Tailoring Google Site Search Brett Lucas Payman Labbaf July 2008
Problem Relying on a separate cataloguing mechanism. Unsatisfactory results prioritisation Unable to index document other than html Too resource intensive to maintain
Why Google CSE Ease of use Well-known Search platform Highly customisable
Introduction to Google Side Search Introduction to Google Site Search Google CSE plus –Ads & Google branding free –XML API – support
Features Customisability –Via a simple web interface –Via submitting an XML file –using XML API XML API Synonyms Data Biasing –based on the age of the documents –based on the domain from which the document is fetched –based on the path of the document Refinements
Tailoring Google Site Search Sign in to your Google account –Provide basic details –Provide all domains which define the scope of the search Create labels for refining or restricting the search Create the search pipeline –Wrap up the request and submit it to Google –Receive the response as an XML –Parse the XML and extract the results –Display the search results Provide final fine-tuning for Data Biasing
Potential Risks Availability Functional limitations Google spec changes? No absolute accuracy guaranteed Contingency Old system on stand-by Loosely coupled modules Documenting spec-related code
Future Tasks Perform Usability Test Improve visual aspects of external navigation Extend the refinements
Resources Introduction to Google Site Search Google Site Search Homepage Custom Search Engine XML Spec Google Site Search XML API