Forms of Implementation

It is always difficult to describe dynamic methods with static communication media like the paper this text is written on. I hope this limitation can be overcome to some extent by looking at the implementation details of the prototype application TUPF presented in Appendix A. The real effectiveness can only be seen and estimated by looking at full implementations on a computer system.

In the following I shall present some implementation topics for TDFs.

Without going into much detail I think it is useful to discuss some implementation issues at this point. Greater detail about the prototype implementation can be found in Appendix A.

As stated in the introduction fast methods are required to access the large data volumes commonly occurring in wildlife applications. Building a user interface (UI) for this purpose means that graphical methods are needed and used. Future UIs may be based on virtual reality applications or further developments, but for the time being they are not available to most researchers.

For the basic form of the TDFs two parameters need to be interactively changed: the position and the width. In a computer program two scrollbars can be used for this purpose, one that selects the position and one which defines the width (figure 4.6).

Figure 4.6: Implementation example of scrollbars as one method to define the parameters for a TDF.

A more sophisticated solution would be the application of a diagram with two axes. The x-axis describes the position, the y-axis the frame width. The user can then use a pointer (e.g. a computer mouse) moving around in this space and simultaneously change both parameters as needed.

For cyclic temporal aspects the form of using two scrollbars can also be used, one for the selection of the time point, the other to determine the width of the time frame. In addition there can be cases where more sophisticated access methods are needed. One example for this is illustrated in figure 4.7. This way the problem of cyclic and circular parameters does not occur as in the application of scrollbars.

Figure 4.7: Implementation example of a graphical selection method for cyclic temporal aspects.

These two examples of implementation issues have shed some light on how TDFs can be used on a high level of computer abstraction in the field of user interface design. It is not the aim of this study to produce user interface design research, as there exists a large literature about this field itself. This look ahead to Appendix A is meant to give an impression of how TDFs can be used in a more concrete situation.

When multiple time aspects are considered simultaneously logical operators as and, or, not need to be defined and implemented to describe the relationship of the TDFs. Figure 4.8 illustrates a possible setup for analyzing data with the combination of three temporal data frames handling several temporal aspects. In this example three aspects are under investigation: The time as a linear component (selection of a two month period around 9.April 1998), the altitude of the sun (dawn and dusk times) and moon illumination (full moon phase).

Figure 4.8: Multiple time aspects as an example in the application of TDFs. Data is selected within a certain time span (green), sun altitude (yellow) and lunar illumination (white).

Further developments of TDFs (e.g. window splitting, inverse selection functions, data shifting) are provided in the prototype application in Appendix A. It is easier to understand and introduce them by providing an example implementation in parallel. I will now provide an example application of the temporal data frames in the next section.