In which I disagree with Brian Deer on the issue of how to deal with scientific fraud