The archive file fairdata.zip contains the files:
| readme.txt | this file of general directions |
| fair.f90 | the complete source code |
| The sample cases provided are | |
| case1.nml | a sample case for y vs. x (non-parametric) |
| case1.out | output from case1.nml |
| case1a.nml | same as case1, but with added joints |
| case2.nml | a sample case for closed parametric data - cross section |
| case2a.nml | same as case2, but with added joints |
| case3.nml | a sample case that turns back (parametric) |
| case3a.nml | same as case3, but with added joints |
| case4.nml | a sample case - drag polar of an airfoil |
The program asks for the name of the input file. The program makes the file fair.out to be viewed or printed and three plotting files. The file fair.dat has the original data points. The file fair.crv has a fairly dense set of points on the spline. The file fair.jnt has the joints. Using gnuplot, you can visualize the fit with the command
gnuplot> plot 'fair.dat','fair.crv' with lines,'fair.jnt'
Anothe file called fair.dbg contains a track of the calculation. You might find the info in this file useful.
The namelist NAM1 has the following variables
| xData,yData | arrays of input data |
| weights | weight of each data point (preset to 1.0) |
| joints | array of joints |
| parametric | .TRUE. if data is to be plotted against a parametric variable |
| .FALSE. if the data is to be treated as y vs. x | |
| closed | .TRUE. if the data closes (last point==first point) |
| .FALSE. |
Note - parametric and closed are preset to FALSE.
I have not really exercised this code extensively. I do see however, that it can go unstable if you try to put too many joints into the fit. There should be at least a few data points between each pair of joints.