|
Boost : |
From: Andrei Tovtchigretchko (andrey_at_[hidden])
Date: 2000-04-29 00:20:06
This approach breaks "Strict Weak Ordering" STL requirement,
that is, you can have a,b,c such that (a == b) && (b == c) && ! (a == c).
So, algorithms such as 'sort' cannot be used.
'Equality Comparable' model is also broken and, since algorithms that
deal with floating point typicaly expect it to hold and assume transitivity
axiom, you can get:
a == (a + delta) == (a + 2 * delta) == ... == Infinity => a == Infinity
etc.
In other words, such objects cannot be safely passed to algorithms
expecting floating point numbers. The safe way is to use full-blown interval
arithmetics.
Andrei
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk