Editing Fitted Curves

Data points define a fitted curve through a process of interpolation, that is to say that the curve passes through all of the data points.

Often data points have been derived from an external source, such as an existing design. In such cases there is usually importance placed upon the accuracy of curves with respect to these points. However, there is also the requirement that curves are fair (judged principally by the inspection of curvature tuft plots). Usually, there is some conflict between these two requirements, meaning that curves fitted to their original data points are not adequately fair.

Various factors may contribute to this situation, for example:

• Stochastic errors (noise) in the original data

• Systematic errors in the original data (for example, from a measuring process)

• Inability of the B-Spline representation to recreate the original curve geometry

• Unfairness of the original curve

An acceptable compromise is required between fairness and an adequate level of fidelity with the original data. This is sought by editing curves via their data points.

The application provides comprehensive means of manipulating fitted curves. Data points can be easily moved, removed and added. Additionally, tangency conditions at data points can be specified.

Data points are represented visually by small symbols, the specific appearance depends on the tangency conditions that are specified at the point:

Figure 3:47. The different data point types.

The 'handles' that emanate from some of the data point types provide control over tangency. for further information about different data point types, see Curves Graphical Reference.

Data point attributes are also represented numerically in the Data Points grid of the Data Bar:

Figure 3:48. Data Points grid of the Data Bar.