23 Jan
2010
23 Jan
'10
9:40 p.m.
Thomas Klimpel wrote:
Find attached a small test program that tries to highlight some of the "usual" floating point traps.
Is it time to raise the idea of a Boost fixed-point library again? It's not a solution for all, or even most, floating-point problems, but there are cases where floats are used only because they are more convenient (i.e. better supported by the language) than fixed-point. Phil.