The implication of modelling concrete fracture with a fictitious crack of zero fracture process zone (FPZ) height is addressed because FPZ height, in reality, is not zero and is bound to vary during crack growth. The ligament effect on fracture energy GF is explained by thenonuniform distribution of a local fracture energy gf showing the influence of specimen boundary and variation of FPZ height. Thenonuniform gf distribution is then used to determine the size-independent GF. The recent boundary-effect model based on a bilinear gf function is confirmed by the essential work of fracture (EWF) model for the yielding of deeply notched polymer and metal specimens. The (EWF) model provides a theoretical basis for the bilinear gf distribution. The principal rationale of the boundary-effect model, the influence of FPZ height on fracture energy, is supported by experimental observations of thickness effect on fracture toughness of thin polymeric adhesives between metals.