> I did a test of CSPLINE interpolation from the GSL package and compared the > test (interp.c) file given in the doc/examples/ directory of the native GSL > package > to the XMGRACE interpolation function and found that they differ > significantly. > here i attach a pdf file which shows this difference.
What does XMGRACE use for end conditions on the cspline? I notice from your PDF that the discrepancies seem to "originate" in the derivatives at the ends of the interval. GSL's end conditions are documented here: http://www.gnu.org/software/gsl/manual/html_node/Interpolation-Types.html#Interpolation-Types - Rhys
