Lucy's model of contact binaries has been serving well since its development in 1968 as an excellent conceptual framework for internal structure and evolution studies of the W UMa type stars. Its over-constrained structural demands were recognized early, but -from the observational point of view -it was particularly appealing in its ability to explain W UMa-type binary light curves. Large numbers of photometric light curves of W UMa type binaries have been solved for geometrical parameters utilizing the clarity and simplicity of Lucy's model to the point of frequent over-interpretation of the meager lightcurve information content. Many, single, photometric solutions do not translate into a satisfactory understanding of these binaries. As the David Dunlap Observatory radial velocity program of bright, close binary stars (conducted in the last 10 years before the observatory closure in 2008) has shown, there are problems when spectroscopic results are confronted with light-curve solutions: Very frequently, the crucial parameter of the mass-ratio determined spectroscopically is entirely different from its photometric analogue (i.e. the one determined assuming a strict adherence to the Lucy's model), the third star incidence is high (which makes light curve solutions difficult, if not entirely wrong, even for totally eclipsing systems) and -last but not least -the model has serious difficulties for the best observed case of the crucially important, very low mass-ratio system AW UMa. One would argue that it is spectroscopy which gives correct results, but it requires access to large telescopes, extensive use of special techniques to analyze strongly blended and rotationally broadened spectra and is simply much harder to do than photometry. If the light curve solutions deceive us into believing in correctness of Lucy's contact-binary model and if only a fraction of (or none among) W UMa systems can be explained by that model, then the subject of contact binaries needs new ideas and new, mostly spectroscopic, high quality observations.