Résumé
The abundance and the heterogeneity of medical knowledge to be represented and organized lead to consider automatic techniques and tools to minimize tedious, repetitive, time-consuming, costly tasks. Such works are often connected to ontological approaches that leads to design medical knowledge organization systems (MKOSs) in laboratory methods and conditions and to pay less attention to the human and socio-organizational aspects linked to the use of MKOSs in concrete situations. These methods seem to be reductive, and the use of automatic tools should be therefore connected to epistemological approaches in which designers search for criteria that are external to the corpus. However, such elements are not provided in advance. The aim of our paper is to assess how and where to find them to perform the right action while using automatic techniques for the elaboration of MKOSs.