Systemic sclerosis (SSc) is a rare connective tissue disorder whose aetiology remains obscure, although environmental and genetic influences are likely to play a role. Disease registries have contributed to enhancing our understanding of this debilitating illness, but without sensitive, specific, and extensively validated classification criteria, accurate comparison between registries and the identification of patients suitable for clinical trials can be problematic. The American College of Rheumatology (ACR) criteria, published in 1980, have become outdated as our understanding of disease specific autoantibodies and nailfold capillaroscopy has improved. In addition, the sensitivity of the ACR criteria is low with respect to limited SSc. Although subsequent classification systems have been proposed, none has gained universal approval. The two- versus three-subset disease model remains a point of debate. Newly derived criteria are likely to draw upon the older classification systems as well as incorporating up-to-date diagnostic techniques and biomarkers. Validation will be critical before their use becomes widespread.