I like story and challenge. When I play a new MMO, I'm the guy that likes to read the quests and the guy that turns down dungeon run-throughs by high level guildies in favor of a PUG with 5 players that have never seen the inside of the dungeon. Unfortunately, in recent experiences, I find many gamers are now on the "rush to max level." They blow through quest content without reading the story and try to finish dungeons without speaking to their PUGs. The sole goal is to reach max level to get to the end-game.
The above mentality makes MMOs far less enjoyable for me, as it is difficult to find leveling companions (since I go at a slower pace), difficult to find PUGs (most competent players are in guilds and getting run-throughs), and there is little point in playing solo, as the single player content tends to not be challenging.
It also creates a problem for developers, as they spend massive amounts of time and money developing non-end game content to try to make leveling enjoyable, when most players will just skip through this to rush max-level and then complain about the lack of end-game content.
The Game Idea
The idea is simple, instead of designing most of the world for leveling with a few high-end areas and dungeons, I'd like to see a game designed with a short progression tutorial followed by a world of end-game content. The starter world would be a guided instanced tutorial or a newbie starting zone to which high-level players cannot return. Instead of being a quick 10-20 minute tutorial that thrusts you into the open world at level 5-6 with 3-4 skills, however, it would be a 10-15 hour adventure that finishes with you at the max level of 1 and gives you every skill you'll have throughout the game.
The starter area could actually be a part of player development. Let's imagine, for example, there is a quest where you and a NPC-ally are tasked with retrieving a stolen item from a certain bandit leader. In a game with class-specific skills, each class will accomplish its task a different way designed to teach its skill. A tank might challenge the bandit leader to a duel and be required to use a new block skill and a counter attack to win. A melee dps might charge in, activate a temporary attack bonus skill to quickly kill the leader, and escape as the entire camp converges. A rogue will be given a new stealth skill to sneak in and steal the item. An archer or mage will get a long-range skill to attack the enemies from afar, distracting them while the NPC ally steals the item, and a healer will gain a shield/healing spell to keep their ally alive. Alternatively, a game with unrestricted skills will let you choose which of these you want, designing your own class through your in-game choices rather than a pre-game menu.
Depending on the game, this area will be a solo instance or an area only open to beginners. Either way, there will be no high-level PvP players to grief those still leveling.
At the end of the leveling zone, you travel through the Rift, land on the planet, leave the starter island, or whatever is appropriate to the story and enter the main town. Depending on the game, it may conclude with a multiplayer event, such as a group dungeon or world boss, to help you learn to interact with other players.
The Main World
As the starter world will be a small theme-park zone, the entire rest of development can be spent on creating a high-level world where every zone can be challenging. You can visit each zone in any order you desire, however, some zones might be too dangerous to enter alone and others might be too difficult based on your skills.
Progression would be created through completing dungeons, with each dungeon providing a specified bonus. One dungeon boss might drop gear, another might guard a rare metal for smithing, while a third might offer a skill/talent/attribute point. What you get will affect your ability to be successful in other zones. The cold mountainous region might offer enemies that are too difficult to defeat until you first find a weapon with a fire-based elemental attack or unless you are a mage that spent a skill point on your fire-based spell. Another zone might have heavily armored enemies that can only be defeated by magic or with a good armor-penetration weapon, while a third zone might have magic-resistant enemies and can only be navigated in a party with a well-armored tank.
The current end-game progression tends to be run Dungeon/Raid #1 over and over until you have enough gear to run Dungeon/Raid #2 over and over, then go back and run Heroic Dungeon/Raid #1 over and over, and so on and so on. In my world you might run Dungeon #1 for a sword, Dungeon #2 for a shield, and Dungeon #3 for a skill point, but you might also run them as Dungeon #3, then #1, then #2.
Another aspect to prevent the end-game from stagnating would be to create scaling dungeon difficulty. If you choose to complete Dungeons 2-5 then go back to Dungeon #1 for your sword, you might find it too easy. Thus, in instances, the game would scale difficulty to the party. A party that has completed 0 dungeons would spawn a dungeon at level 0, while a party that has cleared 10 dungeons would spawn one at level 10. Some dungeons might have higher minimums so they are too difficult for a party with no upgrades. Others might have secret triggers that make the dungeons more difficult and more rewarding. For example, a player that has completed <Holy Quest> might find that Dungeon #1 suddenly has a higher difficult and yields a magic sword instead of a quality sword as a reward. Dungeon #3 might be filled with werewolves that take animal form on the night of a full moon, making the dungeon more difficult.
Ultimately, the game would still have some linearity and a progression to a final dungeon. However, instead of spending ~80 hours to reach max level so you can run the same 5 dungeons over and over, waiting for developers to release dungeon number 6, you would spend ~15 hours to reach max level and be greeted by a world with 20-30 dungeons for you to explore in almost any order you choose.
It's better for players as they don't have to slog through low-level content to reach a repetitive end-game. It's better for developers as they can spend the bulk of their time designing the portions of the game that the players will see most often, rather than designing low level dungeons and quests that many players may never see. The only people it isn't better for are griefers that like to prey on weak, leveling characters without being exposed to possible defeat. While they can still attack players that haven't completed any dungeons, the differences in stats/gear will be much smaller, so skill will play a larger role in PvP.
Perhaps you should make the other players be the questing in the game. The king that needs something delivered to a thief in X town, are both players in the game. The Mining guild that is looking for an adventurer to clear those goblins out of that cave are other players.
Gaining max level in the game is almost in a way "voted on" by the other players. Hence your interactions is what brings you up, not grinding. How to exactly implement that is up for discussion but it would be the holy grail of MMOs =)