"Some time in the fifties car makers discovered that if the bonnet opens while you are mobile bad things can happen.
They decided to have two latches one holds the bonnet closed good and proper the other just holds it more or less down but stops it coming any further up. It also allows it to project a bit so you notice the upsticking and vibrate a bit so you wonder what the noise is
They did this because they knew every so ofter one latch just isn't enough"
We call this method of working a FAIL SAFE do we not?
Just as a point of order, no that is
not 'fail safe'. It
is belt and braces, which is no bad thing.
A 'fail safe' is a situation whereby the failure of the mechanism 'naturally' fails in a safe way. An example would be (if you'll excuse the mention of railways) electro-pneumatic braking system on trains, whereby the natural state of the individual brakes on the wheels is that they are held on using jolly big springs, and the action of the train's braking system is to pull the brakes
off using pneumatics. If any part of the train's system fails (for example the train splits, causing the brake-line to separate, releasing the pressure from the braking system) basic laws of physics (Newton & Hooke) cause the brakes to revert to the applied state rather than any backup mechanism* per se.
A
fail safe cowl latching mechanism would be one whereby if the latch fails, the cowl nevertheless remains in the safest position while in flight (which is to say, closed.) You can envisage designing the cowls such that while in flight the flow of air passing the cowl naturally tends to push the cowl closed rather than tearing it off. (Simplistically, arranging the cowls such that their hinges were at the front and opened like the petals of a flower, would seem to be a 'fail safe' design in respect to the latches.)
(Note I'm not recommending that design - I'm sure there are plenty of reasons they're
not designed like that and plenty of reasons why they are designed as they are, but that would be a design whereby the cowls would seem to fail safe in the event of latches not working.)
* some smartarse (me!) is going to say, 'what happens if the springs don't work'. Well, that's a different failure than the train braking system failing, so doesn't affect whether or not the design of the braking system is fail safe. I suppose one could start to worry about whether or not springs need their own fail safe, but to a certain extent worrying about whether or not springs are going to spontaneously stop being springy is a bit like worrying about whether or not gravity is going to stop working mid-flight. Possibly worth worrying about if you worry about the Bermuda Triangle, but otherwise you have to draw a line somewhere...