Soon after the appearance of CRPG's (Computer Role Playing Games), a similar concept appeared, to present Multi-user CRPG's, what were to be called MUD (Multi-User Dungeons). In 1978 appeared MUD1, a platform to build MUD's: Virtual Worlds full of monsters and users willing to slaught'em. The concept evolved, and, besides the appearance of different types of MUDs, several other Multi-User Virtual Worlds appeared, with different focuses other than MUDs (like, for instance, Talkers or MOOs). In 1984 appeared "Islands of Kesmai", the first MMORPG (heavily inspired of MUDs) to be explored commercially.
Let's forget all the technology advances made on Virtual Worlds: while a good use of graphics, sounds and other technology advances might enhance the experience (specially embodiment) one has in a Virtual World, those aren't a part of the Virtual World per se: the Virtual World is the concept, not its implementation.
Business Models for Virtual Worlds
"Islands of Kesmai" had a business model quite similar of those we usually see nowadays on Virtual Worlds, including MMOGs: at the time users were paying $12 per hour. There isn't much evolution on the "new" business models, and while a lot of research is being made on that field, there are basicly three possible business models: the Time-Based Subscription Model, the Virtual Currency Model and Second Life's Virtual Real Estate Ownership Model. Some new models are being purposed and appearing both in research and real implementations, but they're always inspired by these three, and can be seen as simple derivations.
Defying Virtual Worlds' business models
The problem with the actual Virtual Worlds business models is simply that we are building Virtual Worlds as businesses, and thus modelling that Virtual World in such a way that it can optimise the results of its Business Model. While it's easy to understand the need for a company that decides to create a Virtual World to do things this way (and it's comprehensible if his objective is to make money), we're seeing the problem in the wrong perspective. Users don't want to be a part of a business - they want to be in a Virtual World. The implied limitations of these kind of Virtual Worlds are obvious when you see the difficulty in creating your "virtual persona", your avatar, that isn't just "your avatar in world xx" but is simple "your (ubiquitous) avatar". The need to fix that is seen if you take in consideration the creation of some "structured virtual ubiquitous societies", like the NationStates. So, while there's no wrong on having companies building businesses by the use of Virtual Worlds, taking the Virtual World itself as a business is what is really stopping the evolution in this field. Not the technical stuff (like better graphics, less lag) but the design of the Virtual World.
Building a better Virtual World
"So, if all the well-known VW's are wrong, how could them be right?"
Imagine a World Wide Web where, instead of being just a protocol running on top of a network (the Internet), each website had its own protocol, made by a company, and you had to have that websites client program (the browser), made by the same company, to go there. Cnet's browser would let you visit Cnet's website; Amazon's browser would let you visit Amazon's website. Cnet was making money by selling its browser, as well as Amazon by selling its see-a-list-of-books browser. Insane, right? Well, that's what's happening with VW's nowadays. So, how to fix it?
Well, we could use the parallel between the Web and VW's and say that what we need is to use a network where each node knows each other and that give us the chance of having each node linking to each other, and use that to define space. With it, and using a standard and unique protocol, it would be trivial to map from one layer to the other, and, with it, to have a "Virtual World Framework". Think of Neal Stephenson's metaverse, think about what is needed to reach that kind of ubiquity. Simple, isn't it?