I'm not sure whether this will strike people as staggering or just staggeringly obvious, but I'll just come out and say my thoughts on the matter:
Min-maxing is gameplay.
Just as coming up with a character's personality and background is a creative endeavor in the same sense as roleplaying, the act of character optimization is as much a part of "playing" D&D, Shadowrun, etc. as the combat is.
And when I look at the process of min-maxing in 3.X as a type of gameplay, it lets me come to a few conclusions.
-Character optimization needs to be easy. Or rather, foolproof- damn hard to do wrong. A beginning player should ideally be able to create a viable character, one that won't automatically be overshadowed by the creation of an experienced min-maxer so long as said min-maxer isn't deliberately trying to break the system (i.e. exploiting poor wordings and other "bugs" the game designers failed to catch).
-Character optimization needs to be less important. The question the game designer has to face is: How much importance do they want to lend to the act of character optimization? To what degree should character optimization be the key to victory? I would say 3.5 D&D is an example of a game that goes too far- beyond a spellcaster's ability to choose which spells to prepare and come up with creative uses for them, the *only* major key to victory is a characters underlying bonuses paired with sets of maneuvers that were planned out before the game ever began. And one thing that will be necessary to fix that:
-There need to be more significant in-game tactical decisions. It's only natural; for decisions made *prior* to the game to be less important the mechanics will have to make player decisions *during* the game.
Having had a chance to look over 4th Edition, it's interesting to see how it accomplishes these goals (I'll bet they didn't phrase them the same way I did, but it seems like they worked along similar lines). Character optimization is made easy by presenting you with a pair of probably-can't-go-wrong choices (race and class) and then having ensuing choices be based on those two initial ones. (You're a dwarven paladin? Well, here are the racial feats, here are the class feats, here are the paragon paths, and if you want some of what he's having there's a feat chain that will provide it in a way designed to not screw you over optimization-wise.)
Powers, meanwhile, are used to place an emphasis on "in-game" play via abilities whose uses are flavorful, conditional and/or otherwise limited in use- you have to pay attention and plan out when and where you're going to use them. The limited number of powers works from a gameplay perspective. . .it's just that the kind of gameplay it provides doesn't quite jive with me.
We're now dealing with two major subjects here- "in-game" gameplay and character optimization. The next two posts will delve into how these aspects of 3.5 were deconstructed while developing d20 Rethought. It'll take alot of playtesting before I can offer any definitive opinion with regards to the results of my approaches vs. those taken by the 4th Edition design team, but it's interesting to compare and contrast what's been done.