I'm submitting this in case someone easily spots a less verbose representation of fabs( x - y )< eps than the one shown below and would like to share it. Using lambda (or variants) only (I know I can write my own functor class). Thanks. #include <cmath> #include <boost/function.hpp> #include <boost/lambda/bind.hpp> #include <boost/lambda/lambda.hpp> #include <boost/numeric/conversion/bounds.hpp> typedef boost::function<double(double)> f1_; typedef boost::function<double(double, double)> f2_; typedef boost::function<bool(double)> g_; double eps = boost::numeric::bounds<double>::smallest(); f2_ f2 = lambda::bind( f1_( fabs ), lambda::_1 - lambda::_2 ) ; g_ g = ( eps > lambda::_1 ); BOOST_ASSERT( lambda::bind( g, lambda::bind( f2, lambda::_1, lambda::_2 ) )( 0.0, 0 * eps ) ); BOOST_ASSERT( !lambda::bind( g, lambda::bind( f2, lambda::_1, lambda::_2 ) )( 0.0, eps ) );