Kamikaze Midget said:3e obviously fits this definition, where "Y" is "All earlier editions." 3e sells better than earlier editions, it supports a stronger market, and it appeals to a wider audience.
Err - although I'm with you with "3e sells better than 2e", I'm not so sure about "3e sells better than 1e". Charles Ryan (how I miss him!) indicated that 3e was doing really, really well ("best ever"!), but we don't really have the figures to corroborate that.
So I ask: What is there worth saving in the old editions? What experience can they provide that 3e cannot? What ancient wisdom has 3e discarded in it's hunger for money and high schoolers?
That's an excellent question. Though I'm tempted to say "Gygaxian prose", I can't really say that it universally improved the 1e rulebooks. It was great at giving the grand picture of the game, but when it came to explaining the actual rules, it could get pretty lousy. (See initiative).
When I look back to my early years in D&D, the rulebook that I really admired was the Moldvay Basic D&D book. It did everything a basic rulebook should do. It provided an excellent game. And, most importantly, the rules were clear. AD&D, in comparison, was pretty lousy. What AD&D had going for it over Basic D&D was choices. You could play gnomes! You could play illusionists! You could play elves that weren't fighter/magic-users!
Where AD&D really defined itself to me was in the adventures. They are spectacular. (There's a certain goofiness about Basic D&D adventures that doesn't really appeal).
Choices, they are an essential part of D&D for me, to keep the game entertaining; the variety in approach and challenge. What 3e does better than previous editions is provide the base from which these options can be built without causing more and more rules conflicts. (See monk and surprise).
Cheers!