Exploring the Future Virtual Cosmos With IBM
ComputerworldUK has a nice article up about the possibility that different virtual worlds will one day support a standard that would let users travel freely among them. This is an idea I’ve been hot on since even before starting this blog, so it’s nice to see other people supporting it — especially when they’re people like IBM vice president of standards and open source Bob Sutor, who’s quoted in the piece. Sutor has been putting up a nice series of posts on his blog since the beginning of June, detailing his basic requirements for virtual worlds, his desire for more VW artificial intelligence, some scenarios for moving assets, information and identity among virtual worlds, and the need for worlds to run on multiple platforms. (Sutor will be at a virtual worlds event at MIT’s Media Lab this Friday, apparently, though I can’t find a link.) A lot of what he’s talking about in those posts, if you ask me, points toward the broader future of virtual worlds. But feel free to poke holes in my arguments below. Even if it’s only to complain about the great length of this post.
I was especially excited by Sutor’s vision for a more distributed virtual world, or worldlets, which squares really nicely with some of the stuff Peter Ludlow and I write in our forthcoming book but which actually goes further, which is very cool. Toward the end of the book, we contemplate an open-source virtual universe in which users can host their own sim, or 3D environment. You can teleport from any sim to any other (as you “teleport” between Web pages today), or many sims might sew themselves together into smaller or larger continents.
Sutor describes “a peer-to-peer model where I could link my private local world to someone else’s, and that would constitute our entire universe. Now generalize, and allow multiple simultaneous links. Finally, also allow me to link my local world into the master grid.” He also envisions “a model of many planets, solar systems, and universes that could be grouped in any way people want to do so. These could be by function, by affinity, or by whimsy.”
This seems to me inevitable. As open-source metaverse development accelerates, it will push toward a kind of LAMP stack for virtual worlds, something simple and available enough for anyone to set up. This doesn’t require a new networking protocol, necessarily, but would entail a virtual world protocol that woudl ride on top of TCP/IP (a VWTCP/IP?), and that would allow one 3D space to communicate sensibly with the next. There may be competing standards for this for a while, but I’d imagine that they eventually merge into one, as the network effect takes hold and users want to be part of the broader community, or at least have the ability to interact with it, rather than being sequestered off in a space that couldn’t talk to the outside world even if it wanted to.
The tough problem that needs to be solved in order to get here, in order to create a Web of 3D environments in which I have the ability to represent the same presence in different spaces, is the portability of identity, as well as the management of assets in certain situations. Ideally, identity will be centrally managed in such a virtual universe, if only so that I don’t have to log into each and every 3D space I want to visit. Depending on the technological model, a centralized system may also need to pass some assets and information from one realm to the next — though not, most likely, to store all the assets for all the users of the system; that task is just too big.
This central space, this spot at the midst of the virtual galaxy, is not a place that gets mentioned very often. But it will eventually be perhaps the most valuable piece of real estate in the virtual cosmos. Occupy that spot and you become the traffic cop that directs all the little virtual packets that are flying around the 3D Internet, and can leverage your presence there into all kinds of services. Maybe this function gets taken over by some magnanimous international agency, but I doubt it. Even if it does, there will still be room for service providers to stand there and pick off clients.
My bet is that it’s this spot that both IBM and Linden Lab, makers of the virtual world of Second Life, have their eyes on in the long run, in slightly different ways. I don’t know that IBM wants to control all that traffic, but I do think they want to be able to stand in the eye of the storm and offer their services to all comers. This is pretty much what they’re saying with their “we want to do for the virtualworld what we did for the Web” stuff.
Now, Linden Lab, being all nefarious, probably wants to stand at the midst of it and own all the identity information and charge you for making it available to multiple realms, etc. But wait a minute: How did IBM become the benevolent innovator and Linden Lab become the evil empire? Oh well, I take it back. Kind of. But LL has talked in the past about eventually open-sourcing the Second Life server as well as the client, but perhaps hanging on to one layer of their metaverse stack as a business model.
If that happens, any viable business model would require that at least one aspect of your metaversal interactions pass through the LL service, of necessity. It might not be identity information; it might be economic information, some information about your assets, or simply some Virtual Record Locator information that allows your client to resolve which region you’re heading to next. If LL is going to open-source their only product, then they’ll have to come up with a service to replace lost revenue. Since, short of buying the Electric Sheep Company, they don’t have a lot in the way of services to offer, it’ll have to be some form of traffic-copping the metaverse.
I’m not totally convinced that’s a winning long-term strategy, unless — and here’s where Linden Lab becomes the big bad wolf — they can gain the kind of dominance that Microsoft enjoys. That presumes, however, that theirs becomes the standard for 3D spaces on the Internet. Given the number of competitors just now coming to light and the problems with LL’s software, I’d say that’s a longshot.
Where does this all leave us? No idea. These are really just some long-winded musings on the future of the 3D Internet. Hopefully they were at least vaguely enjoyable and will spark some response in the comments thread. The virtual cosmos is your to enjoy.
Feeds
Hey, great ‘musings’ and good info. links!
Yes a standardized grid environment offering portability would be great. This would be so big that long term it is hard for me to see any single company, or even handful controlling it all. If nothing else, politics would get in the way.
Okay, OT, but cool all the same, just got through using voice in SL First Look Viewer on Sheep Island, one of the few I have found to be voice enabled. Apparently will be rolled out to different locations daily. Bet we will see a thread here on 3PointD soon, LOL.
Hi Mark,
Nice post! I agree with this vision of the ‘multiverse’ as a collection of many distinct universes that we can seamlessly pass between. I also think the web, as we know it now, will be a large part of that multiverse — and that our identity should remain consistent between the 2d and the 3d worlds.
Open Croquet is a great start. It’s LAMP-based, open-source, peer-to-peer, and has the ability to link universes together. Hopefully, it will continue to mature in this direction.
I’m sure the identity issues can be handled with a standard much like OpenID… Or even with OpenID itself.
Keep brainstorming! :)
Hilary: OpenCroquet is not LAMP based as it doesn’t use Apache, MySQL, or any of the P languages. I’ll be very surprised if SmallTalk on a custom virtual machine replaces LAMP any time soon since it’s been shunned by developers for more than a decade.
Mark, quite a few of the ideas that you are talking about have already been implemented and are in fact available today. I outlined a Metaverse that is similar in a lot of ways to what you are describing. I put my thoughts together a white paper, The Metaverse 2.0 http://www.pelicancrossing.com/WhitePapers/TheMetaverse20.pdf which was published a few months back. It is a big document, and I would advise against reading it while operating machinery!
Most if not all of the things talked about in the white paper are doable today using Blink 3D from Pelican Crossing, Inc. Blink 3D is currently only a V1.x product but exciting new features are being added all the time. For example, the current Community Preview build supports the ability to easily add metrics to a Virtual World and then analyze the resulting data using Google Analytics just as you would a Web site.
As for the holy grail of wearing the same avatar no matter where you go, I have to disagree with you on this one. I address this idea in detail in the white paper. But to summarize: it is a big technical challenge, your avatar is going to be limited to the lowest common denominator, a mesh is not a mesh is not a mesh, meaning that your avatars actual mesh would need to be converted along with any accessories, not all technologies support morphing, not all technologies support individual scaling of body parts, if you paid for your clothes in one VW using their money do you have to buy the same clothes again in the next VW you visit, where is your avatar stored, how do other VW’s get access to it, how do you prevent other people from getting access to it, how do you prevent clothes or accessories that have been paid for in one VW from being resold or pirated in another competing VW, how are avatar updates handled, do all avatars have to be humanoid or can someone use a house as an avatar?. Some VW’s will want to control the look and feel of avatars for either consistency within the VW or for performance reasons, keeping the poly count down. Blink 3D’s default avatar is a green Ninja, I think it would look a bit out of place in Club Penguim! or World of Warcraft or a VW that is a Virtual Heritage site. I’m not saying that these problems are insurmountable but I think there are bigger fish to fry out there.
Unlike the majority of VW’s out there Blink 3D actually runs inside a Web browser so I tend to take a broader look at VW’s, beyond the game like “let’s all dress up and buy each others stuff and chat” kind of thing (not that there is a problem with this, on the contary). Applications for VW’s go way beyond these kinds of VW’s. I see the majority of VW’s as being open to anyone who wants to visit them just like Web pages are open to people today, there are no special prerequisites in order to experience the VW. Just like the Web, there is little requirement for Web surfers to log in anywhere, I think this was one of the key reasons that the Web has been so successful because it was so open and free wheeling. Sure there are exceptions where it’s a subscription based member’s only thing and those exceptions can be accommodated. But this openness removes the need to login everywhere. You don’t need to login anywhere to visit VW’s created with Blink 3D. Having said that, the use of something like OpenID would be a good idea for those VW’s that require a login but OpenID does not address a lot of the things that would be needed for this to work.
Nice discussion. Seems like Mark is talking about a certain class of virtual worlds (VWs) where each person has just one avatar at a time, and they use this avatar to interact with and view the VW. Not only is the avatar a 3D representation of a user in the world, but it also usually represents the user’s view, which makes it handy for interacting with other people in the VW. But that the avatar represents the user’s view is a pretty arbitrary rule. Second Life already breaks it to some extent: the camera can fly around independent of the avatar. The SL “open client” architecture will enable the development of even more “omniscient” views. Right now I have 5 tabs open in my web browser. Why not 5 simultaneous view ports into one or more 3D worlds? Looking at it this way, the avatar is reduced to just the insertion of a user-controllable object into the target VW, and it’s not a necessary condition for viewing or manipulating the VW. As we become more adept at manipulating and reading 3D environments through better 3D browsers (teleporting our views everywhere, viewing from multiple angles simultaneously, etc) we may find that the old single-location avatar is no longer the most useful way observe other users.
I’m trying to rewrite my last sentence at least:
“As we become more adept at manipulating and reading 3D environments through more advanced controllers and browsers (augmenting our gestures, teleportizing our views everywhere, viewing from multiple angles simultaneously, etc.), we may find that the typical single-location avatar is no longer the most useful construct for representing ourselves or our attention to other users.”
Okay, that’s a little better. While I’m back, I guess I might as well inquire here about the possibility of getting user editable comments on 3pointD. Is there a way I can edit/ delete my posts? Would it be better for me to post on my own blog and add a link here? Somebody help a blogospheric noob. Cheers.
And, dude, Mark, thanks for all these great links!
I’m trying to rewrite my last sentence at least:
“As we become more adept at manipulating and reading 3D environments through more advanced controllers and browsers (augmenting our gestures, teleportizing our views everywhere, viewing from multiple angles simultaneously, etc.), we may find that the typical single-location avatar is no longer the most useful construct for representing ourselves or our attention to other users.”
Okay, that’s a little better. While I’m back, I guess I might as well inquire here about the possibility of getting user editable comments on 3pointD. Is there a way I can edit/ delete my posts? Would it be better for me to post on my own blog and add a link here? Somebody help a blogospheric noob. Cheers.
And, dude, Mark, thanks for all these great links!
I think it’s inevitable that the multiverse is leading to this ’shared world’ experience. All participants certainly want this to happen, if only the architecture would cooperate. I think we all want it on a subconscious level. The main question is–as you have pointed out–is who will be responsible for maintaining the ‘central grid.’ Would egos allow this to happen, i.e., for the benefit of everyone? Will the subconscious need for a “connected worldwide experience” allow for this abeyance of ego? I think it probably will.
Innerspace is just as unknown and as exciting as outer space. It is fascinating watching this develop.
As far as avatars morphing from one world to the next seamlessly…I think the evolution of the avatar itself is going to continually change to look more and more like the person behind it. In the not too distant future we will likely have a small camera recording our facial expressions and reflect those same expressions onto the avatar. In fact, the avatar in ten years may be indistinguishable from the real “you”–especially as more and more ‘virtual reality’ comes into the landscape. So as long as the avatar looks like the real person, it would seem less likely the clunky avatars of today (and the various architectural requirements for same) would tie us down. All virtual worlds would adopt this same “You.” It would seem to me, then, that the entity who works this out the most realistically would be the best one to manage it (probably IBM). IBM would manage the “look” and the “functionality” of the avatar. The assets could be managed by, say, Goldman Sachs (or someone proven with real world asset management).
All this would have to be united under some centralized non-profit roof (maybe governed by the United Nations?).
As I said, this is all fascinating–and workable.
I totally agree about the need for more freedom and flexibility in running different discontinuous worlds linked to each other. In fact this is the natural and obvious way to proceeed, SecondLife stops short of this (they have reasons to do so: fostering a particular kind of user experience, real-estate revenue model).
However, I don’t think that building *on top of* LAMP will work. HTTP protocols and even just TCP cause problems for real time dynamic systems (http://interreality.org/wiki/VipDocumentation).
Reed
We don’t need one central global authority to host personal identity information (for example, RSA keys, random biographical stuff like name). You can host this information anywhere, for example, on your own server, or an organization can host such information for its employees/members, or competing services can host them on your behalf. We just need a standard protocol and schema for different sites to access that info (like OpenID, though I don’t know enough about OpenID yet to know if it’s sufficent). Then when you join a world, or meet other people in world, you supply the address of this information and the world server can go verify it for its purposes (e.g. authorized to modify the world?) and other people in the world can verify it for their purposes (trust this guy?).
For stuff like avatar geometry, I think we can probably settle on a set of common formats (cf. image formats) that different VW technologies can use. Then the client can supply that directly to others, no need for a central authority to store it.
Reed
Link for today’s MIT event - sorry it’s after the fact:
http://www.research.ibm.com/research/press/virtualworlds.shtml
Mark, this post is very inspiring! Mason’s thought, “the typical single-location avatar is no longer the most useful construct” blows my mind. I think that what Microsoft’s Identity Architect, Kim Cameron has been working on in regards to Identity Metasystems is interesting and worth reading if you have not had the chance. http://tinyurl.com/2kln6a It is increasingly becoming evident to me that the need for a system(s) that can carry digital assets/ID info around with users is a necessity in RL, VR, & 2D web space transit and interactions.