My experience has been that it can be a problem...and how difficult a problem varies from group to group and game to game, but it is a problem at every level. I had a game that went from 1st level in 3e to 28th level in 3.5 over the course of 6 years or so. My players are all gamers who've been playing for decades. They are not inexperienced. They are not foolish and rarely unprepared. Several of them recognize this as a potential problem, as do I. For some, clearly it's not an issue. It's not a game-breaking issue, either...having been on both sides of the equation, I can appreciate the problems both a DM and player face in this regard. Which is why I look forward to it.
At varying levels, the spellcaster will be forced with a 'usefulness' choice. Yes, they can contribute. In fact, under 3e, they can contribute with more effectiveness than ever before. But that hardly means there is no room for improvement, IMHO. This choice, however, is often one of diminishing returns. The spellcaster, more than any other class, has to dole out his effectiveness in limited quantities. The inherent problem with this is that every combat he has to decide, "is this the combat where it's worth using my fireball/lightning bolt/disintegrate/etc.?"
That fundamental choice is undoubtedly part of the fun for some players. But it's also a source of consternation and frustration for others. In some cases, the spellcaster's efficacy is tantamount to the party's efficacy. Do you go into the tomb to fight the ghast or the bodak without the cleric ready to counteract it's powers? You might, if the need is great or the time urgent. But if the party has no other way to deal with it's terrible powers, you may choose to wait until he has them. That hurts verisimilitude for a LOT of people.
In many cases, the spellcaster empties his arsenal...or at least the meaningful parts of it...well before the adventure is done. Maintaining that Remove Poison or Alarm spell until the evening is hardly the stuff of excitement, useful as it may be when it's needed. Alternately, the spellcaster feels the need to be extra-conservative...his most powerful spell may save the day....but when to use it? The bard often had this same problem with his bard-song...until Eberron gave him a feat to add four more uses at first level. Suddenly he could be an essential part of the team's strategy and with five daily uses at first level, he made a significant difference in every low-level combat. He didn't suddenly become vastly more powerful...just slightly more powerful and vastly more useful. His secondary song abilities became much more viable, when he knew he could spare them for other uses. What if the wizard and cleric got similar such boons as the backup powers for a bard? Perhaps they could dedicate a spell-slot for a specific pool of effects (bolt, shield, movement, illusion), inherently weak but always useful? There are plenty of possibilities to make a spellcaster feel relevant without making him overpowered.
Complete Mage features the Reserve Feats and they have proven quite popular. The wizard in my game can cast a storm-bolt every round. It's weaker than a melee attack, more limited in range than a ranged weapon, requires tactical manuevering to be effective and is unique for a spellcasters (having electrical damage as a feature) and it requires dedicating a spell (which also determines it's power). This gives the caster a meaningful resource choice all through the day (must memorize a specific type of spell...use the spell or keep the ability?) without significantly increasing his power. To me, this is a win/win.
The suggestion of switching over to a non-spellcaster ability is fine, but in practical situations not all that exciting. At early levels, a caster can maintain with the fighter in some limited way...but only because the numerical differences are slight. Within a couple of levels, this becomes less likely. And as levels grow higher, having plenty of spells doesn't translate into immediate satisfaction. The choice then becomes one of lesser effectiveness: a magic missle, acid arrow, fireball or ice storm is not very effective at 17th level, generally...especially against the kinds of foes you're facing at that level.
3e, in general, discarded the notion that the classes had to be balanced with differing advancement rates or that classes were meant to underperform at differing points in their careers. It sounds like 4e is taking that notion a step further, by removing the limitation further still.
To me, the question isn't whether or not a DM can compensate for these issues...it's whether I, as both DM and player, WANT to be forced to accommodate what I consider to be an artificial construct within the game if there's a perceived benefit to doing it differently. Whatever maximizes the game's fun is, for me, a benefit. What works for me may not work for others....and when the time comes, I expect people to vote with their dollars. But based on past performance, I'm willing to give the WotC designers the benefit of the doubt. I liked 1e, in it's time. I left it based on issues like these. 3e brought me back. 3.5 did nothing to hurt the game, IMHO, and I expect 4e to enhance my game. If I'm wrong, the market will prove it and 3.5 will remain the champ. Either way, I see it as a win/win.
