In principle the analysis of positions and shapes of diffraction peak profiles can reveal the imperfect crystalline structure of a sample. The parameters 'size' and 'strain' as obtained from the established Warren-Averbach method need to be used cautiously: in essence the order-independent and order-dependent parts of the line broadening are determined. These require interpretation in terms of structure parameters, for which additional information must often be available. In a case of pure size and strain broadening the order-independent part of the broadening can be interpreted straightforwardly in terms of a size distribution, whereas the order-dependent part cannot readily be related to a specific state of strain. Recent progress in this area is discussed. Non-ideal standard specimens introduce systematic errors. Within the framework of the Warren-Averbach and the single-line Voigt methods, it is shown that this non-ideal situation introduces a constant error and that relative determinations of size and strain are feasible.