Regarding The National Academies Discovery Engine ... Because of the vast breadth of material -- published books, public documents, testimony to Congress, proceedings from meetings, and more -- simple, linear search results have been demonstrated to be inadequate to the task of providing useful, relevant results to our researchers and browsers. To address those and other institutional issues, we have developed what we call a "discovery engine," which extracts pertinent information from every search result, and integrates new "knowledge discovery" tools into a unified interface. Discovery tools are different than simple "identification." They include user-driven exploration, enabling "successive approximations" to facilitate "homing in" on what the user wants; include relevant pointers to specific rich resources; include key term extraction and reapplication, and more. These are the sorts of tools for discovery needed within a deeply diverse resource set like The National Academies sites. As you'll see, initial search results are presented with tools to allow further discovery and further refinement. The user has the ability to initiate "find more like" actions from any presented document, or extract key terms from a search and reuse them, or apply the "instant searches" we provide (extracted from the results). It sounds more complicated to use than it truly is--please give it a try. This is the first "production release" -- it's the result of three cycles of usability testing and process improvements -- and will continue to be improved. If you've got feedback, I'd love to hear it: Michael Jensen, mjensen @ nas . edu