I read with interest but no pleasure the detailed report on the collapse of the new Paris Airport building.
With the benefit of hindsight it appears that the constant thickness of the concrete shell - at less than one-hundredth of its span and with some 50% of the width sacrificed for windows - was rather thin.
This would have reduced dead weight and stiffness and hence reaction forces at the supports, but would tend to accentuate stress levels and increase movements associated with resisting loads, some of which - for example wind and temperature - would be reversible.
In the vicinity of the escalator cut-outs the concrete compressive stresses may have been high. This may have resulted in accelerated creep progressively eating away from the ultimate strain capacity of concrete. When combined with reversible effects, this could potentially cause failure.
Although this is only a conjecture, I believe concrete stresses for the serviceability conditions should be controlled to within a reasonable limit, especially for unusual structures.
Furthermore, greater emphasis should be placed on proper understanding of structural behaviour, including second order effects, and not merely on automated analysis of structures which can sometimes be misleading.
Arvind Kumar, (F), 2 Penn Road, Beaconsfield, Buckinghamshire, HP9 2PX