Follow-up on the "how many classes" question. Some posters raised 3 classes (warrior, caster, expert) as an alternative. Let's ponder the expert.
In any game that's recognisably D&D, all PCs are adventurers. It can be assumed that they're all quite proficient in many of the routine tasks of their profession. So, should it really take a distinct character class just to be good at things like observation, athletics, exploration, spelunking, etc.?
The answer to the question really depends on what utility skills have in your system.
In stock 3e the tendency is for skills to be primarily passive means of overcoming obstacles as selected by the GM. What is knowledge really good for? For the most part, it's assumed to be good only for learning some clue that the GM thought to put in the situation and assigned a skill and DC to. What is search good for? For the most part, it's assumed to be good only for learning some clue that the GM thought to put in the stiuation and assigned a skill and DC to. If that's all 'skilled guy' does, he doesn't do enough to really justify a role among the heroes. He's a glorified henchmen in that role.
It would be like if under the stock 3e system, spells didn't actually do anything - they didn't actually have a descriptive text - unless the DM mentioned the spell as a solution in the text of the room and described what casting that spell would do. It would be like if Fireball only made a big fire if the DM specifically described something as a possible target of fireball.
Now, if on the other hands skills can be used actively by skill guy to significantly improve his chance regardless of the situation that he finds himself in, then the skill guy might in fact be largely on par with the sword swinger and the spell slinger.
Another thing to consider is whether or not the game master considers the most common challenges faced by heroes to be combat or not. In my game, I typically have more skill checks in a session than attack rolls. In this situation, it can really matter whether or not you have a bit of skill in various areas or whether you are just a big dumb lug whose only at an advantage in a nice flat enclosed arena holding something big and pointy while facing someone similar. If every challenge you face can be simulated by a combat in a large flat enclosed arena, then chances are 'skill guy' is not going to shine very much. What is he going to have to do?
I'll give you some examples:
What if your skill in movement gave you a higher base movement speed?
What if your skill in tactics gave you free actions each turn?
What if your skill in porter allowed you to carry more weight with fewer penalties?
What if you could use your skill in alchemy to brew potions, poisons, and grenades?
What if you could use your skill in planeswalking to blink yourself out of danger?
What if you could use your skill in astrology to cast divinations?
What if you could use your skill in leadership to counteract mind affecting spells?
What if you could use hypnosis to break a charm?
What if you could use heal to actually recover hit points?
And so forth.
Suddenly skill guy is actually useful because he can actually choose to do things. In the edges of D&D there are some wierd cases of actually being able to use a skill meaningfully because a feat lets you use it meaningfully or because there is this class feature that is written up as a skill, but they are a sorta of wierd, usually pretty narrow damage generators, and not terribly well balanced with other skills. It's not that they are bad, at least within the context of the other brokeness available by late 3.5, but they are just sort of tacked on to the system rather than thinking about that tact from the start