TimeWalker

tools for visualising huge amounts of log data

The goal of this project is to implement a tool for visualising huge amounts of log data.

The visualisation techniques of TimeWalker are based on the paradigm of information visualisation. In this paradigm, the emphasis is on maximizing information density and minimizing end-user effort by always showing detail and context simultaneously. TimeWalker introduces the innovative concept of an hierarchical zoom lens sliding along the time-axis. Because the user-interface is exploiting the typical human capabilities for pattern-recognition, TimeWalker can be seen as a sophisticated looking-glass for visual data-mining in the time-dimension.

In order to optimize usability, there will be close cooperation from the beginning with alpha- and beta-sites. Network and system maintenance is chosen as the first application-domain. Feedback will be based on sites that are representative for producing high volumes of log data within distributed infrastructures.

TimeWalker is also quite interesting as an open-source project exploitingPythonup to its limits as a wide-spectrum language in any stage in the software lifecycle. Python has enabled a high level of elegance and portability to represent implementation, documentation, persistent storage of data and running states, regression tests and profiling. A remarkable scalability and performance could be obtained by just using the generic facilities of Numerical Python and wxPython, instead of coding specfic extensions in C.