What I am, in fact, is a curious about lots of things, including thoughts of the future, in general, technology and virtual worlds. Those who've read some research and/or SciFi on the issue will surely agree with me here: we are doing a really poor job on the virtual world field. There are issues that were explained, studied and dissected for more than twenty years, that highlight problems that you see even in nowadays successful games, like World of Warcraft. Neal Stephenson described the Metaverse several years from now in his SciFi book "Snow Crash", and while he kept developing his own concepts around it (take, for instance, the peer-to-peer and decentralized networking technology he roughly describes in his Young Lady's Illustrated Primer), we're so far from his Metaverse concept that any exciting new online game has a bad taste in mouth for those that, like me, are craving for something like what he described back then. See, I'm not complaining that the made a game conception and design that nobody implemented, nor I'm really interested in a Snow Crash'y version of a Metaverse (or the future). What I want, really, is something as simple as a game at least as well designed as his Metaverse was.
I talked about it before (sorry, no network at the moment to provide you the link): the biggest problem in virtual worlds developing is that these projects are only financially backed up from for-profit corporations, with only profit in mind. I'm not saying that having profit is bad, what I'm saying is that those who are actively playing in this field are restrained of being creative, and instead are compelled to do a product that will maximize profit while minimizing investment. You have tons of well known and profitable MMORPG's, but they're all pretty similar. While the same formulae sells, why would the gaming industry invest in new ones?
The only thing (at least that I can recall now) that I do not agree with Neal Stephenson's metaverse is the (well explained, in fact) necessity he felt for avatars trespassing each other. I'm sorry, but while it's acceptable that characters in Second Life have the ability to fly, it's unacceptable that a character has no difficulty on going from a street to inside a building, while a river of people are between him and the door. Avatars should be able to "pass though" a sword abandoned on the ground, but not a tree, a wall or another avatar.
The first perceptions achieved by babies is that of the existence of world dynamics: permutable, mutable and ever changing. If you want to create a virtual world, you must, above all, emulate that experience, creating what some thinkers call as a "dynamic landscape". You've seen it before: take as an example the old single-player game "Sim City". While you start with a randomly generated environment, you're allowed (with a certain amount of money) of turning rivers into mountains and vice-versa. It's feasible, for instance, to start with two different and randomly generated landscapes and affect them in a way that one is exactly a copy of the other. Now take this experience no new levels - the multiplayer experience. Imagine a new version of Simcity where two players can play together, via a network connection, collaboratively. In that scenario, they could dispute over what was to be done in a certain corner, and, in fact, spend all of their resources by turning a river into a mountain into a river into a mountain. Now take that new level of experience into the Virtual Worlds realm. Think of Second Life, World of Warcraft, Ultima Online, or whatever your favorite virtual world is. Are you able to do that? Wouldn't it be a great enhancement to the gaming experience?
On the topic of Simcity, it also has a characteristic that unfortunately many MMOG's adopted: the end of the world. See, you can design a virtual world that, instead of being a planet, are, for instance, a turtle shell. Yet, heading always in a certain direction should not lead me to "a wall" or any other limit. While it might be easier to see it in Simcity (where the world is basically a small bidimentional squared matrix) if you think about it you'll notice that most "Virtual Worlds" also fail this test, thus not being really Virtual Worlds.
One of the things that are lately being sold as "new" is the concept of in-game economy. The thing is, this concept is not new - at all, and was even implemented for more than a decade. What's yet to be seen is the mindset on the gaming industry that virtual worlds are only good when they give a collective user experience - which is to say that the gaming experience of a single user must be such that the experience itself only have value by being online. In other words, let's kill the NPC's and have a fully player-driven economy. Don't take economy as "money talking" - I'm here talking about the player role and his effect into the world itself. One of the biggest critics I do on WoW is actually that: if you took out every player of WoW out of there, the world would be entirely unchanged, and if you took half of them, randomly or not, the game experience of those others wouldn't be affected - at all. Players of a MMORPG's are citizens on a Virtual World - and should be able to role play as such. That means that they should be able to create, change and break rules, to act alone or as a community (including creating social synergistic concepts, like governments/polytics or religions). In a scenario as the one described, players naturally create relations, bondings and groups that go beyond the "capacity to play" - avatars tend to group up with those with similar value chains, morality and the concept of right and wrong arises, and "rules" (that can be simply viewed as chaotic organization) emerge.
Now, for a change, let's talk about a classic issue: physics. While this is being sold as a big feature, it should be, instead, taken as an obligatory one. What kind of immersion and embodiement is expected when, in a simulation of an Earth-like world rains do not make rivers rise, clothes get wet and dust turning into mud? Just to tease you off, even the first MUD ever created had both the concept of physics - the most obvious being stamina, day time and seasons.
Also, please stop imposing real-world social restrictions and taboos into the virtual worlds - worlds dynamics should not be affected by external sources. This implies that stuff like we're used to see in Second Life are bad for the world itself - things like player bans, kill restrictions, passage of real world laws into virtual worlds and stuff like that. If you want to see a world evolve, you have to let it evolve.
Finally, and probably the most arguable requirement... Death is death. If you died, you're dead. Dead. You shouldn't "respawn", you shouldn't turn into a ghost. Most critics on MMOG's are on the fact that they are "excessively violent". Well, if death meant death in a virtual world, players wouldn't be killing each other (at least not that much). On the other hand, if in real life when someone dies he respawns in another place, people would be killing themselves all the time.