One thing that's striking to me is the emphasis on removing options, even when it's presented as providing them. Players always had the option of starting characters at 4th level or retiring them at 14th. When the "kernel" was relatively simple and modular, players could modify the game in all sorts of ways -- including, of course, with "official" supplements. Plenty of folks have altered things along 4E lines, by their own choice.
Now, it's supposedly the company's job to prevent people from playing old-style 1st- or 21st-level characters. Set aside number-crunching mechanical details and compare this basic attitude to the quite different one expressed in the seminal work published in 1974.
From design to marketing, there's a question either lurking or looming: whose game is it?
Arneson and Gygax designed the original. Gygax designed AD&D, and, whatever his position in the company, D&D remained "Gary's game" in the eyes of the public. Holmes, Moldvay, Marsh, Cook and Mentzer were pretty clearly working not so much on their games as on new presentations of the game.
Then new ownership came in, and Gygax was out. TSR was not in WotC's league as far as interest in market research, but the firm solicited a lot of player input as to what the second edition of AD&D should be like. David "Zeb" Cook was not in the position of creative control that Gygax had enjoyed; 2E was not really "his" (or anyone's) personal expression. It was the first truly "corporate" version of D&D.
TSR owned the trademark, but in a sense the fans owned the definition of the game. That was but a very tenuous sense, to be sure, and while I pretty much missed out on the "edition war" of the 1990s that was in part because I had already lost interest in new releases from TSR. They were in a sense slapping older-edition players with the "let us tell you the proper way to have fun" line.
The origin may be a chicken-or-the-egg question, but there was a significant departure -- and a lot of players were on board with the company. The latter could hardly have kept churning out "splat books" if the former were not buying. Eventually, the supplements would transform the game pretty radically for those who used the whole kit and caboodle.
The core rules, though, remained pretty darned recognizable. The "no difference" line was baloney (and the "improved" one just a matter of opinion) even then, and the objective differences alone are plenty to warrant a preference for 1E or 2E. However, enough fundamental matters remained the same that one could without too much trouble use material from 1996 (assuming one found some worth using) with the rules-set of 1976 -- or vice-versa. There was a foundation of common concepts expressed in mostly the same language.
So, for 26 years (1974-2000) "D&D" was at its heart a reference not merely to a trademark but to a game about as objectively definable as many others. It was from the start meant to be just a start, a framework to be reshaped and built upon from campaign to campaign. That was part of the tradition.
With 3E, tradition lost its primacy and became just one consideration among many. Everything was negotiable -- among the new designers. They were designing their (and the company's) game.
Already with 2E, the process had been turned upside-down. It was not a matter of someone creating something just for fun and then finding that it had a market. There was no Muse (What is the name of the patron Muse of RPGs, anyway?) demanding that it be written; the demand was from Corporate, for a product. Tradition was respected because it was thought to sell.
Wizards held onto that idea rather loosely. I think part of why is that the designers were a new generation, their image of what was definitive of D&D shaped by TSR's 2E-era marketing. There's a feedback loop when folks who "drank the Kool-Aid" are put in charge of making it. That's also going on in the market. The folks still buying the product make up a demographic self-selected for not considering the latest stuff crap. How many former customers have been lost? How many potential customers might have bought the older model?
Maybe WotC's market research breaks that loop. The bottom line, though, is not about fidelity to "D&D" as anything but a trademark. 3E established that one thing that sold was treating D&D players not as fellow hobbyists but as consumers. The "pros" are not just guys lucky enough to get paid for writing game material; they Know Better than You. And that's a good thing, because supposedly writing D&D stuff has turned into something as expertise-dependent as designing computer operating systems or interplanetary rockets. Who designs "game engines" if not engineers?
So, after the terrible Rust Monster Disaster of 2006 (Oh, the humanity!) they swore "never again." Too many 3E games were crashing and burning, endangering not only the players but hapless bystanders. How to prevent such tragedies due to human error? By designing the system to lock operators out of the loop at critical decision points. No more "patches" -- rewrite it from scratch.