For Mike_Bracknell
The rudder failures weren't really down to a fault with the original design, here's the story as I remember it:
The ctrl surfaces are made of a honeycomb-core bonded to the skins, essentially. They originally had a blunt trailing-edge, as was then de-rigeur with supersonic design. At some stage it was decided that a sharp trailing edge was actually beneficial so they had an extension fitted, which had the unfortunate effect of allowing a certain amount of water ingress to the core. Heating and expansion of this lead to disbonding and ultimately failure of the surfaces. (I suspect my engineering colleagues will have a much better and more accurate explanation).
Now - here's the important bit, and another example of this aeroplane's excellent failsafe engineering; Concorde had two rudders, one above the other (same as the 747). Each is driven by one dual-bodied PFCU. You ABSOLUTELY don't want a PFCU endangered by ctrl surface damage so each surface is divided in two, either side of the PFCU control horn.
Visualise the PFCU attached to the centre of two surfaces with an end rib on each, but skinned to look like one surface. Therefore, in the case of the surface suffering damage, it can only spread to a point short of the all-important PFCU. Look at the rudder-failure pictures and you'll see what I mean.
So - far from the 'rudder' breaking up, the reality is that half of one of the rudders had failed.
It was somewhat inevitable that Concorde's control sfcs would suffer, given the horrific loads they endured, and this was dealt with at the design stage. The elevons had the same sort of design.
It does of course look bad when you land with bits missing and this, plus the Regulators and company safety depts ensured that eventually some HUGELY expensive replacements were built.