My reservation with putting the value of specialists too far into the %mod direction is that they become less good at smoothing out a start-up city, and I'd have concerns that they lose too much value in the early game. For instance, I'd look for a way that the first Engineer offers the same amount of yields as it currently does, but the second and onward would be picked mainly for the %mod. Quick brainstorm for this: "[+Yields (maybe era scaling?)] while working at least 1 [Specialist]", maybe on the first building of a specialist-type.
On a more general note, I would prefer to consider Specialists versus Resource tiles, not generic tiles. No-resource mines shouldn't be competing with Engineers in my opinion; you already bought into the specialist slot, why should you need to make a decision over whether that's better than some un-supported improvement? If this means that improvements without resources get weaker to accommodate improved resources getting stronger, I think that's a good direction. I can see an argument that Features (Hills/Forests/Jungles) are sort of like their own mini-resources, but I think that's already supported by Mines/Lumber Mills > Farms.
This also gets to one part of the puzzle: should citizens be slot-efficient early (improve your resources and fill your specialists first), and "over"-growth just squeezes a little more out of the surrounding terrain, OR should citizens be a more consistent value per slot, then scales multiplicatively once you can feed specialists? I'm in favor of the former, but building consensus around either of these approaches will at least get us on the same page.
On a more general note, I would prefer to consider Specialists versus Resource tiles, not generic tiles. No-resource mines shouldn't be competing with Engineers in my opinion; you already bought into the specialist slot, why should you need to make a decision over whether that's better than some un-supported improvement? If this means that improvements without resources get weaker to accommodate improved resources getting stronger, I think that's a good direction. I can see an argument that Features (Hills/Forests/Jungles) are sort of like their own mini-resources, but I think that's already supported by Mines/Lumber Mills > Farms.
This also gets to one part of the puzzle: should citizens be slot-efficient early (improve your resources and fill your specialists first), and "over"-growth just squeezes a little more out of the surrounding terrain, OR should citizens be a more consistent value per slot, then scales multiplicatively once you can feed specialists? I'm in favor of the former, but building consensus around either of these approaches will at least get us on the same page.