(Full disclosure: I used to write for secondlifeinsider.com, a Joystiq spin-off site that covered all things Second Life from the inside, hence the clever name. The site is still there, with tons of archives for you to read through; if you're curious about what really goes on in SL other than what the mainstream press would have you believe, check out the archives.)

Having said that, I'm not about to make the case that virtual worlds like Second Life will necessarily rise up to subsume the Internet and everything within it. For one thing, we're not quite at the saturation point of number of humans on the planet being equivalent to the number of humans online. There are still a good number of homes out there that aren't Broadband-enabled, and that's one of the key components to the All-Net Virtual Community (whatever it's actually going to be called). For another thing, there is no mainstream anymore, no one source for entertainment. Thus there won't be a single place for every user to call home. Finally, any such gathering place will, by necessity, bear a similar structure to the 'Net itself: no one entity will be in control of it all. The ANVC will have to be an unregulated, decentralized network, or it will fall under the whims of some seriously moneyed and out-of-touch megaconglomerate who will try to turn it into a business venture, and that never works well.

And hey, we already have this persistent interconnectedness in the 'Net, so what more needs to happen to make this work?

It seems like those who are betting big on the future of the 'Net as a huge virtual community -- similar to Snow Crash's Metaverse -- are lacking the "killer app", the one compelling product or service that will tie everything together and make it irresistable for the average Joe User to sign up and join the fun. To look at current popular virtual worlds, we can abstract out what their particular strengths are, and see if there's a way to push those ideas forward into a single, all-representative framework that will appeal to everyone.