Some thoughts in reaction to the discussion of balance on Green's site this past week:
It's acceptable for one character template (class, skillset, etc) to be much more popular than another, provided the gameplay does not demand all templates to be present to conquer most encounters. Each template only has to be popular enough to justify its design cost. So half as many players might play necromancers as play knights, but that's alright if enough players enjoy necromancers to offset their design cost. It also must be acknowledged that templates can improve a player's gameplay without that player ever choosing to play that template himself. Just because one loves playing as a barbarian doesn't mean one doesn't appreciate the presence of clerics in the game.
HUES OR COLORS?
Players can vary more like colors than hues. Templates do not have to be equally complex or equally difficult/easy, if the game is designed so. That's a restriction when every template is expected to engage in PvP, but the mere presence of trade professions in MMOs demonstrates this is not a universal expectation. The main factor that limits dissimilarity is player expectations, including in regards to fairness. Even though my fellow player is contributing as much to our group as me, should he receive an equal reward if his gameplay his simpler and easier? On the one hand, it's possible to train players to think "yes".
But it's also possible to reward more complicated gameplay with a more complicated reward system. Should all characters be bound to the same system of punishments and rewards? Perhaps not. Thinking laterally, we should consider the possibility of separate and incongruent avenues of gameplay contributing to a great whole.
In the real world, countless systems which are seamingly separate combine to form larger systems and, ultimately, a unified whole. Gravity, the shifting of Earth's mantle and crust, and the sun's blasting rays are separate systems of very different designs that combine to create the weather system of Earth. Movements in the arts and philosophy combine with movements in scientific research to create culture systems. The same principle may be applied to MMO gameplay. They can have vastly different goals and process, even largely ignore each other, while still combining into a unified world.
Of course, that begs the question: Can players disagree about the ultimate nature and purpose of that world, like people do about the real world?
It's acceptable for one template to be less popular among groups than others if the game is intentionally designed so. Grouping doesn't have to be a core, universal element of gameplay in an MMO. As I've pointed out before, there are a number of reasons solo players are often attracted to this genre, including the expansive worlds (which allow months-worth of exploration) and greater illusions of a living world (other players needn't interact directly with the solo player to improve his or her gameplay). Furthermore, the presence of solo players can improve the gameplay of group-oriented players. How many popular fantasies present one or more "lone wolf" characters as pivotal story figures? Games can be designed so that both types of players benefit more obviously from the presence of the other.
Look at it this way, an MMO is a virtual society. In real-world societies, loners benefit the masses without much social interaction, such as a web programmer or architect, on one extreme, and geniuses like Beethoven and Da Vinci, on the other. You don't have to interact with people directly to be a vital member of their society. As players are increasingly able to impact their gameworlds, the importance of soloers in virtual societies will become more obvious.
All of this is meant less as a critique of current MMOs than as recognition of some viable alternative methods.