DnD Next put out their next playtest package that gives the testers actual information about how to build a character up to level five (the old version just gave us premade characters) and much more refined rules. I also got to listen to an interesting podcast where the Penny Arcade guys did a long interview with Mike Mearls, who is a bigwig in charge of Next. In all that there were some really great things and some really questionable things; the fighter mechanics look fantastic and the rogue mechanics look questionable.
In the first iteration the fighter was a simple autoattack machine. Nobody much liked that and the 'no thinking allowed' fighter got scrapped. The new iteration gives the fighter extra dice to spend on things. They start out with 1d6 and gradually get more and bigger dice up to 4d12 at level ten. Initially those dice can be spent adding to the damage of attacks or reducing damage taken from attacks but each fighter will have other abilities they can spend dice to use. One example is riposte - when missed by an attack, the fighter can spend three dice to make a regular attack against the enemy that missed. This seems like a fantastic mechanic because someone who wants simplicity can simply spend dice to do more damage but if you have an interest in strategy and flexibility you can play a fighter with lots of options.
Rogues get Sneak Attack in combat and they get several mechanics to make them the masters of skills outside of combat. The most interesting is their ability to automatically get at least a ten on any skill roll. After rolling their d20 they simply raise the result to ten if it had been lower. This is problematic because if the rogue can make the check on a ten they have a 100% chance of success but if they need an eleven then their chance for success is 50% and there is nothing in between! Rogues also get many more skills and higher skill values than other characters. This is all fine and good if you want to be the skill character but I question how much fun it is going to be. If anyone else in the party has any chance at a roll the rogue probably succeeds automatically and given that skills use no limited resources I don't think there will be a lot of intensity in skill based situations. Maybe I am just not used to skill systems where success is assured but it doesn't seem right.
Back when I was designing games with Iolo, Sthenno, Hobo and Full Throttle we eventually came to the conclusion that having some characters be awesome in combat and some awesome outside of combat was not a good balancing mechanism. Not in a heroic combat based style, anyhow. Next seems to be aiming to balance things this way and I am hesitant about it in general and don't like the rogue mechanics in particular. The other thing that I really want to know is how the designers intend to balance high level spell casters. Sure, you can give fighters so much damage that they are good in fights next to a wizard but if wizards can fly, go invisible, teleport, scry, etc. then eventually the wizard just beats every challenge by themselves. More than that the wizard completely eclipses the rogue because every rogue skill gets trumped by a spell that does the same thing but better. If you want to build a system that goes from level one to level twenty you first have to sort out how you are going to build a wizard with a classic feel that lacks the classic overpowered.