There is only limited new understanding of the structure of the evolvi
ng dynamic metal/oxide interface itself but there is valuable circumst
antial evidence from cooled specimens and from other interfaces. Ident
ification of flaws for crack initiation, definition of a composite def
ect and elaboration of delamination models, involving a plastically-re
laxed crack tip region, have been undertaken. Loss of adhesion has rec
eived more attention, by the alternatives of film buckling or crack pr
opagation along the metal/oxide interface by wedging, including cases
where stress relief occurs by creep. Mapping of scale failure modes ha
s progressed. Computation and modelling using electronic and atomistic
models are useful, but ''clean'' metal/oxide and contaminated interfa
ces require in-depth work. Stress development and relaxation are compl
ex. Models involve dislocation climb and the role of vacancy transport
and cavity development. Sweeping of various particles by moving bound
aries is receiving renewed analysis. The role of convoluted metal/oxid
e interfaces in promoting or diminishing adhesion has been analysed qu
antitatively. Explanation of reactive element effects by the embedded
atom method is instructive, as are the relative roles of S and P in we
akening interfacial adhesion.