In a recent Legends & Lore post, Mearls explored some of the thought behind magic item design in the new edition of D&D, particularly with regard to attunement. I've seen this post take some flak in other circles, but I think it has a ton of potential and I'm really excited about it, so I'll be providing a contrasting opinion that is largely in WotC's favor. Mearls also touches on item identification, and that's something of a can of worms, since we have many editions of prior experience with item identification as context.
The article opens with a sort of contrasting pair of ideas: that attunement is for the big, important magic items rather than consumables, items central to a character's portrayal; on the other side, items with attunement have a chance to compel the character's will. Okay, that's three ideas, so let's go through them. Yes, it would normally be the worst idea ever to require attunement in one-shot consumable items, since the PCs manage a mere three attunement slots... but I can imagine an awesome consumable item that has one effect when unattuned and a much greater one when attuned, and enough hurdles to attunement that characters can really only start adventures with one attuned - there's really no way to attune one during the course of an adventure. As with any medium, once you understand not just the rules but why they're a good idea, you can break them creatively.
Second, attunement is for conceptually important items; the article's examples are Raistlin's Staff of Magius and that dark elf's porcelain kitty collection. It's funny, I don't actually know how to get PCs to identify with their items to the degree that these characters from novels are invested in those items. It's good to have that close bond of character and gear, but I can't think of a time I've seen that emerge organically in a tabletop game. Oh, God help you if you try to take items away for anything less than fair market value, but PCs sell them off posthaste if something better falls into their laps - and if it never does, they're disappointed. Oh, sure, you can increase or alter the enchantment to be more desirable, possibly melting down the new item in the process so that you hang onto the old one, and having rules for that kind of thing helps. My point here is that it's an area where gameplay looks nothing like the fiction, and I don't really see that workarounds have succeeded in creating that sense of Signature Magic Item.
Anyway, that's the bad side of attunement - the worst that can be said here is that looking to novels for inspiration doesn't get you all that far. I love the idea of attunement unlocking the item's deeper powers, though: in itself, it is (part of) a solution to the issue. Gee, this +1 sword doesn't seem like much. What's that you say? I just had to attune myself to it by going through a quest-based rigamarole? Don't mind if I do! The next step of this idea is that attunement doesn't have to be a strict binary: you could gain the ability to attune at some kind of second, third, and so on rank to unlock more powers.
But seriously, folks. The attunement rules also have this cap at three attuned items - a rule that is all the more important, I suppose, if attunement is easy. I do like having a limit here, and three is a good number of "most favored" items. It bears a certain resemblance to 4e's weapon-armor-neck slot trifecta, and three is probably a good upper limit on noteworthy items that one character can wield before they all start to run together. It's a soft counter to the 3.x Christmas Tree effect - on the off chance you haven't heard that term, it's the thing in 3.x and Pathfinder where you are decked out in a ton of individually-forgettable minor magic items, and the goal is to fill every magic item slot with goodies. That still could happen in D&D Next - a lot of items have unattuned benefits that are still nice to have - but other design choices have diminished its probability.
At the same time, only being able to really care about three magic items at a time (which this rule does only in a sense) is going to do some odd things in terms of adventure loot. In all prior editions of D&D, but especially 3.x, and in Earthdawn, players usually acquire more treasure than they could feasibly use, between weird consumable items that they'll save for the rainy day that never comes (god knows I play this way - I can't help myself) and the magical gear that the bad guys were wielding. The former is especially true of Arcana Evolved (and it's practically the whole point of Numenera), while the latter is most notably true of 3.x and, presumably, Pathfinder.
Starting sometime around 1996, people doubled down on wanting to see monsters explode into showers of treasure, much of which was useless, or useful only to sell to a merchant. When 3.x came out, it embraced rules-as-world-physics, and sociology for that matter, more deeply than any of its venerable forebears. There were detailed rules for the cash value of gear NPCs ought to have, which had the dual effect of massively slowing down NPC statting and increasing the prevalence of useful-but-forgettable +1 whatevers (by preference, armor and cloaks of resistance, because they had great bang for their buck; rings of protection and amulets of natural armor were close behind). Once you've got one of each of these things for every member of the party, they're just big-ticket items to sell to someone who doesn't mind how you got it.
In brief, then, this set of design choices caused most DMs to accept magic items as more of a commodity than Something Special. Books that came out very late in 3.x's development cycle made a game effort at fixing this; I don't think those ideas penetrated the userbase deeply. The solution is dead simple: embrace magic items with more than one function, by way of the Adding New Abilities rules. Interestingly, this is a key part of apparent magic item design in D&D Next: instead of the tagline-like construction of 3.x items, items are presented with two or three abilities and no clear breakdown. DMs are left to their own devices in terms of balancing new magic items, but the balance of D&D Next currently seems far less exacting than 3.x or 4e, so even nominal balancing errors shouldn't ruin play.
How does this relate to attunement, though? If the adventure publishers and the Rules As Written don't back way the hell off on handing out scads of loot, the magic item trade will be worse than ever, because you only need to check new items against the three items you really care about. If it's not an upgrade (or at least a compelling side-grade) of those, it's as good as got a For Sale sticker on it. The crux of the matter - in my experience as a player and DM who values a rich setting and emergent narrative over most other potential goals - goes back to the dynamics of engagement that I talked about a good while back. Attunement plays directly to the problems of engagement: it is another possible story hook. (...which will be fine as long as the gameplay environment doesn't shift to the point that players regard any attunement that can't be handwaved and forgotten as "the DM being a dick.")
At last I get to the third point of my initial summary, where we see that attunement is covering the mechanics for intelligent weapons as well. It's interesting to see D&D Next blend 4e's artifact concordance concepts with "every magic item has personality" from 13th Age. (Also, Wield still has a few days to go in Kickstarter, if you are a hardcore fan of this concept.) Even among very powerful items, it's supported only intermittently the fiction: while intelligent weapons and other artifacts of will and malice are by no means unknown in popular fantasy, they're far from universal even among significantmagicitems. None of this is to say that attunement items having will and purpose, and the ability to enforce same, is bad - the examples in the article are solid, if heavy-handed.
Harbinger's Advice: The second time you hand out an attunement item that forever binds the wielder into the service of Asmodeus, you will a) have players that never again pick up magic items, much less attune them, and b) likely have fewer friends than you did five minutes earlier. What I'm saying is, use the really drastic stuff judiciously, and if you're not sure whether a use meets that standard, assume the answer is no.
The article explains, more or less, that classic items (like Vecna's nefarious extremities) will be presented with defined attunement rules, while "general" magic items won't. For the sake of supporting best-practices DMing, I hope they'll include a sidebar or random table of attunement effects, for times when you just don't have anything prepared. I, for example, run games off-the-cuff most of the time, so I haven't carefully sculpted each magic item months in advance. I'll probably want to adjust my style a bit, or design a stack of ready-to-go magic items, to handle this. Let's not pretend this is easy - good writing takes work.
Next up is identifying magic items. One of D&D's most lasting legacy spells is identify, and like a lot of D&D's "traditional" spells, its design has always had problems. The game has a dubious-at-best relationship with figuring out magic items without an identify spell, suggesting sometimes that it is possible without so much as hinting at rules for same. With highly lethalcursed items also part of the mix, I am accustomed to players treating the identify spell as the sole means of learning about items and an absolute prerequisite to use. (Let's breeze past the game-running implications of items that kill you if you try to use them - the quick summary is that these are a key example of why players don't trust DMs. Even the principle of these items' existence introduces distrust that does more harm than good to gameplay.)
The effect, then, of this design is that the exciting part of receiving magic loot is always undermined by needing a delay (one hour or eight, depending on edition) and an expensive material component that has no other game function (a pearl worth 100 gp). Oh, and no low-to-mid-level wizard is going to prepare that spell without an item in hand to identify, so you can forget about getting that done today - it becomes a downtime activity for once you've returned to town. One might argue that it prolongs the anticipation and excitement, but I suspect that it waters down the experience by separating it from the initial moment of discovery. This also makes the wizard into the gatekeeper of this information; if by some bad luck your 3.x party has no wizard, the sorcerer is going to tell you where you can shove it when you want her to learn the spell, and hiring an NPC to do it only makes the money sink more egregious. (No one ever played bards in my 3.x groups, so until this moment I did not know that it was on the Bard list.) Having said that, 2e and 3.x identify isn't the worst thing ever, but it could be more interesting, and it could avoid being a bump in the road on the Victory and Cheering Turnpike.
D&D Next does a few different things to improve this situation. Prior to this article, it has been shown to us as a ritual in the playtest documents, which has the excellent effect of making it so the mage doesn't need to prepare it or spend a spell slot to cast it. Since this action otherwise took place in downtime, this is not a real increase in power, except that it frees up a bit of the wizard's downtime for spell research. D&D Next promises to have downtime rules, so this puts the mage back on equal footing there. Also, the D&D Next bard may be a good bit more appealing than its 2e and 3.x cousin, though it doesn't give the 4e bard a run for its money. (All down the line, D&D Next doesn't handle the leader role as interestingly as 4e did.)
This article gives us a bit more depth on that: the identify spell is only for items with attunement properties, and then only if you're feeling a bit dubious about the drawbacks. Otherwise, just sit tight for the duration of a short rest, and you get everything but the attunement features. Even just avoiding an identify for every consumable item is a major quality-of-life improvement. Attuning to an item may or may not reveal all of its properties through use, so identify does still have a place.
I think they're about halfway there toward making identify everything that it should be. Pick one or more of the following:
Let it compete for space among a mage's prepared spells by giving it a variable casting time based on what you're identifying. A magic item's attunement features? One hour. The magic on that door over there? One minute. The statue that you're worried might wake up and crush you? One minute might still be too long... For that matter, a few details about a creature might be worth a one-round casting time in combat, or more details afterward, in exchange for extending the casting time back out to an hour.
Make the component cost commensurate to the use. A pearl (average value 100 gp) balances the mystical scales for a magic item's attunement features. (Because really, if the attunement features are anything less than self-evident, it's a safe bet that the item's creator wanted them to be secret. Maybe the cost of that obscurement is, likewise, a pearl?) What might a fine ruby get you?
Turn it into an encounter, like a Commune with a much lesser spirit. An earlier playtest packet described the spell in just this way, and it became canon for Aurikesh. Only one of the mages in the group ever has ever cast the spell, and when he does, a fire lizard appears out of the flame or from somewhere nearby to talk. It usually answers questions about one or more magic items, and offers limited knowledge of the nearby area.
As a result, the player gets a lot more out of the spell than the text specifies, but it's also a lot more interesting, and the player hasn't abused it. (If I felt he were, the fire lizard would probably start squeezing more treasure out of him.) In Aurikesh, a hundred gold pieces is pretty serious cash, so I'm fine with easing the expenditure.
In conclusion, the attunement rules we've seen so far are light on the system side, while rich and open on the story side. That's about right for my money, though it remains to be seen how rules-y they become over the span of the edition. The design underlying the identification of magic items is fine. It could still be punched up a bit on the spell-design side; D&D remains too married to the form and function of the spells it has always had.