Before the noble and learned Lord sits down, may I ask him this? I respect his opinion, for obvious reasons, but one issue he did not address—it was one of my arguments for why these clauses should stand part—is the difficulty of proving the threat at the beginning of an investigation. It is not straightforward. We have to say that someone out there is going to kill this officer or try to attack them—that there is a threat to them in some way. Of course we all make our best attempts to assess whether that is accurate or not. He describes the present system as a blanket arrangement, but actually there is only an assumption, which can be removed, and in the Kaba case was removed. That leaves the officer at risk of that decision being automatic—that is, to be named if they cannot prove otherwise. Why should they bear the risk of being named, when the reverse could allow, first, an assumption they would not be named, and if later that changed, they could be named. What we can never do is name someone, then introduce anonymity—so it is a one-way valve that surely the law might help to respect.