Version: HEAD (using KDE KDE 3.5.1) Installed from: Compiled From Sources OS: Linux When fitting data it would be useful to optionally allow an additional x-vector to be supplied by the user, which would be used to calculate the output y-vector. At present the input x-vector is used to calculate the output y-vector. However, there are many situations when the user might want the output y-vector to be sampled differently from the input x-vector.
This request seems to be in conflict with the current behaviour of returning residuals. If the output x-vector does not overlap (or only partially overlaps) with the input x-vector how would residuals be calculated?
Sounds like the residual curve follows the current behavior. Only the fit itself would be evaluated on the new x vector.
I think my preference here would be to add an additional x-vector. The result and residual vectors would be provided basesd on the additional x-vector. In the case where the additional x-vector is the same as the base x-vector then the existing behaviour is returned. This would be the default behaviour.
In order to maintain backward compatibility the only way to implement this proposed change would be to duplicate all the existing fit plugins, modify their name slightly, and add the additional input x-vector and output y-vector. This would likely be confusing to the user and the proposed functionality can also be created by using an expression.