Or, you could make strategice resource management meaningful.
I imagine that there are people who will prefer both solutions.
RC
I prefer both, provided that it is a conscious design decision from the very beginning, and then that design is implemented fairly well. Those are bigger caveats than they first appear.
For example, part of this whole original topic as it plays out in 3.* implementation is that gestalt rules were tacked onto 3E instead of built in from the beginning. (Not blaming anyone here, since the goals of gestalt are not the goals of the original design. But changing goals midstream has its own problems.)
If I'm playing 3E at all, I'm playing with some kind of limited gestalt. Take the basic gestalt framework, and then knock off some of the sheer power that it gives. Because there are very few characters that you can't get at least a decent representation of using gestalt, and if I'm playing with D&D Lego (which 3E is), then I want to make detailed characters. All the limits of the fighter become moot* when paired in a gestalt, and the advantages of the wizard (and even moreso the cleric and druid) are muted. (You can get an almost GURPS-like slew of options in 3E gestalt by simply saying that everyone has "side 1" of a gestalt which is always single-classed, and "side 2" of a gestalt which can never take the same class twice in a row.)
*Indeed, the fighter is one of the things that has to be reined in with gestalt, and this is not surprising.
But almost all of the problems that stem from RAW gestalt tie into what has been referenced earlier. You can advocate that the fighter is "merely a fighter". He fights. He gets a few bones thrown his way for characterization, but they are always substandard. Or you can say that the fighter is supposed to be a good way (or at least a decent option) for representing a host of "fighting" characters from fantasy media, and thus the class should contain elements that make that possible when played as a single class.
The problem with 3E is that it tries to split the middle on that issue. It can't be split well. If you go the second choice, and say that every class is supposed to be able to represent something with character depth and/or reskinning, you can do something like 4E. 4E is almost rabidly opposed to multiclassing in some ways.

Or you can do something like 1E. There are many possibilities for making single classes viable, and they will all have strengths and weaknesses depending upon what else you want to favor or discourage.
Or you can go the 3E gestalt route, and say that classes are building blocks. Building blocks, by definition, are often small slices of the whole. Design seriously for gestalt, and you will cut out single classes as viable options. As just one example, the fighter (or fighting alternatives) would be a lot more impressive in such a hypothetical system if the casters didn't get so much base attack and hit points. Caster keep those things to remain viable as single classes.
To a lesser extent, this applies to multiclassing as well, and explains why the fighter is as relatively weak as he is in base 3E. The implementation decisions that keeps the fighter from being a veritable treasure of cherry picking goodness in the original multiclassing are the exact same things that keep him subpar as a single-class. And while you can play around the edges to compensate, if you want, by giving the fighter more abilities at higher level, this is explosed by RAW gestalt rules as a multi-classing kludge rather than a real fix. (Though in the context of straight 3E, probably a very effective kludge, and thus worth considering if you don't intend to redesign a gestalt-centric system from the ground up.)