Hi Markus, On Wednesday, 2011-07-20 12:03:08 +0200, Markus Mohrhard wrote:
this patch removes the check for the upper bound of the second parameter of TINV. I checked the algorithm of TINV but there seems to be no reason to limit this number. The only problem is that for values greater than 1E14 the rounding error dramatically increases but in my opinion everyone who uses such large numbers should know about problems related to rounding and floating point arithmetic.
People tend to use and trust such functions anyway.. Does Excel impose some upper bound on that parameter? If so, we should compare its output near such value with Calc's, and if sufficiently "equal" maybe raise the bound. Another approach would be to compare with Mathematica, Regina is known to love such work ;-) Eike -- PGP/OpenPGP/GnuPG encrypted mail preferred in all private communication. Key ID: 0x293C05FD - 997A 4C60 CE41 0149 0DB3 9E96 2F1A D073 293C 05FD
Attachment:
pgpTKuiNlsUx0.pgp
Description: PGP signature