Historically, IBM keyboards like the 84-key on my trusty ol' XT put control left of A where it was on every other DTE/computer keyboard at the time. Unfortunately the AT keyboard put it back where it was on typewriters.

Personally, I remap caps lock on modern keyboards back to control as I use it far more often than even the shift keys. It's also how my finger memory was programmed an age ago. On typical terminals ^H is easier than reaching for backspace, ^[ for escape, ^I for tab, ^U for kill line not to mention all those handy vi and emacs editing commands. I suffer from tendinitis across the tops of my hands which is magnified by reaching for keys more than a keycap away from the home row. Control-below-shift is the worst with my little finger curling up to do the reaching.

A friend of mine drives his 928 is PCA club races. He uses levers to actuate the brake and throttle. He uses a wheel to control the steering.

That's funny and I see you're joking. Maybe you're thinking of the rudder pedals? My RWD cars have always had yaw control. The FWD cars too with some coaxing. Gas, turn, lift, countersteer, gas.... Given enough power you really don't even need to lift.

Now controlling that with the twitchiness of a joystick even if they do get the feedback right is another matter entirely. And that's coming from a person used to a very quick "sneeze and die" steering ratio. Lateral movement controls in a car are largely inferior than rotary (or pushbutton) controls because they are more easily affected by the movement of your body within the car during bumps and turns. I can hang onto a wheel or dial to steady myself and still provide usable control inputs, not so with a slider.

To counter the twitchiness, maybe you introduce acceleration into the movement: precise on-center and progressively faster steering (maybe not accel/decel) as you get to the end of travel. I predict that increases the learning curve and implementation difficulties, particularly between differing marques.