Dermatologists are clever classifiers. We mastera repertoire of approximately 2000 different diagnoses,which can be classified and structured.Classification of diseases is very important. Byclassifying biological phenomena, we structureour understanding of the underlying biologicalcomplexity, and thereby make it possible for usnot only to ask meaningful positivist scientificquestions, but also offer help to our patients.Two threats exist in all classification systems,splitting and lumping, and both are equally serious.Splitting occurs when the same diseaseentity is split into numerous diagnoses dependingon, for example, location, a good examplebeing pityriasis amientacea and scalp psoriasis.This overwhelms the reader with diagnoses thatare not essentially different, but that have beenclassified as different because of an essentiallyrandom aspect of the disease, e.g., location orclinical appearance. This does not allow a meaningfuluse of existing knowledge by direct transfer,and therefore erodes the understanding ofthe underlying pathogenic process as well as accumulationof clinically relevant knowledge.The other extreme is lumping, where all diseasesare lumped together pell-mell in large categories,where little consideration is given tosignificant etiological, pathogenic, and clinicaldifferences between disease entities. It may bespeculated that several of the more commondermatological diseases fall into this trap, asclinical experience suggests significant inter-individualdifferences in regard to treatment responseor prognosis.Nosology should be based on defined parameters,e.g., anatomy (gross and microscopic), etiology,pathogenesis or therapy. It should provideclinically meaningful distinctions between diseaseentities in order not only to promote theunderstanding of the disease and accumulationof knowledge, but also to help practical management.