Résumé
Starting from an example of reverse engineering problem in food science (what is the more relevant packaging for a given fresh food?), we will present complementary solutions based on knowledge engineering methods and food engineering models. A special focus will be given on termino-ontological resources (called TOR) which permit to define standardize vocabularies and knowledge associated with a given domain. We will show than TOR can be used to: (i) gather scientific data about packaging extracted from heterogeneous data sources in a data warehouse, (ii) permit flexible multi-criteria querying of the packaging data warehouse to retrieve the more relevant packaging for a given fresh food.