Summarizing microarray normalized probe expression to gene expression

The usual and fundumental result of the analysis of a microarray data experiment is a list of differentially expressed genes, always according to the statistical design and the biological questions asked. However, what we see most of the times in such a list is a set of differentially expressed probes and not a list of differentially expressed genes (according to official HUGO nomenclature for example). Although this final list contains also gene names next to probes, I have been asked a lot of times to provide a list of genes and discard the probes as they are rather confusing the bench biologists. The latter are asking questions of the type "which probe should I believe" or "why the expression in one probe is so much different from the expression of the other probe corresponding to the same gene"?

The answer is not always easy,