Web 2.0 surpasses all previous technologies in its ability to “explicitly leverage network effects” (a definition of Web 2.0 from Tim O’Reilly). But, while virtual worlds pass another classic litmus test of Web 2.0 – two way participation, they have been, up to this point, largely cut off from Web 2.0 power/network effects.
Persistent immersive virtual worlds, led by Second Life, have done well as niche markets but they remain relatively isolated from Web 2.0, even though they bring something vital and new to the internet – real time interaction and dynamic melded states – in contrast to the current web’s large static files, or small changing files.
The slide opening this post is a modification of a slide from Dion Hinchcliffe’s presentation from his Web 2.0 Expo workshop – Building Successful Next Generation Web 2.0 Applications. Virtual worlds are not anywhere to be found on the original. So I asked Rob Smart, IBM, who has just added JSON support to OpenSim to draw OpenSimulator into this picture. In my interview with Rob, later in this post, he discusses the relationship between virtual worlds and Web 2.0 and how JSON is an important step towards virtual worlds taking up a place in Web 2.0 architecture.
When people think of the current architecture of Web 2.0 virtual worlds do not come to mind. But we are on the cusp of a big change in this regard. Linden Lab and OpenSim, in the Architectural Working Group, AWG, have been working on trust negotiations and the standardization, documentation and use of http (REST enabling). But more work remains on standardizing and documenting where TCP and UDP streams have to be used to create the immersive real time interactions that are the heart of what virtual worlds bring to today’s web (see my upcoming interview with Teravus Oursley, OpenSim, for more on this).
There is a complex network of connections through identity (1st and 2nd life) that have enabled virtual worlds to implicitly leverage the social networks of Web 2.0 (see botgirl’s lovely illustration of this above) The slide above is from W. James Au’s “The Post-Hype State of Virtual World Marketing: What Works, What Doesn’t and Why.”
Mryl (beta) is an application hoping to streamline these linkages with a social gateway for virtual worlds that will provide what KZero terms “outeroperability”. In this vein, Second Life developers have produced a number of interesting high level communications applications, including Chatbridge from Intersection Unlimited, to link Second Life better with the web. I will moderate a panel for Orange Island Innovation Week, Wednesday, Oct 1st, 12 noon PDT, Innovative Communications Devices, with Beyers Sellers, Chase Marellan (Chatbridge), Kevni Koolhaven (Learning Tree International).
But, it is the low level architectural integration of virtual worlds into Web 2.0 (along with improved usability and new User Interfaces) that will weave virtual worlds into the fabric of Web 2.0 and our everyday lives. But unlike Eric Schonfeld of TechCrunch, I see interoperability work (see OpenGrid Beta), and the production of standard protocols (see Open Grid Protocol, OGP) that interoperability work helps negotiate, as an important part of the process.
Immersive virtual worlds are still a long way from mainstream. I attended the Forrester Business and Technology Leadership Forum in Orlando last week to help Oliver Goh, business development executive at Implenia, talk about delivering results with virtual worlds. We found the audience, while familiar with many aspects of Web 2.0 and its business value, had relatively little direct experience with virtual worlds. But, the interest and excitement with this technology was very apparent.
Architectural integration of virtual worlds in Web 2.0 and the standardization of protocols (using existing web standards where possible) will change the picture, creating new opportunities to improve usability, create specific clients for particular needs, facilitate mashups, and leverage network effects, and more! And, JSON support for OpenSim is an important step as it allows virtual worlds to explicitly begin talking the language of Web 2.0.
Rob Smart is an Emerging Technology Specialist located at IBM Hursley where he works as part of the IBM CIO office Metaverse Initiative. In Second Life he is known as Yossarian Seattle and became known to some as the inventor of the translation HUD, which was his second foray into integrating Virtual Worlds with Web applications. The first project was enabling some of IBM’s messaging products to publish events into Second Life, including creation of an RSS Viewer for Second Life. Recently, Rob has been working with clients integrating their internal IT services with various virtual world platforms. His interests now extend to the OpenSim project, with a focus on integration of enterprise data and common web APIs into OpenSim.
Interview with Rob Smart, IBM
Tara5 Oh – on the right (me, Tish Shute) interviewed Yoassarian Seattle (Rob Smart, IBM) in Second Life outside Andy Stanford-Clark’s remote control house on Hursley island (for more see here)
Tara5 Oh: I am interviewing you from the media lounge at Web 2.0 Expo and coincidently it seems JSON is the hot standard here, in fact, the hottest it seems other than RSS for its ubiquity.
Tara5 Oh: As virtual worlds are still isolated from many of the network effects of Web 2.0, at the moment could you explain how integrating JSON support to OpenSim is “Web 2.0 made easy for OpenSim?”
But Second Life and OpenSim so far have poor string handling functions which meant that in LSL, in particualar, parsing anything more than a simple piece of JSON was just not an option.
Lots of coders and developers in Second Life have to run PHP and other scripts on external web servers to act as an intermediary stage in calling thse Web 2.0 APIs.
Thats a real pain, and means you need to have a server somewhere and up full time if others are to use your scripts. Whereas now, with this osParseJSON. function you can forget all that hassle and go straight to the source from OpenSim.
Its a simple but powerful enabler of Web 2.0 technology. I expect it will take people a while to find it and start using it, but it just widens the accessibility for those people who get into scripting in OpenSim.
I’m planning to do a similar thing for XML parsing, but its a bit lower on my priorities at the moment. JSON parsing gives a good quick win so to speak
Tara5 Oh: I just sent you a couple of slides ‘cos one thing I have noticed here at the Web 2.0 Expo is that the understanding of where OpenSim might fit into the architecture of Web 2.0 is vague to zero.
Can you sketch something that relates OpenSim into current understandings of Web 2.0 architecture?
Yossarian Seattle: Really in that first diagram with the APIs etc OpenSim just fits on the level of the web servers. And actually that diagram is a bit wrong as there should arrows between the web servers as sites should really be connected to each other. I’ll stick in here
OpenSim is being REST enabled. At the moment its access to assets, clothes, objects, etc. from the asset servers. But there is no reason that REST interface cannot give access to people logged on, object positions sim layouts etc.
Tara5 Oh: Could you explain the difference between the power of REST for virtual world technology in relation to the power of JSON?
Yossarian Seattle: So REST is really just calling a web URL. You use the tree structure of the URL to indicate your asking for different data. Whereas JSON is an encoding for the actual data that’s returned to you. So they are complementary really. But there has already been some discussion within the OpenSim community about introducing new APIs to OpenSim that allow different clients to connect.
I personally think that VWs are too siloed currently. At the moment in VWs it’s pretty much one world one client. Providing REST or other interfaces to the world data opens up the possiblity of a wider range of clients accessing those worlds. And when i say clients i’m talking about flash interfaces, browser interfaces or other 3d interfaces such as Unity3d clients.
Tara5 Oh: Could you tell me more about Unity3d?
Yossarian Seattle: Unity3d is a game engine. It’s a very flexible engine and adheres to a lot of the 3d modeling standards etc. One of its most interesting features is the ability to deploy the games/applications you make as web brower plugins (as well as windows/mac stand alone). I’ve been messing around with it for a while now, I can show you some demos while you’re over at the VW conf in London.
Tara5 Oh: Another theme at this conference, raised by O’Reilly in his keynote, is that the future is “world to web,” e.g ., sensor projects etc.
Yossarian Seattle: Ah well that’s another favourite topic of mine with regards to VWs Hursley is the home of realtime messaging technology.
At the moment as I say there is pretty much one VW client to VW server and because the only library to acces SL and OpenSim is openmv( formerly libsl) that restricts new clients to being written in c# There isn’t a java library, a flash library , a php library a ruby library etc.
So if in OpenSim we add new connectors, REST ones, talking JSON or XML then we enable lots of new client types and VWs become another mashable service in the Web 2.0 world.
Its about making it easy to get information in and information out. Web 2.0 sites don’t do realtime very well, whereas VWs do. VWs are the real time space that the web often tries to provide but kind of half fudges. Web Servers aren’t built to deal with realtime asynchronous data.
Its interesting how you mention Web 2.0 not really acknowledging Virtual Worlds as when I read the terms of service for a lot of the APIs they’re very specific about use from other web sites but they often dont cover the use of the API from other applications.
Tara5 Oh: Really?
Yossarian Seattle: Yes.
Tara5 Oh: What does this mean?
Yossarian Seattlee: It doesn’t necessarily have any significance for some services. But there is often specific text saying for example that you must use a particular piece of HTML on a page and show the API owners logo etc
I think as time goes on though and more people connect to Web 2.0 services from within VWs then they will be acknowledged as a valid service consumer, after all VWs are platforms that provide novel ways for people to display and interact with data.
Tara5 Oh: I know Hursley and other IBMers have done some nice use case of RL data integration in OpenSim and Second Life. What is your favorite for illustrating the power of Virtual Worlds to bring realtime world to web experiences to Web 2.0
Yossarian Seattle: Andy Stanford-Clark’s remote control house on Hursley island is still a favourite.
I did a hook up ages back with a messaging product MQTT and Second Life. I’d like to revisit that work and extend it. i’m interested in propagating events between platforms whether they be VWs or Web sites.
Tara5 Oh: I am amazed how little play virtual worlds have here at the Web 2.0 Expo.
Yossarian Seattle: Virtual Worlds live somewhere between the gaming world and the web 2.0 world. We see it with the flash social worlds too they edge more towards gaming.
Tara5 Oh: What do you think are the gains of virtual worlds getting more integrated with Web 2.0?
Yossarian Seattle: Virtual Worlds are a platform and and its often said by some that they’re not interested in taking part yet as they haven’t yet seen a killer app for Virtual Worlds. Some of that view stems from the fact that VWs are very isolated it’s hard to get content in and hard to get it back out again.
Virtual Worlds are the shared realtime spaces of the Internet, up until now this position has been filled by IRC chat rooms and instant messaging apps. Neither of these forms lend themselves particularly well to group interaction. VWs are streets ahead in terms of rich social interaction and sharing of content and experiences.
Tara5 Oh: You mentioned you just started working on OpenSim development and becoming part of this growing effort.
Yossarian Seattle: Yep thats right. There is a very vibrant community around OpenSim.
Tara5 Oh: Why did you decide to put your energy into OpenSim at this time?
Yossarian Seattle: I’m now working for IBMs CIO office Metaverse initiative and investigating all of the relevant VWs is one of our remits. OpenSim is my chosen focus.
Tara5 Oh: What is CIO?
Yossarian Seattle: One of the IBM CIO office responsibilities is to look at and provide technologies and tools that improve the productivity of IBMers world wide. But as you know IBM has several people working on the OpenSim project and there is an interested community internally. I’m looking at how we can hook up OpenSim to our existing web 2.0 services internally.
Tara5 Oh: What kind of internal Web Services?
Yossarian Seattle: We have a number of internal Web 2.0 based systems that provide APIs for data sharing, things like Blogcentral our internal blogging platform, Fringe which contains customizable profile information, Beehive is a social networking platform helps people share their interests, track and schedule events within IBM. We also have a platform called TAP (Technology Adoption Program) where people can share services and applications they have created with other IBMers. In addition we have Cattail, a file repository that allows easy sharing and tagging of all types of file. There are many more useful internal services than this even all of which could be integrated with OpenSim.
The nice thing is though that OpenSim affords that flexibility to integrate it with our products and with existing web systems, and provide value back to the community at the same time.
Tara5 Oh: So do you have any thoughts about the path to standards for virtual worlds?
Yossarian Seattle: In terms of standards I think it’s a case of look at whats out there and successful at the current time not just in terms of 3d models, but in terms of real time chat protocols like XMPP things like JSON, REST as well and pick those for the relevant components
The reason for this is every time you introduce a new standard, you have to wait for the communities to catch up and write language specific APIs for that standard.
[Better to use existing ones where possible and give the communities that will build the tools and the extensions a head start.
Tara5 Oh: This is also some of why top down standards like MPEG-V have issues?
Yossarian Seattle: Yep, standards often work best when they’re bottom up, like JSON.
As I mentioned before because the messaging structure currently for OpenSim and Second Life is proprietary (although open) and the only library is libsl (openmv) thats stopped a lot of potential innovation by restricting client/bot developemtn to the c# language.
Tara5 Oh: why is client/bot development restricted to c#?
Yossarian Seattle: Because currently the only library you can use to talk the Second Life libsl (openmv) is written in c#
Tara5 Oh: What do you see as the way through this obstacle?
Yossarian Seattle: If for example the messages that went between your SecondLife client and the OpenSim/SecondLife servers was a standard protocol which had a bunch of libraries for a variety of languages. Then you could start logging into VW servers from all kinds of clients
Tara5 Oh: Aren’t there plenty of standard messaging protocols to use?
Yossarian Seattle: Yep, but at the moment they’re not being used. There are some technical reasons for that. like reducing the amount of data to be downloaded etc. But there’s a balance to be had somewhere.
Tara5 Oh: But in a modular architecture like OpenSim what is to stop them being implemented?
Yossarian Seattle: There isn’t anything to stop them being implemented in OpenSim Which is why i like it
Tara5 Oh: I hear a lot about people wanting to change the physics in OpenSim/Second Life (the linking to the physics simulation in particular). Do you have thoughts on this or is it not on your agenda currently?
Yossarian Seattle: There are a few different physics modules already. Though to be honest i don’t think its the most important area to focus on, for me at least.
But obviously a high end physics engine is going to benefit anyone who wants to do any kind of simulation.
And that’s the beauty of Open Source, someone else will have that as their priority.
Yossarian Seattle: I think there’s a lot of work that needs to be done around ease of adoption still. i’d like it to be easy for people to write new clients for OpenSim.
When we get to that stage then people can produce simplified cut down clients to suit their precise need, so if you’re a retailer and just want to showcase products and let people shop you have a UI to reflect that.
Tara5 Oh: What about the OpenViewer project?
Yossarian Seattle: It’s a wider problem and piece of work. Well notice that openviewer is written in c# That’s because they’re limited again to using libsl (openmv). libsl recently changed it name by the way which is why i’m bracketing it.
Tara5 Oh: So it doesn’t address the underlying issue of messaging and open API’s for OpenSim.
Yossarian Seattle: Not really. But they have made the wise choice of releasing it under a free BSD license, which will enable more people to work on the project.
Tara5 Oh: Intel is working on breaking out openmv into smaller building blocks and basic types. How will this contribute to efforts to integrate OpenSim with Web 2.0?
Yossarian Seattle: Yes they recently hired John Hurliman who wrote a lot of it. i’m following what they do with interest.
Tara5 Oh: John wrote the original openmv?
Yossarian Seattle: He started the project back in 2006 .
Tara5 Oh: How will the work he is doing on openmv now help with the goal of making it easy to write new clients?
Yossarian Seattle: Well if they provide libraries in different languages that would be a good start and breaking it into chunks would allow anyone writing a client to pick and choose between the function they enable in their custom client.
However I’m not sure that’s tackling the root of the problem still.
Tara5 Oh: ‘cos the root of the problem is the messaging protocols which restrict you at the minute to C# for the client?
Yossarian Seattle: The standards need to be applied at the server end, to make it truly accessible.
Tara5 Oh: And these messaging standards need to allow for more than C# development?
Yossarian Seattle: exactly.
Tara5 Oh: well is seems like something quite doable, just time?
Yossarian Seattle: and careful thought
Yossarian Seattle: A lot of people are focusing on issues such as object portability in VWs but i’m not sure those are the ones to be concerned about right now, the games industry seems to have settled on collada as a standard for that. These VWs platforms are complex beasts and the games industry has already solved a certain amount of problems. However in terms of social interactions the VWs industry is ahead, a blend of games and social media.
Tara5 Oh: But games platforms have not solved either the web 2.0 effects or the web to world have they where things get most interesting now?
Yossarian Seattle: No and the games industry is playing catch up in that sense.
Yossarian Seattle: Little Big Planet will be the game that brings user created content into the mainstream for 3d worlds.
Yossarian Seattle: did you read this article? http://eightbar.co.uk/2008/09/10/moving-cubes-from-world-to-world/ that’s not a hack or anything in there.
That’s a full publish subscribe messaging client embedded in unity3d, realtime events across worlds.
Tara5 Oh: What do you think are the most interesting world to web ideas that Andy’s house points too?
Yossarian Seattle: Well the fact that the communication is two way, both in and out of world and also that its real time. when something happens in Andy’s real house it happens here too.
Tara5 Oh: Yes I am very interested in the development of the paraverse!
Yossarian Seattle: There is a personal project Peter Finn has been looking at in IBM, which is actually called Paraverse and is taking real world data including geospatial mapping information and applying it in OpenSim.
Unfortunately our interview ended here, at a very interesting point (I had to go to a panel at the Web 2.0 Expo, NYC). But James Governor’s post/essay – a superlative ode to the paraverse - prompted by his first look at Microsoft ESP visual simulation platform produced an interesting debate on the potential of the Paraverse in the comments that includes a response by Rob. So check it out!