Hi there again, a bit more information. MSVC10 is complaining about that it cannot find the correct copy algorithm implementation since for some reason the iterator categories are messed up in its opinion. I doubt that. Here is why: Copy takes an input and output interator type. My input type is: bit_aligned_pixel_iterator< const pixel_t >. The output is: bit_aligned_pixel_iterator< pixel_t >. Obviously the iterator category for both types is the same and when asked for it results into: boost::detail::iterator_category_with_traversal<std::input_iterator_tag,boost::random_access_traversal_tag>. Now when calling copy the compiler eventually has to decide what copy implementation to use based on the iterator categories. For reasons I don't understand right now it cannot choose between: _OutIt std::_Copy_impl<_InIt,_OutIt>(_InIt,_InIt,_OutIt,std::input_iterator_tag,std::output_iterator_tag) _OutIt std::_Copy_impl<_InIt,_OutIt>(_InIt,_InIt,_OutIt,std::random_access_iterator_tag,std::random_access_iterator_tag) I would think the second one is the one to go with. I'm pretty sure this behavior is new with MSVC10 and I cannot recall this problem with older MSVCs. Unfortunately I don't have access to older versions right now. Anyone can help me out here? Here some simple test code: #include <vector> #include <algorithm> #include <boost/gil/gil_all.hpp> using namespace std; using namespace boost; using namespace boost::gil; int main(int argc, char* argv[]) { typedef bit_aligned_image1_type< 1, gray_layout_t >::type image_t; typedef image_t::view_t view_t; typedef view_t::reference pixel_t; typedef vector< unsigned char > buffer_t; typedef bit_aligned_pixel_iterator< pixel_t > it_t; image_t img( 10, 10 ); image_t::view_t v = view( img ); vector< unsigned char > buf( 100 ); it_t begin = it_t( &buf.front(), 0 ); it_t end = it_t( &buf.front(), 0 ); typedef view_t::x_iterator x_it_t; x_it_t it = v.row_begin( 0 ); std::copy( begin, end, it ); return 0; } Thanks, Christian