The educational resource search engine finds new wings in this article focused on OER disocovery. The problem with existing approaches, say the authors, is that "these methodologies depend on human annotated metadata for approximating the usefulness of a resource for a particular need." Even adding simple tags or keyswords can be a problem. "The OERScout text mining algorithm is designed to 'read' text based OER documents and 'learn' which academic domain(s) and sub-domain(s) they belong to." Results are sorted using a keyword matrix and desirability (based on license and search engine ranking fo the resource).