Version: 1.5rc1 (using KDE KDE 3.5.1) Installed from: Ubuntu Packages Hi, Whilst looking at this report comparing accuracy in Gnumeric and Excel: http://www.csdassn.org/software_reports/gnumeric.pdf , I decided to try some of the tests in KSpread. Sometimes KSpread calculates the correct result, but shows it incorrectly if the number format is set to "Number" or "Scientific" but correctly if set to "Text" format. eg: "=POISSON(0;200;0)" should give "1.38e-87" but displays 0 in KSpread with number or scientific format, but gives a correct result in text format. My guess is that the threshold for rounding results up/down to 0 is a little too high?
I can confirm in svn trunk
SVN commit 884606 by mecir: Do not display small values as 0. Use scientific format if abs(value) < 1e-10. BUG: 124527 M +8 -5 ValueFormatter.cpp WebSVN link: http://websvn.kde.org/?view=rev&revision=884606