As mentioned above, for each word, w, in a collection of documents, an inverted file stores a list of documents that contain the word, w. In addition, search engines typically order the list for each word by a ranking score. Modern search engines must be able to answer more than just single-word queries, however. Describe an efficient method for computing a list of the documents that contain two words, w and u, ordered by the ranking scores assigned to the documents. What is the running time of your method in terms of n_w and n_u, the respective sizes of the lists for w and u?