The paper examines the fundamental aspects of diffraction techniques when applied to the investigation of the structure of non-ordered materials, with the aim to assess the meaning and limitations of the information obtainable. The need of using structural models as representation of the tridimensional objects under scrutiny posses questions which are discussed in some depth. One aspect of the problem regards the diffraction results and their reliability when used as experimental reference for testing theoretical models. Another side of the problem is related to the model itself and to its capacity of representing the structure of systems which are far from thermodynamic equilibrium. In this connection, the concept of ideal glass is revived and the hypothesis that the model may be viewed as its representation is discussed.