“…Their idea is to split the error into two components where the first error is bounded by a computable dual-weighted residual and the second one, claimed small in [5], is estimated via equilibrated fluxes. An important focus whose rigorous investigation has been started only recently is the theory for nonconforming, discontinuous Galerkin, and mixed methods: let us cite in particular Mozolevski and Prudhomme [34] and Dolejší et al [13,15]. Finally, to the best of our knowledge, with the exception of [13,32,[42][43][44], all the above-cited results rely on the assumption that both the primal and the dual discrete problems are solved exactly.…”