In the recent Legends & Lore article, Mike suggests that ability score improvements could be a simple alternative for feats. This is not something I would use, because I don’t think it’s a well-crafted trade-off: a bland bonus that is likely to affect the thing you do most often versus something specific that might add some texture. It also encourages numerical inflation.
The suggestion that it will be common for characters to “raise their key ability to 20” seems particularly unfortunate. Players often use ability scores as a form of personality profile. How boring is it that all melee fighters will ultimately have 20 strength? Isn’t class level (with attack bonus and so forth) supposed to capture the idea of progression in class competency?
Having a prestige class be essentially a preselected feat chain is not a bad idea though (sort of the high level equivalent of the lower level “specialties”) and is easy to tie into the setting diegetically, which I like. It also allows people who enjoy complex character builds the opportunity to mix and match feats but means that players who are not interested in that minutiae can just go by high level flavor, like Warhammer careers.
I don’t think there should be any mechanical prerequisites for feats though, except maybe level. Needing to consider the dependency graph for feats in earlier editions is a big contributor to the overwhelming complexity of the feat system. The model for magic-user spells is a good one, from a game design perspective. Distributing the spells over multiple levels keeps the initial complexity down, but allows for significant individualization of characters over the course of play.