Computational modelling of damage in brittle and quasi-brittle materials needs some coupling between micro- and macroscopic crack initiation and evolution, up to their non-negligible softening behaviour. Most such approaches contain ad hoc evaluations, with some physical and engineering motivations, namely those connected with massive application of steel fibre-reinforced concrete and similar composites in building projects, but without any proper mathematical existence and convergence analysis for the time development of damage. This paper presents a possibility of such deterministic analysis on a selected model problem of structural dynamics, supplied by comments to useful directions of generalization. Several application examples document the feasibility of such approach, up to its software implementation and real data validation.