In many cases either/or works just as well and there is plenty of anecdotal evidence posted here and elsewhere to support this - but it really depends on the species and its particular niche adaptations and phenotypic plasticity etc. But overall, and if there is any doubt, it remains sound horticultural practice to feed both locations - especially in the absence of relevant research on all 400 or so species of macrophyte.
With regards substrate leaching nutrients in to the water column I think that's a two way door. I strongly suspect that eventually nutrients will stop leaching out of the substrate as it becomes exhausted and the substrate in turn will become a net importer of nutrients - if it has a high CEC like Amazonia. But the net flux will probably vary for different nutrients, not least due to demand.
This is another reason why good flow and distribution are so important, it helps to drive nutrients from the water column in to the substrate where they also become available for root uptake regardless whether a substrate is inert of not - so in other words by dosing the water column we are inadvertently supplying plant roots with nutrients as well. I think this particular mechanism is often overlooked.
Conversely, I also think that the nutrient contribution made by dissolved solids etc in tap water, and from fish food is underestimated in many cases - even Walstad type tanks may get a fair amount of inadvertent water column dosing. Overall its entirely possible that in one way or another we've all been feeding both locations without necessarily being aware of it...
...What was the OP's question again...I've kinda lost track...Oh yer, it was about CO2. Definitely go for it...you'll have a carpet in no time. Oh and don't forget to dose.