1) On the one hand we have a desire to have a simple, streamlined game to appeal to "old school" aesthetics and new or casual players. On the other hand we have a desire for a more complicated, robust game to appeal to 3E / 4E -type fans. Sounds like we could set up a D&D and AD&D dynamic: two separate games with some common ground. How did this work out the first time when there was both D&D and AD&D? Should it be implemented again?
2) I for one am a big fan of the Castle Ravenloft Board Game (and, I assume, its kin), especially the stripped down 4E mechanics. You have a few character choices, a few power choices, every PC feels distinct from the others, and a character sheet could easily and clearly fit on one sheet of paper. Now the board game rules are fine for what they are, but I think with very little effort the player and monster mechanics could be the foundation for a very basic D&D game--a modern take on an old school-type game. Heck, even the tile system for measurements (as opposed to 1" squares) points the way to a more abstract version of 4E without the need for miniatures or grids (although I think miniatures and maps are great for new, young, or casual players). I think this could be the basis for the "baseline" of D&D in the future.
3) I think I like Mike's modular ideas as follows: Basic D&D is a separate game, stripped down to the studs, and is the baseline D&D. AD&D is a separate game, using the same studs but building on them to the extent that we reach the complexity level of 3E or 4E. I'm all for sub-systems (complex ship-to-ship combat or mass combat or kingdom management rules), optional rules (skill challenges or gaining followers or creating custom spells), or variant rules (different item creation rules or treasure distribution methods), but those are things that should come up in Dragon or Dungeon or in individual published adventures, campaign settings, or supplements. There doesn't need to be an explicit integration or explicit choice of those alternate rules in each campaign. This is how it used to be. There are the general rules (which are complex), and you can tack on something else when it is appropriate. I think I agree with the heart of what Mike is saying--appealing to more types of players and offering robust options--but I'm put off by the structure of his ideas. I don't want every component of D&D to be a mix-and-match, buffet-style affair--we as players will then cease to be playing the same game in a sense. Something about what Mike is saying seems too mechanical and intrusive to me. And whatever D&D becomes, it has to be it's own creature. It might appeal to fans of 1E, 2E, 3E, or 4E, but it cannot become all of those editions simultaneously, or it will collapse under its own weight. That being said, I think a company can create and maintain two different versions of the same game--and only two if one is very simple--and keep quality up.