FFB is always an interesting discussion! These are just my personal thoughts, and don't necesarily fully reflect what we at Reiza will end up doing.
For cars with downforce, it would be a good idea to have a speed dependent FFB multiplier, so hairpins and fast bends both 'max out' the FFB, giving you as detailed feedback as possible. This is not technically too hard either.
There is low speed and low speed; on the race track low speeds will still be 30mph / 45km/h. FFB behaviour near standstill is a different thing but not really important as race sims are not parking sims. The latest realfeel can mix some Leo with some TechAde so you get some resistance at standstill, so it doesn't feel completely weird! For GSC and in real life, at low speeds you encounter fewer track bumps per second, and their impact on the car is less. Road information through the wheel will fade out as you go slower, but that is logical. Other than that, I can't really see what could be 'missing' from low (30+mph) speed, only that downforce cars have weak FFB at these speeds. So my question is, what tends to be missing in sims / low speed FFB?
Adding effects to compensate for not being in the real car is something I never felt the need for. I use experience and my eyes / ears to 'feel' the car. FFB doesn't tell me, because it simply doesn't always work like this, if I've reached the maximum grip or when the back end is sliding. It all depends on what is happening. FFB is a force that tries to keep the front tires rolling in the direction the car is moving or sliding. This is great because the center of steering isn't always 'straight ahead'.
Getting in oversteer, the center of steering moves towards opposite lock. FFB will guide you towards this new 'center' and helps you apply just the right amount of opposite lock. Too much opposite lock will make the FFB force change direction again. So in short, opposite lock is first aided by FFB, then after you've reached the ideal amount of opposite lock, FFB will try to prevent you from applying more. FFB will try to catch the silde for you!
Getting out of oversteer is hugely important. Doing it badly can cause nasty tank slappers and expensive crashes. FFB should have prevented you from applying too much opposite lock, but as you take throttle away and the car begins to straighten out, FFB again will help you find the new stable center position of the wheel, which is likely to be more or less straight ahead. Both the timing and the amount of steering is very important and a good FFB wheel in a 'sluggish big car' might even get out of oversteer by itself!
Letting go of the wheel in oversteer is NOT a good idea (most of the time), as you probably want to manually apply some over or under correction combined with your experience as a driver to get the car to do what you want. FFB is just a HUGE help, it tells you where the car is happy so you can make small adjustments from this happy place to get the car doing what you want. This goes for drifting but just as much for racing at smaller angles. FFB tells you where the 'center' of steering is, and all the action happens say within 45 degrees of steering wheel lock around this center.
All I really care about here is to feel forces build up as I move away from the center. Ideally this is stong enough to really let me feel where the wheel wants to be, but all I care about is "a force" in "the direction towards the current FFB center". I do NOT look for any specific micro detail. There is no wheel that will tell you the tires are 80 or 90 degrees, or if they are 20kpa over inflated. I'm sure my Bodnar wheel will notice it, but while driving, us humans are busy with a load of things. A few % change in forces are unlikely to be noticed.
The other part I find very important is locked brakes. FFB will go very light when you lock both front tires. It doesn't always tell you the inside wheel is locked because this is usually a lightly loaded tire that only has a small effect on the total FFB. The outside front, highly loaded tire is the one causing the main forces, but this is the case in real life as well.
So I just want there to be a force around the 'center' of steering, where the center can move about depending on under/over steer. I really don't look for 1% changes in FFB because of tire wear or heat, as even though my Bodnar wheel will measurably feel different, us humans won't really feel the difference. On a G25 class wheel, subtle force effects are simply more likely to exist between the ears rather than between the hands.
Even something seemingly simple as the FFB telling you how much grip the tires have isn't so easy. In steady cornering, FFB will start to fall when grip is still increasing. Sure you can feel the FFB is increasing and stabilizing, or falling off, but that tells you nothing about your G forces. If you only felt the tire part of the FFB, some cars would have NEGATIVE force feedback by the time you reach maximum grip. Odd eh! So FFB is a rubbish G meter.
Given limitations of the G25 class wheels, adding speed dependent FFB for downforce cars makes sense. You would feel a decent amount of FFB in hairpins rather than feeling pretty much nothing, while having no clipping of forces occur at higher speeds. I see no real reason for adding other effects, like vibration around 'peak grip' or canned bumpyness or kerbstone effects. If the circuit modeling is good, you will feel kerbs and the road just fine.
But that is my 2 cents (well, more like 2 euro!) on the matter. Again, even with my G25 I never really wanted to feel specific things. It did the main job that I explained above. Nowadays with my Bodnar wheel it does the same, just a lot better and stronger so it is even more of a help telling me where 'neutral stering position' is. Then its up to my experience to make the car go quickly around the circuit..
But that doesn't mean there aren't any good ideas to improve FFB, so if you have ideas, I can offer my perspective on them.