I personally would have preferred bigger drawbacks for the blueprints. They thus would be sidegrades and specialisation instead of direct upgrades.
Just a random example: while an overcharged weapon can do up to +70% damage, why does it still have to have its regular firing speed? Wouldn't it make sense if bigger shots / more energy used per shot takes a longer time to reload/charge the weapon? So the +70% advantage of damage could also come with a ROF reduction of -40%. That would -still- leave the weapon with a gain of 2% damage over time and with +70% of the punch, while requiring less ammo.
Ideally (for me at least), the solution would be to put enough thought into the mod's that there's actually a reason to deliberately choose the mid-tier mod's instead of just aspiring to G5 in all things.
If you want to build an all-rounder then you'd deliberately choose G3 mod's and you'd only push on to G5 if you wanted a ship that's seriously focused on one role.
Used to be a bit like that with PPs and PDists, before XFX were a thing.
A G5 PP provided a lot of power but it'd toast your ship and a G5 charge-enhanced PDist would recharge super-fast but would only give you a few seconds firing with a Beam laser.
Now all the mod's improv your ship in almost every way, with barely noticeable drawbacks.
On topic, I suppose I should say that I'm not sure module sizes is really a problem that especially needs fixing.
I mean, back when we had the ADS and DSS, it just wasn't possible to build, say, an exploration ship without wasting some slot-capacity.
Now, with the extra slot(s) and the removal of the ADS, the only C1 module that you really "need" is the DSS and all ships now have at least one C1 slot to bung it in.
If you want to fit the training-wheels, that's up to you but if you decide to get rid of them you won't be wasting the slot-capacity they require.
I guess some kind of "Rack Unit" would still be handy for things like limpet controllers though.