Inversion of gravity profiles by use of a Backus-Gilbert approach; discussion and reply