This is qEKHAD, one of the Maple packages accompanying the forthcoming book "A=B" (soon to be published by A.K.Peters, Ltd.) by Marko Petkovsek, Herb Wilf, and Doron Zeilberger. Also try out the other packages: EKHAD and multiEKHAD The latest versions are always available by anon. ftp to ftp.math.temple.edu in the directory: /pub/zeilberg/programs or, on the Web, at http://www.math.temple.edu/~zeilberg For general help, and a list of the available functions, type "ezra();". For specific help type "ezra(procedure_name);" Warning: q is a global variable qfac(k) means (1-q)*(1-q^2)*...*(1-q^k), and NOT (1)(1+q)*...*(1+q+...+q^(k-1)) Version of April 1996 We wish to thank Tewodros Amdeberhan for suggesting improvements The CF in 2-3 satisfies a 3-term recurrence A PROOF OF THE NAME RECURRENCE by Shalosh B. Ekhad, Temple University, ekhad@math.temple.edu I will give a short proof of the following result( REF ). Theorem:Let b(n,k) be (1/2 k (k + 1)) q qfac(n + k) ---------------------------- 2 qfac(k) qfac(n - k) and let SUM(n) be the sum of b(n,k) with respect to k . SUM(n) satisfies the following linear recurrence equation n n n 2 4 n 2 a2 (q q - 1) (1 + q q) ((q ) q + q q + 2) SUM(n) n - ------------------------------------------------------ + a2 (- 2 - 2 q - q q n 2 2 n 2 n 2 2 n (q ) q (q q - 1) ((q ) q + 2 + q q) n 2 3 n 3 n 2 4 n 4 6 n 2 5 n 3 6 - 3 (q ) q - q q - 3 (q ) q + (q ) q - (q ) q + (q ) q n 4 7 n 3 5 n 2 2 n 3 4 n 5 9 n 5 8 + (q ) q + (q ) q - (q ) q + (q ) q + 2 (q ) q + 2 (q ) q 7 n 5 n 6 9 n 6 10 n 7 11 + 2 q (q ) + (q ) q + (q ) q + (q ) q ) SUM(n + 1) / 3 n 2 n 2 n 2 2 n / (q (q ) (q q - 1) ((q ) q + 2 + q q)) / n 2 a2 (1 + q q ) SUM(n + 2) - ------------------------- n 2 3 (q ) q =0. PROOF: We cleverly construct G(n,k) := n 3 5 n 2 4 n 2 3 n 2 n k n 5 8 ((q ) q - (q ) q + (q ) q - q q + 2 q q - 2 + q (q ) q k n 4 7 k n 3 6 k n 4 6 k n 3 4 k n 3 + q (q ) q + 2 q (q ) q + q (q ) q + 2 q (q ) q - q q q k n 2 k n k k k 2 5 n 3 k 2 4 n 2 + q q q - q q q - 3 q q - 3 q + (q ) q (q ) + (q ) q (q ) k 2 3 n k 2 n 2 3 k 2 2 n k 2 + 2 (q ) q q - (q ) (q ) q - (q ) q q - 2 (q ) q) a2 k n (1/2 k (k + 1)) (- 1 + q q q) q qfac(n + k) / k n n 2 n 2 2 n 2 / ((- q + q q) (q q - 1) ((q ) q + 2 + q q) qfac(k) qfac(n - k)) / with the motive that n n n 2 4 n 2 a2 (q q - 1) (1 + q q) ((q ) q + q q + 2) b(n, k) - ------------------------------------------------------- + a2 (- 2 - 2 q n 2 2 n 2 n 2 2 n (q ) q (q q - 1) ((q ) q + 2 + q q) n n 2 3 n 3 n 2 4 n 4 6 n 2 5 n 3 6 - q q - 3 (q ) q - q q - 3 (q ) q + (q ) q - (q ) q + (q ) q n 4 7 n 3 5 n 2 2 n 3 4 n 5 9 n 5 8 + (q ) q + (q ) q - (q ) q + (q ) q + 2 (q ) q + 2 (q ) q 7 n 5 n 6 9 n 6 10 n 7 11 + 2 q (q ) + (q ) q + (q ) q + (q ) q ) b(n + 1, k) / 3 n 2 n 2 n 2 2 n / (q (q ) (q q - 1) ((q ) q + 2 + q q)) / n 2 a2 (1 + q q ) b(n + 2, k) - -------------------------- n 2 3 (q ) q = G(n,k)-G(n,k-1) (check!) and the theorem follows upon summing with respect to k . This took 8.650 Seconds of CPU time Then, L and B in 2-5 can be found after multiplying the last equation throughout by n 2 3 n 2 n 2 2 n (q ) q (q q - 1) ((q ) q + 2 + q q)/a2.