The DnD Next public playtest is pretty much over it seems. Wizards is going to put out one final playtest package in September and then work in house on polish. Mike Mearls blogged about what the design team learned from the playtest and some of the broad goals they settled on after getting all kinds of feedback. I am going to talk about those ideas and goals.
-You like simplicity. You want to jump into the game quickly, create characters, monsters, NPCs, and adventures with a minimum of fuss, and get down to the business of playing D&D.
This is true to an extent. People want to be able to build characters and design encounters quickly. Unfortunately Next has really missed the mark on this one. In 4th edition there were hundreds of feats to choose from and some were absurdly powerful while others were almost entirely useless. To be good you needed to spend hours poring over the lists trying to make sure you didn't miss "do 20% more damage". That was terrible. However, in Next as it is they simply stripped out all the choices. The middle ground where they provide a small number of meaningful choices is actually where they need to be. If you provide players with six sets of simple choices (race, class, class subspec, skills, background, stat layout) there will be substantial variety but nobody will feel overwhelmed. Each individual choice is bounded to a small number of easily understood options.
-You like that every class has the potential to contribute in most situations, but you're OK with some classes being better at certain things if that fits the class's image.
I think this is generally true. People do want fighters to be able to do useful things out of combat but they expect rogues to be better at sneaking and stealing and such. Unfortunately with skills being removed from Next those classes that are grounded in physics seriously get shafted and those with magic will dominate non combat situations. There simply isn't any way to build a fighter that can do anything significant outside of combat (mechanically, of course, since roleplaying does not rely on class) but spellcasters can't avoid having all kinds of awesome tricks.
People do want classes to have niches but combat shouldn't be one of those niches. Certain classes being garbage when they are in a fight always leads to problems even if they can shine outside of it. Next needs to find a way for martials to have exciting contributions outside of combat and be reasonably balanced in combat. We really don't need another edition where fighters are tough and then go get everyone else sandwiches in between brawls. Figure out some way to let thugs do cool stuff outside of combat whether it be class based powers, skills, rituals, or something new.
The remaining points are mostly trivial. Mearls points out that people don't want to spend their time fiddling with math (very true), that players aren't looking for a remake of an old edition but rather something that works well (not exactly true, but certainly should be the goal), and people want flexibility in rules. Flexibility sounds like a strange way to spell confusing and vague but honestly I believe him when he says people want this. 4th edition was very precise and people didn't like it as much as the random and inexact 3rd edition. I think it is possible to make rules both precise and fun but it is certainly a challenge.