Wednesday, July 06, 2011

Federated & Faceted Search

In the last few months CLEBC has shipped a new search engine. The upgrades have been significant and these improvements include;
  • performance and currency has been greatly improved - search results occur quickly and the indexes are updated within 12 hrs of new content being added to any information silo.
  • navigation made more intuitive - navigational facets provide standard approaches that are currently being used for search and faceted search.
  • alignment of taxonomies across all information silos - the top level taxonomy has been reconciled across all information silos for better browsing of results across silos.
  • search applied across all silos including the CLEBC store and CanLii - search results include references across all silos with the inclusion of the CLEBC store (to include the coming learning events) and CanLii to include broader results.
  • ability to both search and discover (with the ability to narrow the search criteria within both of these approaches) - the use of faceted search is powerful when it provides the ability to both directly inquire and discover information.

All these new search features are due to using both a federated and faceted search approach. The rest of this post describes these two concepts as they relate to search and why they are important to legal search.

Federated Search
Federated search spans multiple data sources, where the search indexes include results from these many different, and related, information sources. CLEBC already has many high quality content sources and many of these sources are related through practice areas, taxonomies and other common and legal attributes. The design process of this federated search included analysis of each information source and reconciled the common attributes to provide a unified search result across all sources. The result of this federation is to have results from all information silos appears as one result set.

The effectiveness of federated search lies in how single search terms can span multiple information sources and how the results can be organized by date, relevance and information source. Finding related and relevant results across information silos covers more contexts a provides greater insight from different perspectives.

Faceted Search
Faceted search is best understood in a comparison of discovery based searching vs. direct searching. This is the difference between "I'd like to find a beach resort in the British Virgin Islands" (discovery) and "I'd like the contact information for the Long Bay resort in the British Virgin Islands" (direct). With direct searching you know exactly what you want you just need further details; with discovery searching you don't really know what you want and like to browse to find it. A search facet provides this discovery. As  show with the image on the right there are two facets, one for Content Type and the other for Practice Area. As you search for a general (or more specific term) the contents of the facet changes also with the number of results within each term to the right of the faceted list. This allows discovery where you can drill down on terms until you find specifically what you want... discovery based searching.

Special Thanks
Many great projects happen in the open and with many peoples contribution. In this project I would like the thank Susan Munro, Laura Selby and all of CLEBC who provided guidance and insight in how legal search could be improved. I'd also like to thank LexUM for their expertise in building faceted and federated search.

This search patterns book from O'Reilly is an essential read from understanding search. One of the authors, Peter Morville, has put together an excellent collection of flickr images showing the different search patterns.