I'm one of the advocates of the player entitlement that's a big part of the D&D game these days. My agreement with Stephen Radney-MacFarland's "Yes" mantra is complete. I think Scott Kurtz, perhaps channeling Chris Perkins, sums it up well in this PvP strip. The point is that players in my games quickly learn that they can try anything and that they can tell a story about their characters, and we'll all work with any problems that might crop up.
It's not that I didn't play enough old-school games. I just believe that letting my players exercise creativity and interject narrative control is a good thing. In my gaming group, when a mechanical element causes a problem, I find that allowing this freedom pays me back. My players trust me, and they're all savvy enough to see when a rules element isn't working ideally. Sometimes they won't point it out, but if something is interfering with play, we work together to house rule it.
That sort of approach isn't always appropriate, because when I feel motivated to say "no" to a rules element or a player, it usually has little to do with the rules as written. It's often a matter of personal taste. Like all DMs, I'm looking to create a specific experience at the table, and my players have similar tastes and desires, so steering the campaign in a particular direction inevitably improves our experience.
For instance, I don't like magic items that put the responsibility on me, as DM, for constant tracking. I have enough to do. Take the cloak of distortion. It imposes a constant –5 on ranged attacks made from more than 5 squares away. In the best case, I remember the cloak before attacking a PC wearing it, because it's clear the monster can see the cloak's effect. Usually, however, I announce a ranged attack on the PC only to be reminded of the cloak. Then I have to either suck it up or change targets, neither of which is satisfying.
When I create a campaign, I encourage my players to use what they know about the campaign to build their characters. More importantly, I usually ask them to create relationships among the characters. Ultimately, since the D&D game is cooperative, character building is really party building.
Sure, I've played games where I threw the party together. I started one campaign in which the PCs were captured as fodder for an eldritch machine that ensured an aging wizard's life would last decades longer. Add a malfunction that kills the wizard and turns the characters loose, and the first adventure is a romp that requires the PCs to escape from his fortress.
It was fun. But it didn't work as well as campaigns I've run in which the party members have ties that bind. The PCs had no reason to stick together after they survived their escape. The game just works better if the characters have similar goals and interests.
Even when I ask for this, it doesn't always work. Each of my players has a different place on the spectrum of committed versus casual. One of my players treats his character backgrounds with all the serious thought of a dedicated novelist. These elaborate concoctions work only until he comes up with a concept that thwarts another player's desired character choice or he makes his character an island. For example, he might create a character with a hatred of the infernal so intense that he won't tolerate an infernal warlock another player wants to make.
No matter how much good potential a character story has, I have to say "no" when that story steps on the desires of other players. I do so by guiding the discussion about the new character toward compromise and party integration -- toward a solution I can say "yes" to. Having a cohesive party is too important to allow my storyteller player total freedom with some of his lone-wolf ideas. Fortunately for all of us, he never seems so attached to his ideas that we can't come up with a compromise. Even the most dedicated roleplayers usually recognize the importance of the group bond.
Now, if I could only influence the players who have a casual attitude to be a little more like my hardcore storyteller with regard to their character stories. Maybe if they knew I'm more likely to kill PCs who have no important story connections ... but I digress.
In all seriousness, I try to reward the story work my players do for my game. I prefer the hardcore storyteller's details, even the troublesome ones, to nothing. It's easier to trim the fat from detailed character stories than it is to work with a PC who has no apparent motivations or history.
In a roleplaying game, less is not more in this context. Detailed PCs who have detailed connections to one another make for a game that is more immersive and personal.