Mike Mearls has a design column up talking about Wizards. There are a few interesting things here, and also a few possibilities that I don’t think would suit the kinds of games I like to run. But before I talk about those things, let me observe that there seem to be an awfully large number of things that are still up in the air considering that the first public play test is in just over a week.
The aspect that is potentially most problematic from an old school point of view is the treatment of cantrips (basically, at-will powers by another name). This is because having unlimited uses is fundamentally at odds with the resource management that is core to low-level traditional D&D. It is possible to make this work, but the cantrip powers have to be chosen very carefully. For example, there can be no light cantrip. I’m not 100% opposed to something like an at-will attack power (for example, see this post about cantrip scrolls) but an at-will attack does fight against the perception magic as strange and special. This ultimately comes down to a setting question: high magic or low magic?
Traditionally, D&D magic is reliable (with the possible exception of spell interruption). Dangerous magic (spell fumbles, insanity systems, etc) is flavorful and fits much fantasy literature and mythology, but can be hard to model for a game about problem solving. I think both of these styles can work well, but I’m not sure how they can coexist. It seems like a decision needs to be made here. Maybe dangerous magic should be saved for another class such as the warlock?
There are a few points that I am fully on board with. For example, I have never much liked enhancement spells (stoneskin, haste, etc) because in my experience they lead to excessive preparation before any possible conflict. The casting of such spells does not represent interesting strategic or tactical planning. It’s just finding a way to stack bonuses. Once these bonus spells start to feel mandatory, something is wrong.
I like what Mr. Mearls has to say about the creative use of spells (for example, using grease to help a rogue escape). This comes back to the idea of associated or disassociated mechanics and fluff as crunch. That is, in the design process does the effect of the spell come first or the meaning of the spell come first? (Tangentially, I usually hate the terms fluff and crunch, but that roles/rules post also implicitly shows why those words can be so harmful to game design.)
The other possibility that I like is a decrease in the number of spell slots, especially for higher level wizards. Just in terms of practicality, tracking all those spells and deciding which to prepare per adventure is a lot of work. A smaller number of slots makes consumable magic items more valuable as well. Also, having too many slots doesn’t fit either Vancian or mythological literature very well; magic is more often portrayed as more limited. Having many spell slots also doesn’t fit much recent fantasy (like the One Power of The Wheel of Time or the Force in Star Wars). Those types of magic would probably be better served by a mana point system (which I have no problem with as a supplemental class, just not for the core wizard).