OFFSET
-1,2
COMMENTS
R. W. Gosper asks if the coefficients are well-defined. Until this is answered, the sequence is only conjectural. This sequence is very close to A178451, but presumably different from it.
REFERENCES
R. W. Gosper, Posting to the Math Fun Mailing List, Dec 21 2010
LINKS
EXAMPLE
e^(Pi*sqrt(163)) = s^3 + 744 - 196884/s^3 + 167975456/s^6 - 180592706130/s^9 + 217940004309743/s^12 - 19517553165954887/s^15 + 74085136650518742/s^18 - ... where s = 640320. Now set t = 1/s^3.
PROG
(bc)
/* GNU bc code, computes a(0) through a(7) */
define trunc(x) { auto sc, t; sc=scale; scale=0; t=x/1; scale=sc; return(t) }
scale = 200; pi = 4 * a(1); r = e(pi * sqrt(163)); s = 640320;
c0 = 1 + trunc(r - s^3);
c1 = -1 - trunc(((s^3 + c0) - r) * s^3);
c2 = 1 + trunc((r - (s^3 + c0 + c1/s^3)) * s^6);
c3 = -1 - trunc(((s^3 + c0 + c1/s^3 + c2/s^6) - r) * s^9);
c4 = 1 + trunc((r - (s^3 + c0 + c1/s^3 + c2/s^6 + c3/s^9)) * s^12);
c5 = -1 - trunc(((s^3 + c0 + c1/s^3 + c2/s^6 + c3/s^9 + c4/s^12) - r) * s^15);
c6 = 1 + trunc((r - (s^3 + c0 + c1/s^3 + c2/s^6 + c3/s^9 + c4/s^12 + c5/s^15)) * s^18);
c7 = -1 - trunc(((s^3 + c0 + c1/s^3 + c2/s^6 + c3/s^9 + c4/s^12 + c5/s^15 + c6/s^18) - r) * s^21);
CROSSREFS
KEYWORD
sign,more
AUTHOR
N. J. A. Sloane, Dec 22 2010, based on a posting by R. W. Gosper to the Sequence Fans Mailing List, Dec 21 2010
EXTENSIONS
Cullen link, bc code, and a(8) from Robert Munafo, Dec 23 2010
STATUS
approved