So, with D&D “fourth edition”, Wizards is planning on putting out a new Players Handbook, Dungeon Masters Guide, and Monster Manual each year. These will be expansions rather than replacements. They’re considering them, however, “core” rather than supplements.
The distinction is a bit subtle. They typically avoid using a lot of stuff from supplements in adventures, preferring to stick to the core rules. Also, calling these expansions “core” make it easier for them to choose to delay certain things—e.g. gnomes, half-orcs, and druids—that they would normally consider something that should be “core”.
As I said, the distinction is subtle.
I actually think, to some extent, this isn’t a bad idea. It is quite a task to develop a game as involved and with the scope of Wizards’ brand of D&D. It seems daunting and risky to try to fully develop and playtest it before releasing anything.
They did test some parts of “fourth edition” as “third edition” supplements and in the Saga edition of their Star Wars RPG. But I’ll put that aside for now.
I think, however, that I'd go a more classic D&D BECMI→RC path.
The quick explanation: Classic D&D was released in a series of boxed sets: Beginner, Expert, Companion, Master, & Immortal. These were later compiled into a single Rules Cyclopedia.
The first year books would’ve been the Heroic PHB, Heroic DMG, and Heroic MM. They would cover only levels 1–10. (Which “fourth edition” calls the “heroic tier”.) In exchange, they would’ve covered more classes and races.
The second year would see the Paragon versions of the three core books. The third year, the Epic versions.
Then, once the game is fully developed, the fourth year would see all three tiers compiled into the Complete PHB, Complete DMG, & Complete MM triple along with errata and any other fixes that developed over the years.
I think it would be easier to develop and test more classes/races over a single tier than fewer classes/races over all three tiers.