One of the major interests of ecology is to understand and predict species diversity in a community. Two models have historically been proposed: the niche assembly theory, based mainly on the differences of the species use of resources; and the neutral theory of biodiversity, where similarities between species, not their differences, explain the high diversity of many natural systems. There was later on a momentary agreement between scientists when it was considered that both models could generate similar patterns and intermediate models were proposed unifying both approaches. A recent proposal indicate that unification might be justified by simply considering the process of the introduction of new species by dispersion and/or speciation. Nevertheless, shortly afterwards, a close to the biodiversity neutral model was proposed, where species co-occur in a community not because, but in spite of, ecological differences, and which seems to be the best predicting approach considering existing empirical data. Despite these efforts, controversies on the basic concepts of current models and their relationship with empirical evidence make necessary their re-evaluation, to include other important components like species turnover. This revision suggests that extension of the current models or new proposals (if necessary), including and assessing the impact of new components in current models, will allow us to understand and predict more accurately diversity and community composition, its maintenance and changes over time and spatial scale.