The authors study the equivalence between two common normalisations used in
the derivation of shot-noise in optical digital receivers. The role played
by the bit rate in both approaches is discussed. They conclude that both a
pproaches are formally equivalent, but one is conceptually preferrable. Fur
thermore, no absolute result can be given on the influence of bit rate on t
he receiver shot-noise performance. Any result is necessarily relative to t
he normalisation used.