If none of the characters in your 6th level party uses a long bow, don't put a 10th level longbow in your dungeon as treasure. A great way to make sure you give players magic items they'll be excited about is to ask them for wish lists. At the start of each level, have each player write down a list of three to five items that they are intrigued by that are no more than four levels above their own level. You can choose treasure from those lists (making sure to place an item from a different character's list each time), crossing the items off as the characters find them. If characters don't find things on their lists, they can purchase or enchant them when they reach sufficient level. Each set of ten treasure parcels includes one less magic item than there are characters in the party. That's not meant to be unfair, just to make sure that characters gain magic items at a manageable rate. Make sure that over the course of several levels of adventuring, you award items evenly to all the char acters, so that over the course of, say, five levels, every character has acquired four useful and exciting items.
And I thought this was a nice bookend to what I was writing about just 7 days ago.
Having players write down wish lists and then giving them what they want is just one of the many faces of the Give-Your-Players-Orgasms school of GMing. It's precisely the opposite of what should be encouraged: which is players trying to get what they want within the game, facilitated by the GM, and thus being its engine. If one of your players wants some Special Snowflake magic item, he should be trying to find out where one could be obtained, and then going out and obtaining it, in-game. He shouldn't be saying "Pwetty pwease" to the GM and looking at him with wide, innocent brown eyes, and expecting to be rewarded for being a good little boy and going through the motions of beating up some orcs.
I likened the 4e approach to Farmville in the comments on Zak's post. I was mostly being facetious - I recognise that Farmville is infinitely worse in almost every respect. But it attracted this reply [from Zzarchov, who I'm sure won't mind me reprinting it here]:
When I play, if I player says "I would like a box of Arm & Hammer" I think I will strive to ensure they have a 4/5 chance of finding such a thing after an epic and evenly balanced quest.
Which is a joke, but sort of hits the nail on the head. The 4e DMG advice sounds altogether too much like a Skinner box: put the D&D player in the cage; have him perform an "epic and evenly balanced quest"; give him a food pellet. Why do we encourage, and accept, such a reductive approach?
If your player wants "Arm & Hammer", it's his job to go out and get it. He needs to ask the GM if he knows of anywhere that might have Arm & Hammer. He needs to have his PC go out and ask around, and see if he can find somebody who might know. He then needs to get off his arse and go and get it. Isn't it far more fun and interesting that way? Doesn't it engage the player so much more? Doesn't it force him to invest thought and energy into interacting with the game world in an interesting way?
The approach I'm advocating isn't supposed to be some sort of Iron Man, macho response to adolescent wish-fulfillment. That's not my point. My point is that if players are the engine for the game - if they decide on their own goals and then work to achieve them - it makes everything better. It is, frankly, the best way to run D&D. Instead of waiting for the orgasms and food pellets like good little rodents, going through the motions of what the GM has put in place - which is essentially a passive approach, what we surely want is active, proactive players making things easy for the GM, and genuinely getting interested in the setting, by working to achieve their own goals. Isn't it? It's not about being 'hard core'; it's about what works.
I sometimes think that, actually, the real reason why RPGs are in a downward slump is because of the prevalence of the assumption that the GM has to lead the players by the nose, rather than the other way round. That puts pressure on the GM and incentivises passivity on the part of the players. It just isn't as good that way.