Question about the interpolation method used file2array() in pair_eam.cpp

Dear LAMMPS authors,

I was reading the file2array() function in pair_eam.cpp. I found that piecewise 3rd order Lagrange interpolation was used to transform different potential table to a unified array:

cof1 = -sixthp(p-1.0)(p-2.0);
cof2 = 0.5
cof3 = -0.5p(p+1.0)(p-2.0);
cof4 = sixth

These four coefficients found in file2array() are clearly Lagrange interpolation. I understand that this was used in DYNAMO code. But wouldn’t it be better if cubic spline interpolation was used here?

Thank you very much for your work! I learned a lot from the code.