A Semantic Approach for Learning Objects Repositories with Knowledge Reuse Isabel Azevedo 1 , Rui Seiça 2 , Adela Ortiz 1 , Eurico Carrapatoso 2 , Carlos Vaz de Carvalho 1 , 1 Instituto Superior de Engenharia do Porto, GILT, R. de São Tomé, 4200 Porto, Portugal {iazevedo, adela, cvc}@dei.isep.ipp.pt 2 Faculdade de Engenharia da Universidade do Porto, R. Roberto Frias, Porto, Portugal ruiseica@gmail.com, emc@fe.up.pt Abstract. In this paper we discuss how the inclusion of semantic functionalities in a Learning Objects Repository allows a better characterization of the learning materials enclosed and improves their retrieval through the adoption of some query expansion strategies. Thus, we started to regard the use of ontologies to automatically suggest additional concepts when users are filling some metadata fields and add new terms to the ones initially provided when users specify the keywords with interest in a query. Dealing with different domain areas and having considered impractical the development of many different ontologies, we adopted some strategies for reusing ontologies in order to have the knowledge necessary in our institutional repository. In this paper we make a review of the area of knowledge reuse and discuss our approach. Keywords: Semantic Web, ontology selection, ontology reuse. 1 Introduction Formal ontologies have been seen as a way to state content specific understandings for many knowledge-sharing tasks, in the form of conceptualizations of the world of interest. In a study from 2007 [1], the participants were asked about what motivated them to use ontologies. The answers showed that the two main reasons were the need to share a common understanding of the structure of information among people or software agents (69.9 percent), and the requirement to enable the reuse of domain knowledge (56.3 percent). However, the creation of ontologies leads to some costs, not only time, and it is often not considered a simple process, even if many methodologies have emerged. The creation of ontologies cannot be done automatically and requires, to a