Users of search tools who seek educational materials on the Internet are typically presented with either a web-scale search (e.g., Google or Yahoo) or a specialized, site-specific tool. The specialized search tools often rely upon custom data fields, such as user-entered ratings, to provide additional value. As currently designed, these systems are generally too labor intensive to manage and scale up beyond a single site or set of resources. However, custom (or structured) data of some form is necessary if search outcomes for educational materials are to be improved. For example, design criteria and evaluative metrics are crucial attributes for educational resources, and these currently require human labeling and verification. Thus, one challenge is to design a search tool that capitalizes on available structured data (also called metadata) but is not crippled if the data are missing. This information should be amenable to repurposing by anyone, which means that it must be archived in a manner that can be discovered and leveraged easily. In this paper, we describe the extent to which DiscoverEd, a prototype developed by ccLearn, meets the design challenge of a scalable, enhanced search platform for educational resources. We then explore some of the key challenges regarding enhanced search for topic-specific Internet resources generally. We conclude by illustrating some possible future developments and third-party enhancements to the DiscoverEd prototype.