Femoral stem modularity in hip replacement was first developed to connect a ceramic head to the stem, then extended to metal heads using the Morse taper principle. Is it a good thing, or a necessary evil? It contributes to improving lower limb length and lateralization setting, at the cost of fairly rare complications such as dissociation and fretting corrosion, which can exceptionally lead to ARMD (Adverse Reaction to Metal Debris). Modular necks were later recommended, with a double Morse taper: cylindrical for the head junction, and more or less flattened for the stem. Is this one modularity too far? Dual modularity in theory perfectly reproduces the biomechanical parameters of the hip, but is unfortunately associated with fractures and severe corrosion, leading to ARMD and pseudotumor, especially in Cr-Co necks. Moreover, it provides no functional advantage, and no longer has a role outside dysplasia and other femoral deformities. Metaphyseal-diaphyseal modularity is not widespread in primary implants, and is it really necessary? Only one model has been widely studied: S–Rom™ (Depuy®). It features a metaphyseal sleeve adapting to the anatomy of the proximal femur, with a stem fitted via an inverse Morse taper. Its only interest is in case of congenital dislocation; like all metal connections, it incurs a risk of fracture and corrosion. On the other hand, modularity is widely employed in revision implants. Does it really help these procedures? The connection between a proximal femoral component of variable geometry and a diaphyseal stem with press-fit distal fixation provides a real solution to problems of length, lateralization and anteversion. Early models encountered high rates of fracture, but current implants and rigorous surgical technique have reduced this risk. Corrosion is a less serious problem, as the Morse taper undergoes only axial stress, without the friction undergone by other models subject to varus stress.