On the need for changing the ENDF/B convention for the representation of cross-sections in the unresolved resonance region of fertile and fissile nuclei