Semantic Web technologies have been chosen to structure and publish data of the Henri Poincaré correspondence corpus and to provide a set of tools for exploiting it. Two major issues have arisen during this work. The first one is related to the manual editing of triples in order to populate an RDF database. It is a tedious task for users with an important risk of error. That justifies the development of a dedicated tool to assist them during this process. It uses and combines different methods which are presented in this article. The second issue is related to the exploration of the corpus database. The need of an approximate and explained search has emerged. An engine based on the application of SPARQL query transformation rules has been designed. Different research tracks are considered to pursue this work which intends to apply in other contexts than the Henri Poincaré correspondence corpus.