| [AD] Floating Point Precision |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
Some floating point comparisons are made by miscelleanous routines : they
involve the use of a small number (usually EPSILON) to get rid of round-off
issues. So far, each function uses its own EPSILON no matter whether it has
already been defined anywhere else. (Some functions don't use an EPSILON at
all and make a brutal comparison x == y between two floating point numbers
which is, from a numerical point of view, the worst thing to do). May be we
should define a global constant AL_EPSILON in order to keep floating point
code consistent through the library ? May be we should allow the user to
modify its value ?
Bertrand.