The Henri Poincaré correspondence is a corpus composed of around 2100 letters which is a rich source of information for historians of science. Semantic Web technologies provide a way to structure and publish data related to this kind of corpus. However, Semantic Web data editing is a process which often requires human intervention and may seem tedious for the user. This article introduces RDFWebEditor4Humanities, an editor which aims at facilitating annotation of documents. This tool uses case-based reasoning (cbr) to provide suggestions for the user which are related to the current document annotation process. These suggestions are found and ranked by considering the annotation context related to the resource currently being edited and by looking for similar resources already annotated in the database. Several methods and combinations of methods are presented here, as well as the evaluation associated with each of them.