When we designed 3rd Edition D&D, people around Wizards of the Coast joked about the "lessons" we could learn from Magic: The Gathering, like making the rulebooks -- or the rules themselves -- collectible. ("Darn, I got another Cleave, I'm still looking for the ultra-rare Great Cleave.") But, in fact, we did take some cues from Magic. For example, Magic uses templating to great effect, and now D&D does too. (To be clear, in this instance, I don't mean templates like "half-dragon," so much as I mean the templating categories such as "fire spells" and "cold-using creatures," then setting up rules for how they interact, so that ever contradictory rules for those things don't arise again, as they did in previous editions.) Magic also has a concept of "Timmy cards." These are cards that look cool, but aren't actually that great in the game. The purpose of such cards is to reward people for really mastering the game, and making players feel smart when they've figured out that one card is better than the other. While D&D doesn't exactly do that, it is true that certain game choices are deliberately better than others.  Toughness, for example, has its uses, but in most cases it's not the best choice of feat. If you can use martial weapons, a longsword is better than many other one-handed weapons. And so on -- there are many other, far more intricate examples. (Arguably, this kind of thing has always existed in D&D. Mostly, we just made sure that we didn't design it away -- we wanted to reward mastery of the game.)  There's a third concept that we took from Magic-style rules design, though. Only with six years of hindsight do I call the concept "Ivory Tower Game Design." (Perhaps a bit of misnomer, but it's got a ring to it.) This is the approach we took in 3rd Edition: basically just laying out the rules without a lot of advice or help. This strategy relates tangentially to the second point above. The idea here is that the game just gives the rules, and players figure out the ins and outs for themselves -- players are rewarded for achieving mastery of the rules and making good choices rather than poor ones.  Perhaps as is obvious from the name I've coined for this rules writing style, I no longer think this is entirely a good idea. I was just reading a passage from a recent book, and I found it rather obtuse. But it wasn't the writer's fault. He was just following the lead the core books offered him. Nevertheless, the whole thing would have been much better if the writer had just broken through the barrier this kind of design sets up between designer and player and just told the reader what the heck he was talking about.  To continue to use the simplistic example above, the Toughness feat could have been written to make it clear that it was for 1st-level elf wizards (where it is likely to give them a 100 percent increase in hit points). It's also handy when you know you're playing a one-shot session with 1st-level characters, like at a convention (you sure don't want to take item creation feats in such an instance, for example).
Ivory Tower Game Design requires a two-step process on the part of the reader. You read the rule, and then you think about how it fits in with the rest of the game. There's a moment of understanding, and then a moment of comprehension. That's not a terrible thing, but neither is just providing the reader with both steps, at least some of the time.
While there's something to be said for just giving gamers the rules to do with as they please, there's just as much to be said for simply giving it to the reader straight in a more honest, conversational approach. Perhaps that's what the upcoming D&D for Dummies book will be. I hope so.