A brief review is given of what may be meant by a "catastrophic engineering failure", particularly as it relates to situations involving fracture. Whether "fast fracture" is necessarily the same as "unstable" fracture--and whether it makes any difference anyway--is discussed.
Since recoverable elastic energy, and irreversible remote plasticity, are volume-dependent quantities, but fracture work depends only on area, cube/square energy scaling principles are inherent in all mechanics of fracture. These different dependencies translate into the well-known experience that components and structures, made of materials which are appreciably ductile in laboratory-size testpieces, behave in a progressively less ductile fashion the larger they get. Eventually, above some critical size, their behaviour is globally elastic, and fractures are brittle. Conversely, when deformation zones are kept very small in normallybrittle materials, plastic deformation is possible, as shown by limiting sizes in the communution of powders and by the ability to micro-machine glass.
Analytical studies and experimental investigations of scaled cracked bodies, covering the whole range from globally elastic in the large, through to quasi-rigid plastic in the small are described and assessed, including consideration of dynamic progressive fracture in frameworks.