The discussion centers on the best software for fitting data distributions, with an emphasis on transparency and source code accessibility. Some participants argue that software with hidden processes is less useful for scientific purposes, while others highlight the practicality of using established packages like MATLAB, SPSS, and others, even without full source code. The importance of the software choice is linked to the scrutiny the results will face, suggesting that custom solutions may be best for critical applications. Chi-square goodness of fit tests are recommended for unsupported distributions, and GNU/Octave is mentioned as a suitable option for many users. Ultimately, the choice of software should align with the specific needs of the analysis and the expectations of reproducibility in scientific research.