What I'd like to see:
1. Expand the base number of levels from 20 to 100 by spreading the power gain of the existing levels 1-20 over the first 80 levels, weighted in favor of levels 5-12 (which would be appx. levels 10-55). Play would generally be expected to go from level 5-10 to level 60, about like a typical console RPG, but the rules would be designed to go up to 100. Bonuses exceeding rolls would be a feature, not a bug, IMO.
Chances of happening: essentially none.
2. Rebuild the existing classes using a point buy system. They should all end up with the same number of points at their 'capstone' level, and ideally within 5% of their points at each individual level (perhaps 10% at 1st level). Include these rebuilt classes in the PHB, and the system used to make them in the DMG. Make it extremely clear that building new classes is the DM's job (and WotC's, of course; got to sell them splatbooks

), not the player's.
Chances of happening: poor.
3. Do the same thing for races as was done for classes. I think this is actually a bit more likely, because there seems to be more discontent with the LA system and the 'balance' of the core races than with the class system, especially in the core. See: dwarves vs. half-orcs/half-elves. Possibly make half-@@ a feat or template (in general, a LA+0 template) that can be applied to multiple creatures.
Chances of happening: decent.
4. Do the same thing for spells as was done for races and classes. The point-buy system for spells should be effects-oriented like the in the HERO system. Of course, include the prebuilt standards in the PHB and the system in the DMG. Spell research should use these guidelines, of course.
Chances of happening: very poor.
5. Remove the arbitrary need for a 'healer' and rebalance the cleric and druid accordingly. Healing should be a bonus, not a requirement, just like any other ability. On the flip side, make healing spells *much* stronger. WotC of all companies should have known how efficient life gain has to be to be worthwhile. An action taken to cast a heal spell (which
delays your losing the battle) is an action that could have been taken to cast an offensive spell or attack (which causes you to
win the battle).
Chances of happening: good.
6. In fact, remove all forced party roles. Design the game in a way that makes a combination of classes interesting and useful, rather than leaving in artifacts (like the rogue's trapfinding ability or the need for healing) that force a party to include certain classes or suffer gravely for it. The 'magic-user' and 'fighting-man' roles don't need to be forced; the 'cleric' and 'thief' roles either shouldn't (cleric) or don't (thief/rogue). Electronic games have found ways to make using a mix of characters beneficial and fun despite having fewer useful roles, so this isn't impossible by any means.
Chances of happening: good.
7. Remove Vancian magic and x/day abilities in favor of either x/encounter, SP/MP (which regenerate over time), unlimited but weaker or unlimited but with drawbacks spellcasting and abilities. x/encounter seems to work the best.
Chances of happening: poor for spells, very good for x/day abilities.
8. Make counterspelling effective. This actually ties into (4.) - effective counterspelling should be the major check on magic's power, not limiting it on a per day basis. Choosing to counter a spell should be a viable choice and should be made at instant speed, as in Magic the Gathering. The mechanics (swift action casting times) are already in place. Like healing, counterspelling must be very effective if it consumes resources or, worse, actions, since it only prevents/delays losing rather than winning in its own right.
Chances of happening: decent.
9. Include action points in the core rules. Then expand them to at LEAST the level of M&M hero points, if not Cinematic Unisystem drama points. In short, make them the mechanism by which the players, normally only in control of their PCs, exert narrative control. One action point may only give a bonus to a die roll, two may count as a minor success, but a sufficient number should give the player Wushu-level control over events. Like all abilities, action points should not refresh on a per day basis. I like Mutants & Masterminds' refresh system, wherein hero points are given out when the DM fudges for a villain in the way a hero point would be used or brings a hero's complication into play.
Chances of happening: very good for inclusion, poor for expansion, decent for refresh rate/system.
10. Clean up the creature types, with a heavier emphasis on subtypes. Suggested types: Beast, Humanoid, Spirit. Then make a skeleton Humanoid (Undead), an angel Humanoid (Outsider, Good), an air elemental Spirit (Elemental, Air) and a dragon Beast (Dragon). Don't be squeamish about a creature having 4-5 subtypes if needed - they are far easier to key abilities off of and should be used extensively. Creature type cleanup has been ongoing in Magic for years, and D&D could use the same thing if not more extreme.
Chances of happening: decent.
11. Clean up the social skills. Make them clearer, possibly a cleaned-up version of Rich Burlew's diplomacy variant, and include good DM advice for using them - as well as for the option of reducing their role. Discuss the main methods: roll first then roleplay results, rolling only, roleplaying only, and roleplay for bonuses/penalties to the roll.
Chances of happening: good.
12. In general, make skills only a few levels behind spells. A 20th level rogue's Hide and Move Silently (possibly combined into Sneak) should be better than a 3rd level wizard's
invisibility and
silence, but the wizard's spells should be better than a 3rd level rogue's skills. Similarly, Diplomacy and Intimidate should lag a few levels behind
suggestion and
dominate person - but their ability to influence actions should eventually be just as powerful.
Chances of happening: poor.
13. Reduce the importance of equipment. Make an equipment focused character an option, not a requirement. Ideally, make the game effects of equipment a class feature. King Arthur's weapon is important to his mystique; Lancelot's is not. Aragorn's weapon is; Conan's is not. Cloud's is; Terra's is not. Reduce the reliance on assorted glowy knick-knacks like Boots of Striding and Springing and Bracers of Armor.
Chances of happening: poor.