Apologies for the unfortunate overuse of the same letter `q’ for denoting both the set Q of features contained in the query and the probability qi of the presence of a feature in irrelevant documents, but there is no intended, direct connection between these two quantities.