The problem of BL Lac classification is a long standing one and it is mainly due to the subjectiveness of selection criteria used in the definition of BL Lac samples. For instance, an object will undoubtedly be classified as a BL Lac if it shows flat radio spectrum, high optical and radio polarization, featureless optical continuum with weak or absent emission lines, and variable flux and polarization. However, the problem arises when the object shows some but not all of these properties. In face of this difficulty, different authors (Stickel et al. 1991, Stocke et al. 1991) have tried to make a systematic analysis of the data and it has been common to classify as BL Lacs those objects whose strongest emission lines have equivalent width (EW) ≤ 5 Å. Another common criterion is to require the 4000 Å break contrast to be ≤ 0.25. Nevertheless, both of these criteria are rather arbitrary and more directly related to practical observational considerations, than they are to any physical distinction between objects. What is proposed here is a slightly different approach; it is proposed that we take a step back from common classification and that instead of imposing strict selection criteria, we create a multi-observational parameter space to investigate any breaks in the distribution of observed properties that will help clarify the distinction between BL Lacs and other flat radio spectrum sources.