A query engine processes user queries and returns matching answers, in an order determined by a ranking algorithm. The algorithm gives a score representing the importance of the answer with respect to the query. This score is a combination of query-independant and query-dependant criterias. Query-independant judge the document regardless of the query; exampels being the length, vocabulary, publication date, and connectivity-based data such as number of links pointing to the page (called in-degree). Query-depentant criteria include a cosine measure for similarity in the vector space model (well known from traditional IR techniques) and all connectivity-basede techniques.