Disclaimer: I never really researched, so I am honestly asking...not meant as a "counterpoint" for arguements sake...
Your #1 & #3 can be accomplished with stock arms, yes? Defiitely more convenient to just bolt them up vs. needing a press, etc. So I get that point. They certainly would not need to be as strong when moving the spring on the same axis as the strut.
It's the #2 that I'm really after. Are stock arms really that weak? They appear to be overly robust by comparison to the aluminum arms I see on many vehicles.
I guess I'm approaching this from a "what would I REALLY gain over the stock part" as the divorced spring / strut setup is rediculously inefficient from a spring rate required / weight perspective. It's the cost of said arms that makes me say "why would I want that?".
I don't know if tall ball joints are available to fit stock arms, never looked into it myself. Cost no object the three best ways to address this would be spindles (maintain the already very good stock arm geometry and retain full strut travel) or a tubular K-member that compensates for lower ride height by relocating the inboard pivots upward (and forward, in at least one application) or keeping the car at stock height because the folks who engineered it did a good job and compliance does not equal bad handling.
Tall ball joints are mostly a budget-conscious way to still do a better job than
just lowering ride height for cosmetic reasons with stock geometry.
Ford Racing M-3075-D control arms (stock on 03-04 Cobra, OE performance upgrade briefly available in the Ford Performance catalog) have slightly stiffer rubber bushings to increase steering feedback and reduce deflection. Aftermarket tubular arms like BMR have urethane bushings that also are smaller in diameter for much higher stiffness - this is always going to communicate more of the road to the chassis and steering wheel which might be better in competitive scenarios but worse in everyday road driving. I think a lot of people aren't honest with themselves about how they're going to actually drive and don't give the rockstar engineers who developed the car enough credit for the balancing act of making a sellable product.
The stock arms are not weak but they are also not as rigid as a tubular arm because ultimately they are a flat piece of steel that's had a texture pressed into it. You can imagine the difference between a piece of paper you've folded a couple times vs. one you've carefully rolled into a tube and taped - same idea here. The stock arm is a lever with its fulcrum in the middle (the bottom of the spring) and it's held in tension vertically but stiffness in every other direction is kinda hamfisted and probably accounts for the manufacturer just throwing steel mass at it. The ball joint prevents direct twisting force but since the bottom of the coil bucket, the center of the ball joint and the instant center of the pivot bushings will almost never be directly inline with whatever force is acting on them they will always twist a little bit and that means deflection.
This is of basically no importance to actual handling but can mean a lot to steering feel and communication. The weight saving and
possible reduction in strut bind of going to a tubular arm would likely be responsible for most or all improvements in lap times,
if any.
I speculate they were designed that way for economy, somebody could throw a sheet in a press and pull out a formed arm in 4 processes (stamp, fold, coat, insertion) , probably under 30 seconds net work per arm vs 10+ minutes per arm manufacturing welded tubular ones. I do not know enough about the state of aluminum casting in the 90s but I think it's likely even if material cost is identical there's no way you could manufacture aluminum arms with as little labor or process cost as stamped steel.
To give the original engineers credit the divorced spring and strut arrangement means that the coil spring is traveling half as fast as it would on a coilover and since kinetic energy scales logarithmically your net (dynamic) unsprung weight on a divorced spring can be potentially less than on a coilover even if the actual spring is heavier. This also makes the K member the main stressed element in holding up the car's weight which (likely) provided some weight savings in engineering the firewall/pan/fender structure and brought some of the twisting forces acting on the chassis inboard to make the whole chassis torsionally stiffer. 70s car engineers might have been under a lot of economic constraints but they weren't dummies.