You know there was a D&D edition that did what you describe. 4e, and my understanding is 5e has solidly trounced it in terms of popularity. It's possible when designing games in general to be too balanced, and the D&D player base has straight up proven it's okay with or prefers a level of imbalance if it means they get more variety with their characters. So the trick becomes creating that variety while keeping it as balanced as possible and not breaking verisimilitude too badly. Which brings me back to my chart from before and the question about long rest, non-magical resources. The only ones I can think would be more narrative focused like calling in a favor, or some kind of Schrodinger's tool that a character picked up in the past and it becomes what it needs to be now. I am not sure either of those work that great.
So that brings us back to trying to modify things so the short/long rest balance is better, both in the classes and DM guidance. Sorcerers using Spell Points on a short rest is easy (I keep thinking they should be Spell Points in general, I never thought much about the rest cycle). I suggested a Warlock-like Druid class (and provided an example starting point in another thread). Converting 1 of the Ranger, Paladin, and Artificer to some kind of short rest alternative to their spell slots would be ideal, but I am not sure which one would work best, and that may be the most involved conversion.
It's not theoretical rules. It's assessing the value of something by comparing the cases with and without it. It would be like comparing a Fighter class without the Extra Attack progression to the standard Fight class to see how much difference the Extra Attack progression makes on its damage. The spellcasting multi-class changes what you get when multi-classing into a full caster. The nature of that change as a buff, nerf, or something else is informative. You clearly see it as a buff, hence why you think casters get some special positive treatment.
You may be right the multi-classing rules on spellcasting on average are more of a net buff than nerf. I am still on the fence, but's not as clear cut as I thought. The locked in ability to share spell slots is definitely a buff though. That's what makes Sorcadins really possible. Which makes me wonder if there is something to the reverse of this thread title. What if Extra attack still didn't stack (though it did have some alternative bonus if you had more than 1 instance), and spell casters didn't have a shared slot progression? By that I mean each caster has its own spell slots and those can't be used to fuel the abilities of another class. Warlock Pact slots also wouldn't be usable to fuel the spells or abilities of another class. It would probably be a bit of a housekeeping for PCs that were built as dual fullcasters, but it might make multi-classing more balanced in general, and it would certainly be more consistent. Something may still need to be done with the 1st 1 to 3 levels of class being more valuable on average encouraging dips.