize (or eliminate) the amount of code that must be written to achieve their goals?
Let us revisit the example of Betty and Alice again, now considering it through the lens of the three questions just mentioned. First, I have already established some of the domains of interest here: computer input devices and digital video. For this toolkit to succeed it clearly must present some reasonable interface to these types of media, not to mention other related media such as the state digital audio inputs and other sensors. Further, all of these domains reside on the input side of the application: gathering the information. The toolkit must also address output domains such as digital audio, digital video, and the graphical user interface so the user can see the results of the computations.
In the previous example, the activity tool measures several quantities such as the amount of keyboard and mouse activity and the amount of change in a video scene. Abstractions must be present to compute all of these values; and these will be the operators for this tool. An operator which will be necessary in this example, but which is not as obvious as the others, is some type of security operator which operates over any domain. This operator should allow Betty to control when and by whom her information can be accessed. Finally, to be able to construct such an application for Alice a tool will be needed which will allow the application developer to easily " fit together " these operators and their domains into a result which Alice can use effectively.
This framework is based on the idea of connecting a source domain to a sink domain. This connection is accomplished by connecting operators that perform the desired computations. (This type of notation itself is not new and has been used successfully elsewhere.) Figure 1 is a simplified diagrammatic representation of the parts of my architecture working together to create a visualization like the one Alice uses to stay aware of Betty's activities. The two domains of the keyboard (input) and the Graphical User Interface (output) are represented by triangles and the operators are represented by the squared off ovals. The black rectangles on the sides of the operators and the GUI domain indicate that these objects can accept an input. The arrows represent the flow of data from one part of the architecture to another.
As mentioned before, one of the goals of this work is to allow application builders and researchers to build prototype applications quickly using this architecture. Therefore, I am developing a tool that will allow users to construct prototypes in a visual representation similar to the one used above in Figure 1. It seems reasonable that within the scope of informal communication and awareness a tool could be built which would allow new and interesting applications to be built without requiring the implementation of a general tool for constructing arbitrary programs. Other disciplines have had success with tools which allow for the visual construction of programs in a domain, such as Iris Explorer and DEC's AVS in the area of Scientific Visualization.
In conclusion, the ability of coworkers to communicate informally and stay aware of each others activities is important to the success of their work. When coworkers are separated by space, their communications break down due to the inaccesability of informal information. This work is exploring the commonalities between the problems that result from this spatial separation. These commonalities are being used to develop an architecture and toolkit for the construction of applications which solve the communication problems.
[2] Mantei M., Baecker R., Sellen A., Buxton W., Milligan T., Experiences In The Use Of A Media Space. In Proceedings Of Conference On Computer Human Interaction (CHI) `91. p. 203-208.
[3] Smith I., Hudson S. Low Disturbance Audio For Awareness And Privacy In Media Space Applications. To appear in ACM Conference On Multimedia `95.
[4] Smith I., Hudson S., Mynatt E., Selbie R. Applying Cryptographic Techniques To Problems In Media Space Security. In Proceedings Of Conference On Organizational Computing Systems (COOCS) `95.