Sorry, I won't be able to respond to everything right now; I may do more on edit later.
One could argue that in such a case there is little point in having an FoF, and that we should instead instead add a compromise component to FoW/FoE or have them consume gauge while FoF be just... no stance at all. Of course, we could say the same for the present situation. Nonetheless, I'm glad you're considering them in new ways.
You can kind of consider what you're moving towards as what's called "back-loaded" or "back-end(ed)" effect, whereas immediate Damage and Mitigation bonuses would be called "fully dynamic", and Haste "to-GCD dynamic". (For comparison, the current FoW, in the form of GL4, is a "to-GCD dynamic" effect with "windowed/limited access" since one can only enter it on a Coeurl skill.)
You could, for instance, go entirely back-ended, generating something over time, from attacks dealt, potency dealt, relative potency dealt, or any combination thereof, such as by having Fire generate Attack Power (affects all sources of potency, including even Second Wind), Wind generate Skill Speed (AAs, GCDs, and DoTs a bit further), and Earth generate (Magic) Defense, each of which fades over time. The problem is two-fold: (1) an back-ended Defense buff is going to feel useless or entrapping until one has a fight fully memorized AND somehow needs the added mitigation AND can get more out of sacrificing damage during the time prior to proximity AoE than by moving that little bit further away (unless of course there's a damage component to FoE as well), and (2) it's just far less intuitive and harder to track.
A logical alternative, then, would be to build something generalistic, and then have them spend from that same resource pool for their (additional) effects. In that way there's no sense of entrapment. Swapping to FoE before a big hit will consume just enough resource to keep you alive -- no more, no less -- while FoF may consume resource based on potency dealt (its damage bonus refunding an increasing part of that resource cost as not to make its value even across all GCDs), FoW per strike, etc.
This is another way to go, and my personal preference if ever we returned all jobs to using MP in some way. I despise wasted UI components, so I'd like to see that shift regardless.