The development of modern shoulder replacement surgery started over half a century ago with the pioneering work done by CS Neer. Several designs for shoulder prostheses are now available, allowing surgeons to select the best design for each situation. When the rotator cuff is intact, unconstrained prostheses produce reliable and reproducible results, with prosthesis survival rates of 97% after 10 years and 84% after 20 years. In patients with three- or four-part fractures of the proximal humerus, the outcome of shoulder arthroplasty depends largely on healing of the greater tuberosity, which is therefore a major treatment objective. Factors crucial to greater tuberosity union include selection of the optimal prosthesis design, flawless fixation of the tuberosities, and appropriate postoperative immobilization. The reverse shoulder prosthesis developed by Grammont has been recognized since 1991 as a valid option for patients with glenohumeral osteoarthritis. Ten-year prosthesis survival rates are 91% overall (including trauma and revisions) and 94% for glenohumeral osteoarthritis with head migration. These good results are generating interest in the reverse shoulder prosthesis as a treatment option in situations where unconstrained prostheses are unsatisfactory (primary glenohumeral osteoarthritis with marked glenoid cavity erosion; comminuted fractures in patients older than 75 years; post-traumatic osteoarthritis with severe tuberosity malunion or nonunion; massive irreparable rotator cuff tears with pseudoparalysis; failed rotator cuff repair; and proximal humerus tumor requiring resection of the rotator cuff insertions).