Hi,

Thank you for the submission as discussed per private email earlier.

I see where you're heading to, however, I don't find the provided example practical.  For one, there is a more efficient way to learn the width and height of a PNG image, than fully loading the image in memory first.  That would involve creating a `reader` instance and calling its `read_info` method.

I could add this as a tutorial, but maybe it would make more sense to provide an additional example program?  It is this short in the nutshell:

    std::ifstream file(filename, std::ios::binary);
    png::reader< std::istream > rd(file);
    rd.read_info();
    std::cout << "width:  " << rd.get_width()  << std::endl;
    std::cout << "height: " << rd.get_height() << std::endl;

Your examples with `getWidth/Height()` functions doesn't seem really practical too, unfortunately.  I would not include them in the tutorial.