An infinite-precision floating-point function library for Miranda and type for Haskell
Newcastle, 14 Nov 2007
Version 4.1 fixes the ubn_add bug mentioned below. It only took me two years to get round to releasing it...
Newcastle, 15 Nov 2005
There is a bug in the implementation of the overly complex ubn_add2 function, which was only detected by the (forthcoming) faster arcsin implementation giving wrong results from the 80th or 99th digit onwards, but it may affect other calculations. An immediate fix to the source of all versions is to replace the line ubn_add = ubn_add2 with ubn_add = ubn_add1.
Newcastle, 6 Nov 2005
bignum/BigFloat now has a Home Page and Download Site on SourceForge.
Nijmegen, 5 Oct 2005
BigFloat participated in the "manydigits" competition and came last for speed. Some of the winning entries obtained inaccurate answers very fast.
Sittingbourne, 26 Feb 2004
Bignum's sqrt(2) calculation has exceeded the existing publicly-available value (1,250,000 figures). It managed 1930026 decimal places having run from 20 Jan - 26 Feb 2004. Then some idiot managed to turn the power off in the house.