Steve, the pragmatic approach definitely seems to have some merit. However - and please forgive me if (a) I haven't understood all of the issues here, and (b) seem to be labouring the point - the question "when is a room a room" seems applicable to a wide range of problems. Ken's response seems to suggest a door-based assessment and some "established practice" here. I've looked at the other thread.
It may well be that a FRA-based approach to inspection removes some of the blanket "inner, inner room? just say no!" reactions, but I can't help feeling that there will be a presumption that an FRA ought to be at least based upon, or reflective of, the "established practice" that has gone before. If an FRA ever goes to court, I suspect (with some legal knowledge, but less FRA knowledge) that wording would be important. So, for instance, if one considered "Room X" in the FRA (which would be an inner, inner room), opposing counsel would almost certainly seize upon this as "a room" because the FRA says it is, and then argue based on previous pracitce based on "inner, inner rooms". Conversely if one called it "Sub-area 3 of Room Y" then one might very well be asked to justify the decision not to call / classify it as a room. As we all know, courts aren't always rational, and the case could end up hinging on this, even if it were ultimately irrelevant to the question of whether means of escape, AFD, etc. were actually appropriate to the risk.
Therefore I suppose what I'm really getting at is, in current established practice, how is the distinction made? I'm not trying to argue the merits of one approach to risk assessment or warning over another, I'm just trying to establish what the current understanding of the term "room" really is, since I don't seem to have been able to find a specific definition.
Thank you, all, for your patience here!