DocEng 2011: Interoperable Metadata Semantics
2The 11th ACM Symposium on Document Engineering
Mountain View, California, USA
September 19-22, 2011
Interoperable Metadata Semantics with Meta-Metadata: A Use Case Integrating Search Engines
Yin Qu, Andruid Kerne, Andrew M. Webb, Aaron Herstein
Presented by Yin Qu.
ABSTRACT
A use case involving integrating results from search engines illustrates how the meta-metadata language facilitates interoperable metadata semantics. Formal semantics can be hard to obtain directly. For example, search engines may only present results through web pages; even if they do provide web services, they don't provide them according to a mutually interoperable standard. We show how to use the open source meta-metadata language to define a common base class for search results, and how to extend the base class to create polymorphic variants that include engine-specific fields. We develop wrappers to extract data from HTML search results from engines including Google, Bing, Delicious, and Slashdot. We write a short meta-search program for integrating the search results, reranking them, and providing formatted HTML output. This provides an extensible formal and functional semantics for search. Meta-metadata also directly enables representing the same integrated search results as XML or JSON. This research can profoundly transform the derivation and representation of interoperable metadata semantics from a multitude of heterogeneous wild web sources.