Mark Ruzon wrote:

Suppose you want to compute 3 - x for a vector of floats:

using namespace std; vector<float> v; transform(v.begin(), v.end(), v.begin(), boost::bind(minus<float>(), _1, 3.0f));

This compiles fine on the MSVC 8.0 compiler. But I want to compute x - 3, and replacing _1 with _2 in the above expression yields a mess of errors, starting with:

c:\program files\boost\boost_1_34_1\boost\bind.hpp(278) : error C2679: binary '[' : no operator found which takes a right-hand operand of type 'overloaded-function' (or there is no acceptable conversion) c:\program files\boost\boost_1_34_1\boost\bind.hpp(205): could be 'float boost::_bi::list1<A1>::operator [](boost::arg<I>) const' with [ A1=float &, I=1 ]

Anybody have any thoughts on this one?

transform( v.begin(), v.end(), v.begin(), _1 - 3 ); transform( v.begin(), v.end(), v.begin(), 3 - _1 ); _1 represents the first argument to the final lambda (block, functor, whatever you want to call it). _2 the second. You cannot have a second argument without also having a first. Assuming that minus() takes left and right as arguments one and two then your example should already be x - 3. Whichever expression you really want, the way to do it is to swap the arguments to minus. You want one of these two: boost::bind(minus<float>(), _1, 3.0f) boost::bind(minus<float>(), 3.0f, _1) The result of this is a unary lambda whose single argument is represented as _1 in the expression. K