Currently we have a demonstration search that works by telling AlltheWeb to limit results to pages that link to Creative Commons licenses. While useful, this is far from our vision of a metadata-aware search engine.
The first requirement for a Creative Commons license-aware search engine is that license metadata (RDF embedded in pages) must be indexed. It wouldn't be necessary to index arbitrary RDF initially -- indexing only Creative Commons license metadata would be a good first step along the path to a Semantic Web-enabled search engine. Please read the detail page for further information Add license search to a major commercial search engine.
Once you start indexing license metadata, you can do two obvious things with it:
Provide users with an interface to filter their results by license or license characteristc. The aforementioned AlltheWeb demonstration interface is an example of the latter.
Display license information in search results. This could be done even if a query does not involve a license filter. If you have license information for a result, display the license in proximity to the result.
As you index and understand more metadata, you'll be able to go beyond these basics, with enhanced format or domain-specific searches and richly annotated results.
We've started a benevolent cycle by making it painless and natural for people to publish quality metadata (automatically included in HTML generated for cut&paste publishing by our license selection application). Continue the cycle by enhancing search with metadata -- providing stimulus for users and other applications to generate yet more and better metadata.