In biomedicine, biodiversity and other fields of research, large databases are used. Assuming that a proper statistical procedure has been chosen, a crucial point is the selection of the right software to compute the data. The available software has to be sufficiently proven and having the guarantee that it is reliable. Currently, it is easy to obtain free software for most statistical procedures. We agree that a free software is especially useful because as a large number of researchers can take benefit of it. However, in several repositories, software has not been sufficiently proven, and could yield to erroneous results. This situation could lead to dreadful consequences, for instance, when studying cancer or complex genetic diseases. We propose that researchers should be especially accurate in their software selection, and also the control levels should be improved in order to upload new software in a public repository.