h_b_k, didn't see your reply. However, now there's even a link to the query. ;)
4:15 pm on May 15, 2003 (gmt 0)
Ok my first guess was a fair way off. I'm reading things well beyond my 1:45am brain :P
4:17 pm on May 15, 2003 (gmt 0)
The model creates a space in which query is represented by m-dimensional vector, where m is the number of unique terms in the query and the length of m-th component of the vector is associated with the term frequency in the query. Then, a vector similarity function, such as the inner product, is used to compute the relevance between the query and a document in Goo index.
7:08 pm on May 15, 2003 (gmt 0)
Thank you for posting such an interesting question. I have now even more brain food for this evening.