No one in my group ever cared to play a 4e sorcerer, so my concept of the wizard/sorcerer distinction is mostly the 3x version. The way it seemed to me was that wizards and sorcerers shared a spell list because in the end, they were bringing about the exact same phenomena. Wizards did it by careful study and understanding of the phenomenon itself, sorcerers just sort of winged it with instinct. IIRC, the PHB said some sorcerers "claimed" a draconic or extraplanar bloodline, not that this was in fact the established source of their power. Having known an autistic savant who could play the piano like a master with no formal training whatsoever, the idea of a sorcerer just somehow figuring out how to do what a wizard spent years learning never sat ill with me.
Anyway, this to me is what the 3.x mechanics tried to model. The wizard had to undergo an intensive and life-long study of how magic worked and as a result, could learn a lot more spells. The sorcerer was stuck with whatever he could figure out with his own instincts, hence fewer spells and slower advancement. Whether this is a good way to model the difference or not is obviously up for debate. It doesn't (to me anyway) provide a solid rationale for a wizard only being able to use a prepared spell once without tacking on the fluff of "fire and forget" or "preparation means the spell is 90% cast and just waiting on the last component to set it off."
All that is to bring me to a slightly different idea. Consider two things:
First - Part of the goal for Next was to recapture some of the original feel of older editions (0D&D, 1e, 2e). Initially, the basic arcane class was called "magic user" specifically to avoid any connotations association with terms like wizard, sorcerer, warlock, witch, conjurer, magi, etc.
Second - So far, the idea seems to give each basic class build options. Fighters get fighting styles, rogues get schemes, and clerics get domains. Each class got two builds. It seemed like the arcane traditions were slated to occupy this spot for wizards. What if instead, the casting mechanic were to occupy that spot?
So in comparison to what we have so far with the other classes you might get a class called "Magic-User" (yes, they'd have to rename the background currently using that title) with the basics: d4 hit dice, poor weapon attack bonus, good magic attack bonus, no armor, few weapons, spells per day table.
Then add "arcane tradition" as the analogous choice to fighter style, rogue scheme, or cleric domain. The two basics could be "wizard" and "sorcerer".
Wizards would have a spellbook and the ability to record any arcane spell of a level they can cast in the spellbook. But they have to specifically prepare spell for each slot in the spells per day table.
Sorcerers would not have a spellbook and be limited to knowing a number of spells for each level equivalent to their slots per day, but could cast any spell they know by expending a slot of the same level (or higher).
The wizard gets more spells to choose from with the trade off that he has to prepare them and run the risk of choosing a spell that turns out to be suboptimal or possibly even useless in the day to come.
From there you could add a warlock tradition, an illusionist tradition, and so one each with a mechanic that modeled the fluff of the class.
Incidentally, if the above were to come about, I'm not whetted to calling the class "magic user" and making the traditions "wizard" and "sorcerer" and "warlock" and so on. If want to call the basic arcane class "wizard" and let the sorcerer and warlock be something totally different, that's fine.
I guess the major weakness here is that it rolls the different casting methods into wizard options and leaves the cleric out.
