It is clear that for the identification of pulsation modes a necessary first step is the unambiguous detection and extraction of frequencies from observations. It is no less clear that this task is arduous and full of obstacles. To name just a few: irregular sampling introduce correlation between frequency bins in the periodogram; noise level cannot be estimated properly when the statistical properties deviate from a uncorrelated stochastic process with Gaussian distribution; the false alarm test, which is based on a proper estimation of the noise level, also suffer from biases due to assumptions like an exponentially distributed noise; different fitting methods produce different results, etc. Here I will discuss some of these issues and introduce an ambitious project aimed to formalize a set of consistent criteria for a reliable frequency extraction in pulsating stars observations. This will produce a living recipe with the purpose of serving as a reference for the present and future data analysts of the ultra-precise observations gathered by asteroseismic space missions. That is, the necessary criteria to be a connoisseur of stellar light curves and identify pulsation modes with seismic models.