Keep cool. The program is work in progress and there will be always bugs. Write a brief bug report and sent it to the author of the program. If possible sent the .all file too.
Check the number format in Windows. Especially the decimal point and use this format for importing data. The program is well tested by using the “.” as decimal point. Most of the other files, like optical constants files are working only with “.” as decimal point.
The best option is to begin with Parratt to make the structural fits and then use Zak- or the matrix-formalism to refine the model.
The 5 main errors in descending order are.
The kinematical approximation is another very easy way to calculate reflectivity spectra. Actually it uses following approximations.
The main advantage is that one obtains an easy analytical formula in which the reflectivity is expressed as the Fourier transformation of the z-derivative of the electron density. In principle the reflectivity curve can be inverted to get the density profile. (see e.g. K.-M. Zimmermann, W. Press, Phase determination of x-ray reflection coefficients, Physical Review B 62, 15 (2000)). The major drawback is the usable energy range in which the absorption is very small. e. g. off-resonant hard X-rays. In soft X-rays the absorbing part of the refractive index could be too large. If there is demand to implement this inversion algorithm don't hesitate to ask.
The second beam is used in the two calculation modes dichroic and asymmetry. It just means that in one calculation the ray has polarization 1 and in the other one polarization 2. In dichroic mode this means that the two separate reflectivities are calculated and plotted. The asymmetry mode does the same but in this case the weighted differences of these two curves are plotted. The matrix-formalisms are faster as they depend on the incoming polarization only in the last step of the calculation. In the Parratt-Formalism the whole algorithm have to run two times.