Hello, -> The PL2 scheme is ready and can be merged. I've cut down on the number of calculations and have optimized the upper bound as well. ->Even though I've eliminated multiple calculations and logarithms from BB2, I could not optimize the upper bounds in BB2. -> However, I've corrected the code and have introduced checks to ensure thet sumpart < max_part and upper bound does not become "NAN". It was becoming NAN as the number inside the logarithm for stirling value was sometimes becoming negative because of the values the document statistics had. I now check for this first, and take necessary steps to ensure a positive value (at the cost of a greatly increased upper bound :( ). Th code and documentation to show what I am talking about can be found here:- https://github.com/aarshkshah1992/xapian/commit/cd7dcf856a82fb1a87309aa2facbbec100eb2961 -> Both PL2 and BB2 can be merged now. *-> *I am now working on BO1 query expansion. Will also work on the feedback for BB2 and PL2 when I get it so as to get them merged before mid-term evaluation ends. Regards Aarsh -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.xapian.org/pipermail/xapian-devel/attachments/20140623/b35a0a53/attachment.html>