next up previous contents
Next: Introduction to SGML Up: CASCADE Previous: Functional Overview of

Data Streams in CASCADE

  As indicated in design philosophy, data streams and data stream manipulation are important features of CASCADE. While the functions described in this section are not intend for end users in there current form, the data being examined here is exemplar of the kinds of data streams that will be used to drive CASCADE process and agents. A brief overview is provided to give the reader a sense of where CASCADE will be going in the future. There are three kinds of data being collected and analyzed in CASCADE at the current time, user activity, gross actions, and micro activity data. They are discussed below.

User Activity

Figure gif shows an example of a how user activity is currently visualized. It is available to users of CASCADE as the private eye function, although its primary purpose is to aid researcher in understanding connection patters for the purpose of developing agents that might have to analyze this data stream. In its intermediate form, the analysis shows the number and duration of CASCADE sessions for any group or subgroup of users over any number of days and over any period during the day. A histogram of current connection time at the bottom of the display shows possible good times for setting up synchronous meetings. The data might also be used to better understand which users are active on the project and who is inactive, etc.

Figure: User Connection Visualization

Gross Actions

CASCADE maintains a log of activity for the documents that it manages. See section gif for a description of the report generation component of this feature. Upon login the user is informed of the amount of activity since his/her last session. This activity could include people adding new documents or comments about existing documents. The user can choose to review a detailed list of this activity in order to find new documents/comments that the user is interested in. Some of the activities that CASCADE logs are comment creation, document creation, editing, and link deletion.

Micro Activity

In a collaborative environment it is beneficial to know if people are really looking at the information in the document. It is easy to find out how long a document was open on a desktop, but it is much more difficult to know if people really read it. In an attempt to answer these kinds of questions, CASCADE keeps track of user navigation throughout a document. CASCADE has the capability to record Mouse Clicks, Scrolling, keypresses, and other user motions. This information is currently kept in a file named ``mymainlog.dat'' in the user's personal directory. In future implementations, this information could be used to make statistical inferences in order to determine how actively a document is being looked at. As patterns are detected in these movements, it may even be possible to tell when an inexperienced user needs assistance. An example of the visualization of such micro activity is shown in Figure gif.

Figure: Micro Activity Visualization

next up previous contents
Next: Introduction to SGML Up: CASCADE Previous: Functional Overview of

Michael Spring
Fri Jan 31 13:59:00 EST 1997