... And 5E will be a golden opportunity to re-unite the fan-base.
"One game to find them,
One game to bind them..."
I do not beliee such a game exists, that it existed at all was a heppenstance of history and circumstance and sundered elements will never again unite in the way they once did. Unless they invent the holodeck.
--making the game modular so that folks can play different versions of the same game/edition and still be "official."
That could be done in 4e as it stands with some effort on Wizards part. If they want to make the effort.
I also wouldn't be surprised if we see a 4E Revised before 5E. In other words, errata-ed, re-formatted, and revised versions of the core three, fully compatible with Essentials. Maybe that's why Sword & Spell is on hold.
I would not either but they could put that out with 5e announced. I think also we will see some more experimental books in the Book of 9 Swords style before 5e will be announced, probably something with classless D&D.
However, I thing 3 years is too soon, I thin 5 years is the minimum and here is why: It will take 3 years to create a fully integratred suite of DDI tools and at least 2 years of observation to see how people use them.
In that 2 years they can develop the new tools to support 5e. When that toolset is nearly ready then they announce 5e for a years time.
I think that they are now capacity constrained in software develoopment and while the can output stuff for the current game and develope a new game in 3 years I am not convinced that they can develope new tools for the existing game and tools for the new game at the same time. Hell, it is not obvious right now that they can complete the development process for the tools in the existing game right now.
A 5e with no eTools having gone as far as they have gone with etools for 4e is pretty much unthinkable, and would have very little credibility.
Especially as they have demonstrated the interest in the marketplace. Failure to deliver would certainly open the door for others to succed where they have failed.