H II regions, ionized nebulae associated to massive star formation, exhibit a wealth of emission lines that are the fundamental basis for estimating the chemical composition of the Universe. For more than 80 years, a discrepancy of a factor of around two between heavy-element abundances derived with collisional excited lines (CELs) and the weaker recombination lines (RLs) has thrown our absolute abundance determinations into doubt. Heavy elements regulate the cooling of the interstellar gas, being essential to the understanding of several phenomena such as nucleosynthesis, star formation and chemical evolution. In this colloquium, I will talk about the solution we found for this problem. The best available deep optical spectra of ionized nebulae show evidence of the presence of temperature inhomogeneities within the gas. These inhomogeneities affect only highly ionized gas and cause the abundance discrepancy problem. This work implies that most metallicity determinations, those based on CELs, must be revised as they may be severely underestimated. The evidence indicates that this effect could be greater in regions of lower metallicity, such as the JWST high-z galaxies. Fortunately, since the temperature inhomogeneities are concentrated in the highly ionized areas of the nebulae, it is possible to correctly determine metallicities using CELs, following our proposed empirical relations.