I don't think arcane recovery should exist. Rapidly recharging magic is not the wizard thing.
I mean, if I had my druthers, yeah, I'd be there with you.
I don't think it is bad per se, but it could be much better. But as the core concept is "has a lot of spells" it is pretty easy to do passably, which they have done. But it is not terribly interesting.
A core concept that is fundamentally not very interesting should not be a core concept in the first place. Isn't the point of offering a class at all that that class should be interesting to play?
I'd argue that hunting for ancient scrolls to copy spells from is the sort of Indiana Jones style research which fits an action adventure game. But there could be more, and I'd gladly swap the arcane recovery for something that would support this theme. The issue is that logical direction is altering and customising your spells, but that basically what metamagic is and that's currently the sorcerers only unique thing, so we cannot give it to the wizard.
See, I completely disagree that it makes sense as "altering and customizing" spells that exist. Instead, I see it as meaning that Wizards should have a build-your-own-spells toolkit. Such spells should never be
as good as a spell that already exists...
initially. You have to work on it. Perfect it. Find all the foibles and fix them. Add correction terms to the calculations. Perform perturbation analysis. Maybe do a Fourier transform to see if there are missing harmonics. (I could regurgitate more random math/physics terms here but I think you get the point.)
Wizards shouldn't be cribbing in the margins, making tiny changes to
someone else's work. They should be, as I said, pushing the frontiers of magic. That's what they do; that's what they live for. Let the Sorcerers and Warlocks be stuck with whatever spells their biology or sugar-parent saw fit to endow them with; the Wizard
creates her own power.
The issue with them really isn't the power, it is the blandness. D&D spell schools are not particularly evocative (yes, not even the evocation) or interesting conceptually (apart the necromancy) so building subclasses solely around them is a bad idea.
Oh, you won't hear any argument from me on that front either. But therein lies the rub: traditionalism is actively in conflict with developing better, more interesting, more worthwhile flavor. You'll
never get an alternative that doesn't have at least 30% of fans committed and aggressively opposed to
some part of it, and that unity will ensure its downfall in the 5e playtesting space.
Is the topic of this thread just referring to the Sorcerer class in general (regardless of which D&D edition or 3pp it came from)? Or is it just referring to the Sorcerer class in D&D? Which is it because there probably are versions of this class that could make-up for the amount of dislike I have read on this thread.

One such version appears to be Laser Llama's take on the Sorcerer class.*
Any thoughts about the Sorcerer classes from PF1, PF2 or Level Up?
*I really love their Magus class.
I genuinely adore the 4e Sorcerer as a mechanical structure (it's a rock-solid "simple damage-dealer" design with interesting mechanics in each "subclass" in 5e terms), and if there were a way to merge it with the D&D Next playtest version, it would instantly become my second-favorite class ever (after Paladin, natch.) The whole transform-into-your-sorcerous-soul thing was just
oozing with flavor, and I'm still just....angry and disappointed that it got completely and totally canned, never to even be looked at again, because a vocal minority didn't like that it was too different from 3e,
an edition where Sorcerers weren't great.