Roof windows of the Velux type are not usually considered to be an adequate means of escape. Firstly they can be difficult to get out of, either because of their height or the way they open (or both). With some it is very easy for the window to close on someone getting out and if they then try to get back in they get stuck.
Secondly, the escape is onto a downward sloping roof. It is often difficult to get a grip on the roof tiles and the head first downwards exit is difficult, frightening and, unless there is a flat roof within feet, dangerous in itself.
Moreover in a fire if a roof window is opened it creates an airflow which draws more oxygen onto the fire below making it more severe and bringing more combustion products into the room. In attic conversions, especially non-compliant ones, there is rarely any way of closing the attic off from the rest of the house - it can only be done from below (if at all) so unlike a normal room you can't close a door before opening a window.
Death in fires usually occurs due to asphyxiation. Contrary to popular belief it isn't the smoke which kills but the relatively transparent gasses which often precede the smoke (hence the use of "products of combustion" rather than "smoke").
Opening a roof window often produces an immediate enhancement of the fire which encourages panic in the occupants of the room. It also draws more poisonous gasses into the room. Unless people can get out very quickly they will soon lose the ability to escape.
There are however a vast range of individual designs. For any particular one the only reliable way of assessing its potential for escape is to try it. There are some basic rules -
- If you need something portable to stand on (other than perhaps the bed itself) to get out it is unsuitable. Sods Law makes it a certain that whatever you need to stand on won't be there if its needed.
- If the window can't be fully opened without fiddling with special catches it isn't suitable.
- If it doesn't have a flat roof immediately below which you can safely roll someone on to, it isn't suitable.
When you try your escape test do it blindfolded and have someone spin you around a few times in the middle of the room first. Try the same with the youngest person you might expect to have to escape and the oldest. If you don't find it easy to get out in these simple circumstances you can be pretty sure you won't in a fire. Children in particular panic easily and don't naturally run to escape from danger - they hide from it. After fires you find their remains in wardrobes, cupboards or under beds - even when there had been time for them to escape through smoke.
Risk analysis really needs to be taught to people. I know of a family who are completely paranoid about their children's safety and terrified that any accident will befall them. They have little plastic shields in every electrical socket, plastic corner protectors on every table and escort them across a sports field 100 yds to school and back every day. The two children (twins) sleep in an attic conversion. When you look down the stairs from the attic you can see the open plan living room and kitchen two floors below. The attic room has two Velux windows with child resistant catches which open onto a steep roof with nothing but the gutter before you hit the ground.
These idiots actually had the conversion built like that, they didn't try to get it done properly because "enclosing the staircase would have spoilt the open space".