Ontologies have been proven invaluable tools both for the semantic web and for personal information management. In the context of a historical archive an ontology may provide meaningful and efficient support for search tasks as well as be used as a tool for storage and presentation of historical data. The creation however of such an ontology is complex, since the digitized archive documents are not in text format and the concepts that must be captured may vary among different time periods. This work presents a user-centric methodological approach for extracting the ontology of an historical archive focusing on the evaluation issues related to this process. The approach is exemplified through cases from its application in the University of Athens Historical Archive.