[math-fun] Can anyone fix DLMF 15.8.23 ?
http://dlmf.nist.gov/15.8 It seems to need another term on the right. --rwg
* Bill Gosper <billgosper@gmail.com> [Aug 05. 2015 07:43]:
http://dlmf.nist.gov/15.8 It seems to need another term on the right. --rwg [...]
Works for me (fixing parameters because of Pari's limitations with symbolic computations): -------------------------------------------- \\ http://dlmf.nist.gov/15.8 \\ eq. 15.8.23 /* \[\mathop{F\/}\nolimits\!\left({a,1-a\atop c};z\right)=\left(\sqrt{1-z^{-1}}-1% \right)^{1-a}\left(\sqrt{1-z^{-1}}+1\right)^{a-2c+1}\left(1-z^{-1}\right)^{c-1% }\mathop{F\/}\nolimits\!\left({c-a,c-\tfrac{1}{2}\atop 2c-1};\frac{4\sqrt{1-z^% {-1}}}{\left(\sqrt{1-z^{-1}}+1\right)^{2}}\right),\] */ \r hypergeom default(echo,1); N=15; z = 1/2/'z^2+O('z^N); c=1 a=1/2 L=hypergeom([a,1-a], [c], z, N) S=sqrt(1-1/z) R = (S-1)^(1-a) * (S+1)^(a-2*c+1) * (1-1/z)^(c-1) R *= hypergeom([c-a, c-1/2], [2*c-1], 4*S/(S+1)^2 , N) L - R quit; -------------------------------------------- The expression 'L - R' near end gives: 93990019574025/147573952589676412928*z^-30 + 25145962430625/18446744073709551616*z^-28 + 1690195005625/576460752303423488*z^-26 + 457028729521/72057594037927936*z^-24 + 7775536041/562949953421312*z^-22 + 2133423721/70368744177664*z^-20 + 147744025/2199023255552*z^-18 + 41409225/274877906944*z^-16 + 184041/536870912*z^-14 + O(z^-13)
Hypergeometric2F1[1 - a, a, c, z] == (-1 + Sqrt[1 - z^(-1)])^(1 - a)* (1 + Sqrt[1 - z^(-1)])^(1 + a - 2*c)*(1 - z^(-1))^(-1 + c)* Hypergeometric2F1[-1/2 + c, -a + c, -1 + 2*c, (4*Sqrt[1 - z^(-1)])/ (1 + Sqrt[1 - z^(-1)])^2] it = % /. {a -> 3/5, c -> 1/3}; Table[{z, Equal @@ (Chop@N[List @@ it] )}, {z, -1/2, 1/2, 1/24}] gives all 'True' except an indeterminate for z = 0 Wouter. -----Original Message----- From: Bill Gosper Sent: Wednesday, August 05, 2015 3:31 AM To: math-fun@mailman.xmission.com Subject: [math-fun] Can anyone fix DLMF 15.8.23 ? http://dlmf.nist.gov/15.8 It seems to need another term on the right. --rwg _______________________________________________ math-fun mailing list math-fun@mailman.xmission.com https://mailman.xmission.com/cgi-bin/mailman/listinfo/math-fun
DAWK! My bad! The additional term was there, but scrolled off the bottom of my screen! Apologies for the braindamage. --rwg Maybe there should be some sort of typographical "formula hyphen". This could happen at the bottom of the page in an actual book. 18th Century texts duplicated the first line of the next page at the bottom of the current page. Nowadays, we could do that with a miniaturized display. In browsers, too. On Tue, Aug 4, 2015 at 6:31 PM, Bill Gosper <billgosper@gmail.com> wrote:
http://dlmf.nist.gov/15.8 It seems to need another term on the right. --rwg
HOLY CR@P! Not only did I misread 15.8.24, I also mistyped it as 15.8.23!! DLMF must think I'm a raving crank! --rwg On Wed, Aug 5, 2015 at 6:25 AM, Bill Gosper <billgosper@gmail.com> wrote:
DAWK! My bad! The additional term was there, but scrolled off the bottom of my screen! Apologies for the braindamage. --rwg
Maybe there should be some sort of typographical "formula hyphen". This could happen at the bottom of the page in an actual book. 18th Century texts duplicated the first line of the next page at the bottom of the current page. Nowadays, we could do that with a miniaturized display. In browsers, too.
On Tue, Aug 4, 2015 at 6:31 PM, Bill Gosper <billgosper@gmail.com> wrote:
http://dlmf.nist.gov/15.8 It seems to need another term on the right. --rwg
participants (3)
-
Bill Gosper -
Joerg Arndt -
Wouter Meeussen