Your comment about the Sorcerer brings up a thought I've had regarding classes in 5E (and modern D&D).
I kind of feel like a lot of classes in 5E exist solely as a means to express a unique mechanic in the game.
So, all the classes in 5e have a history in prior editions. The Warlock is the 'youngest' being introduced in 3.5 and appearing in a PH only in 4e, followed closely by the 3.0-vintage Sorcerer, and more distantly by the 1e AD&D Unearthed Arcana Barbarian. Everything else in the 5e PH goes back to the 1e PH, at least as a sub-class or proto-PrC-like-classoid. And, everything in 1e PH was introduced in some 0D&D booklet at some point.
Even so, some of them were seemingly created just to introduce mechanics - the Thief (rogue precursor) brought de-facto dungeoneering skills, labeled 'special abilities,' to the game way back in the Greyhawk supplement, the very first of the booklets to follow the original boxed set, the Illusionist introduced specialization to the Magic-User. But others were character concepts - the Ranger was Aragorn, the Monk was Quai-Chang Caine, the Paladin was Lancelot & Percival/Galahad, the Barbarian, obviously, Conan.
I wonder if there is a chicken and the egg thing going on?
The Sorcerer is no longer necessary as Wizards get to spontaneously cast from their prepared spells. So they invent sorcery points to justify the existence of the Sorcerer as a class. Did the existence of the Sorcerer class force the creation of the sorcery point mechanic? Or was the sorcery point mechanic something that they wanted to be put in the game, hence it got stuck on the Sorcerer class?
The Sorcerer isn't there to express the wonder of sorcery points, it's there to differentiate the sorcerer form the other spell-casting classes (which is all of them, now), mechanically. The Sorcerer's a particularly egregious example, though, because if you go back two eds to it's origin, it sure looked like it was created soley as a vehicle to introduce spontaneous casting as an alternative to the much-criticized 'Vancian' model - an alternative that wasn't as utterly broken as the popular un-official spell-point/mana variants that proliferated back in the day.
In a sense, it failed, because wizards &c never stopped memorizing/preparing spells, in another it succeeded wildly, since everyone now casts spontaneously in 5e.
Is this a situation like the board game Root, where there needs to be different mechanical interactions for each class? Why is this required?
It's not required, it's at worst, perhaps, expected. It's a pat sort of game-design approach. Why is character A different from character B, because it uses mechanic X instead of mechanic Y.
What drives the need for a Sorcerer to exist when the Wizard class handles the magic spell-caster with tremendous versatility, already?
Prettymuch just the fact it existed in 3.0, really.
I'm ok with just the magic-user, but I do see why people want a little more differentiation. But why does the differentiation have to be so strongly expressed mechanically? Why does a warlock need a completely different mechanic to 'feel like a warlock' instead of a wizard?
Proliferation of classes, to me, really means proliferation of mechanics.
It does mean the game is more complex than it might otherwise need to be, sure. But, 5e, at least, has avoided some of the fall-out for such needless complexity by holding the line and not introducing new classes, spells, and the like at a breakneck (or even slightly neck-bending) pace.
I often wonder what is the point? Why does D&D need to constantly add new mechanics? Isn't the idea of playing in an unlimited fantastic world of magic, monsters, treasures, and the unknown enough?
5e isn't adding a lot of new mechanics as it goes. It kinda just dumped them all in the PH, and is letting them ride. Psionics has taken years to roll out and still hasn't appeared in print, for instance.