Philip Rosedale:
Open Source, Interoperable Virtual Worlds

Fri, Sep 26, 2008

Metanomics host Robert Bloomfield interviewed Second Life founder and Chairman of the Board, Philip Rosedale, at the Second Life Community Convention in Tampa, Florida.  The Rosedale interview is available here (pictures above are Philip Rosedale and his avatar).

Rosedale talked about Linden Lab’s long standing commitment to open source and open protocols in one segment of this interview and Robert asked me to post a brief reaction. The full interview covers a wide range of topics and Robert has gotten responses on different parts of the interview from Wagner James Au, Christian Renaud, ‘Bettina Tizzy,’ Nic Mitham and ‘Dusan Writer,’ and Benjamin Duranske as well.

A System Without an Owner is A beautiful Thing

While Philip Rosedale’s comments may not, at first glance, appear to be saying anything new, they are in fact a very cogent summary of the important and crucial role Linden Lab has played, and continues to play, in moving virtual worlds out of their walled gardens and bringing them closer to that beautiful thing – a system without an owner.

Only a system without an owner can unleash, for virtual world technology, the kind of creative, world changing power that we have seen on the 2D web from http and html.  Anyone with even a vague idea of the history of the internet understands that it is only through openess, open source, open protocols, open standards, and open APIs, that we will get from here – the alpha days of virtual world technology, to their coming of age of age as a mainstream phenomena.

It is very much to the credit of Linden Lab that, as Rosedale says, they have never been afraid of openess: “I don’t think that the open grid will impact our revenues any more than open sourcing the client,”  he says. While there have been criticisms of licensing choices and ways Linden Lab handles contributions back to their viewer from the community, I think that overall Linden Lab has made very important and visionary moves, first to open source, and now to open protocols.

Open sourcing the viewer at a relatively early point in Second Life’s development created an enormous opportunity for the rapid development of an open source re-engineering of the server side, OpenSim.  OpenSim with the Second Life viewer is the most complete, open implementation of a persistent virtual world.  Without the head start from the open source Second Life viewer, and the connection to the thriving developer community of Second Life, the light speed progress of OpenSim would have been considerably more difficult.

Now OpenSim is getting closer to breaking free from the Second Life viewer. And, standard messaging protocols between client and server are, perhaps, the next step. Rob Smart, IBM, discussed this with me recently (see my upcoming interview with Rob Smart, “Web 2.0 Made Easy in OpenSim,” and see his post by this title for more).

As, Rob Smart, IBM, notes, “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.”  (for more see my upcoming post, “Interview with Rob Smart, IBM: Web 2.0 Made easy for OpenSim.”

Open Standards Will Emerge From Rough Consensus and Working Code

There are some that subscribe to the view that standards will arise in a virgin birth from an ivory tower, i.e., professors and captains of industry, removed from open source developer communities, will produce long documents that describe all of the fields, and every one of the messages, and all the APIs in detail prior to implementation.

But as, David Levine, IBM. Mike Mazur, 3Di, Mic Bowman, Intel, Justin Clark-Casey, and Adam Frisby, Deep Think/Sine Wave cogently argued, on the “Open Source and Interoperable Virtual Worlds” panel at the Virtual Worlds Conference and Expo in LA, this top down approach to standards, (or “vapor standards”), does not, typically, produce good results. For more on the the virtues of creating standards from “rough consensus and working code” as opposed to top down there is a full recording of the LA panel here.

Thus, in my view, Linden Lab’s current focus on open protocols, OpenGrid (for more see here), and interoperability is another key step towards the creation of open standards for virtual worlds. And Linden Lab are again leading the way in creating an environment that fosters innovation.

OpenGrid creates a testing ground where protocols can be worked out, and it enables the kind of heterogeneous ecosystem to develop that can nurture the creation of standards. I  agree with Rosedale when he says content makers will have an important role in driving interoperability and standards. The creation of standards is certainly a social as well as technical process. And as Rosedale notes content creators will have compelling reasons to move their content around in an open metaverse.

David Levine’s (IBM), described in detail in LA (again see recording here) the importance of interoperability and parallel innovation for the creation of standards. OpenSim has already produced an extraordinary amount of innovation, realXtend, Tribal Media and more. Also see my interview with Mic Bowman, Intel, for more on the role of open source/open standards in fostering innovation and in moving virtual worlds into “the fabric of everday computing.”

While Linden Lab only have a small team working on OpenGrid, it is a vital one.  And, with MarkLentczner (Zero Linden in Second Life) leading the Architectual Working Group for Linden Lab, and a collaboration with IBM led by David Levine (Zha Ewry in Second Life) driving the interoperability effort, plus the OpenGrid project, Linden lab has a high powered, agile, lean, machine working for an open future.

So with no more ado, here it is: Robert Bloomfield’s interview with Philip Rosedale!

Rosedale on Open Sim:  Pandora’s Box Was Already Open

Introduction from Robert Bloomfield

Naturally, a major topic of my interview with Philip Rosedale was on the implications of OpenSim and the Open Grid project, which both involve creating open source server-side implementations of virtual worlds that can replicate Second Life’s funcationality.  As a relative newcomer to this corner of the tech industry, I still find myself asking what a company would essentially create its own competitor.  Here is what Philip had to say; I have asked Tish Shute of UgoTrade to comment, as one of the people who has covered the OpenSim/OpenGrid movement with more detail and passion than just about anyone.

PHILIP ROSEDALE: I just really hold true to the strategic belief that there’s going to be a tremendous amount of consolidation and interconnection between these worlds because the content development process is so challenging that the content developers are going to push us all together. They’re going to say, “Give me a file format. Give me an interchange format. And let me move that chair from this grid to that grid. I’ve got to be able to do that because I’ve got a customer here who wants to buy it.” And so I think that that consolidation is going to happen, and it’s going to happen earlier than people would have thought.

ROBERT BLOOMFIELD:  And this is looking at the success, the energy around OpenSim, open grid.

PHILIP ROSEDALE:  The energy, yeah. I think, at this point, we’ve got an appropriate level of energy – I think that’s exactly the right word – around exploring how quickly we can generalize all this stuff and open and interconnect everything together. I really think that’s going to continue.

ROBERT BLOOMFIELD:  [D]o you feel like you might have opened Pandora’s box and that it’s not really under your control now?

PHILIP ROSEDALE:  I think that Second Life has, in many ways, not been under our control from the beginning and that it’s been a basic operating assumption that to create the kind of incredible place and business opportunity, and social opportunity more broadly, that Second Life would require a certain lack of control. And that was true with the content from day one.

So for us, oh, we open-sourced the client a while ago, and now we’re trying to do the same thing with respect to operating standards to interconnect grids. This is a pretty logical progression, using worlds that we’re pretty familiar with. I mean we’ve always felt that, if you have a compelling use proposition, which certainly Second Life does, in other words, if there’s real utility, real fun or real business or real whatever in what people are doing, then there should be a way, as a company, to be open, global and still make money on an hour-to-hour or a user-to-user basis or whatever on what we’re doing. And the economic aspects of the business have been fantastic from the very early days, and we don’t really even worry about them.

Our ability as a company to find a way to make a reasonable amount of money per hour that people spend in Second Life, it’s really never been that much of a problem. It’s actually been fascinating as we’ve changed pricing and as we’ve changed the ways that we make money. Introducing new ways of making money –  like selling currency on the LindeX – it’s been amazing how stable our revenues have been as a function of usage hours. It’s one of the things that we sometimes marvel at. It’s almost an emergent effect, if you will, that the company’s business, its operating revenues are really very stable.

ROBERT BLOOMFIELD:  Even though they’re coming from different streams.

PHILIP ROSEDALE:  Even though they’re coming from different streams. And sometimes the requirements of the platform and decisions that we make will really substantially change the nature of those streams, but when you put them all together and you divide them by the number of usage hours, it’s like a constant. It’s almost a magic number. And it’s a magic number that allows us to be profitable, and therefore, is certainly adequate to make a business in the future. I don’t think that continuing to open Second Life up as we have been is going to impact that. Again, I just think there are so many opportunities to make money that we shouldn’t have to worry about that too much in the company. And, again, I think that’s a lot like the early internet. I mean if you step back and look holistically at the internet – you look at PayPal, the payment systems, auction systems, transaction systems, posting, naming – you look at all the businesses that comprise the internet, well, those are all the kinds of businesses that we as a company can be in, in this emerging market. There’s no business that’s denied us. We are in the hosting business. We can continue to be in the hosting business long term, putting servers up and providing access to them.

We can certainly be in the naming business. We’re in the currency and transaction support business. It’s funny, it’s something that’s often discussed. We worry much more about improving the scalability, stability and the usability of the system: reducing that initial user experience, reducing the time associated with it, making it easier. That’s got to be the lever that drives more growth in the overall industry, more revenues for us. So it’s really all we worry about. But I don’t think that the open grid will impact our revenues any more than open sourcing the client did.

categories: 3D internet, Architectural Working Group, free software, Intel in Virtual Worlds, interoperability of virtual worlds, Linden Lab, Metaverse, Open Grid, open metaverse, open protocols for virtual worlds, open source, open standards for virtual worlds, OpenSim, Philip Rosedale, realXtend, Second Life, vapor standards, virtual world standards, Virtual Worlds, Web 2.0, Web 3D, Web3.D, World 2.0

4 Comments For This Post

  1. Yohan Launay - ConceptSL Says:

    SL and OpenSims interoperability can be a good thing, but what about other platforms ? I can’t help but think that general VW platforms interconnected is not really what people are looking for right now. What worked well with Internet (and made it the Internet that we know today), was purposeful applications / websites – Internet has been litterally built around those.

    Plus, there is the political / strategical problem. All VW platforms don’t see an interest in being open or interconnected. Why would WOW want people to jump to SL ? (Although I would understand why SLifers would like to jump to WOW). Its like putting ads on your website : you just drive traffic to a third party. And unless they find real value to it (not just answering some tech geeks requests), they won’t make the leap.

    Interesting POV, thanks for sharing.

  2. Tish Shute Says:

    Hi Yohan,

    Thanks for you comment! If you have the time please do listen to the panel on “Open Source and Interoperability” – full recording here – many of your points are answered more fully there. But here are a couple of short answers.

    1) The interoperability and Open Grid effort is by no means limited to OpenSim and SL. We made call during our panel in LA for other virtual worlds to get involved in the effort. And when I have talked to Linden Lab, Lively and Vivaty engineers and Vast Park (just happen to be the ones I have met recently) there seems to be much interest in broadening the effort.

    As close cousins, SL and OpenSim, are an obvious pair to begin with. However, creating a heterogeneous ecosystem for parallel innovation is what will be beneficial for the eventual emergence of standards. Again this is explained in much more detali by David Levine, IBM, Mic Bowman, Justin Clark-Casey, Intel, Adam Frisby, OpenSIm/ Sine Wave, and Mike Mazur, 3Di

    But, for good measure, here is one of the points David Levine, IBM, made on parallel innovation again:

    “The point about parallel innovation is that there are several ways to evolve things. Open Source, tends to foster the addition of features, in common to the main trunk of development. Its very much about everyone building together. Sometimes, we need to or want to build stuff which works with that main trunk of development, but doesn’t share code, or has to fork. That’s where standards come into play.”

    Also note this post on eightbar for a deeper understanding of virtual world interoperability.

    2) Also re your point about the “purposeful applications” one of the direct results of the emergence of standards will be that the current isolation of virtual worlds will come to an end and there will be an amazing flowering of innovation and useful applications. I hope you will read my upcoming post on “Web 2.0 Made Easy for OpenSim” about the recent integration of JSON support to OpenSim. Part of the emergence of standards for virtual worlds will involve integrating support for existing standards like JSON.

    3) Yes there is resistance to standards from some within the virtual worlds community. And perhaps that is because, at this early stage, many of the business models have arisen around a proprietary closed model. But the situation is changing fast. And, the fundamental importance of virtual worlds as the first real time collaboration over the internet in a rich, immersive environments is too valuable to remain an isolated niche market. That is why I try in this post to give credit where credit is due to Linden Lab for their vision re open source and open protocols so vital for the emergence of standards.

    I think small mindedness has resulted in a lot of misunderstanding about how standards will emerge in virtual world community. For example, there seems an inclination to put off the process to some distant future set of “vapor standards” to be decided by a committee somewhere, instead of rolling up the sleeves and getting down to doing the work as Linden Lab, OpenSim and IBM have begun doing.

  3. Ben Goertzel Says:

    Yohan — just a quick response regarding your comment on “purposeful applications.”

    My own concern, regarding virtual worlds, is precisely with the development of purposeful applications based on advanced AI technology. And from this perspective, it’s very clear to me that interoperability and open standards will be Very Good Things.

    Right now, for each virtual world or game engine I want to hook my AI up to, I’ve got to do way too much customization work, both on the client side (graphical assets are obviously not the same in SL as in most other platforms, for instance), and on the server side (building a proxy to intervene between my AI code and each particular world or game server). So, the process of creating purposeful apps is slowed down a lot by the fact that rolling one’s app into each particular world requires so much fiddling-around work (that is really not critical to the particular value of one’s app).

    To take an analogy, I can make an animation in Flash and then embed it in all sorts of websites without needing to write a proxy between my animation-widget and the specific webserver or application server running that website. But with virtual worlds this level of simple, obvious interoperability is lacking, and not for any terribly good reason, just because virtual worlds are early-stage, and didn’t develop from some centralized government project but rather more organically.

    Relatedly, Dan Miller and others have been helping me think through the problem of integrating robot simulators and virtual world engines. We’ve come up with some nice, concrete ideas in that regard … but, wouldn’t it be nice if we could avoid reimplementing them separately in every world engine/client combination? … and avoid making this kind of integration specialized only to one platform (OpenSim or RealXTend or whatever…).

    So my overall point is: in many cases, the interoperability is what’s going to foster the creation of the apps. Interoperability and open standards provide some direct, obvious end-user value, but they also have a more indirect (and maybe more dramatic) value to the user in the way they create an environment making it more economically viable for individuals and companies to create purposeful virtual-world applications…

    – Ben Goertzel

  4. Tom Hoff Says:

    One point that may be worth consideration is Linden Lab already has a rather large scale infrastructure in place for supporting their hosting business, and unless a major technological leap radically changes the design of such an infrastructure or renders it obsolete, they should have economies of scale on their side for several years to come. In this case they have another advantage where their revenue stream could finance the efforts of the talent they have attracted and employed to continue to enhance their product and help them maintain a market lead. Viewed from this perspective, opening up the platform specs and allowing competition may be to their advantage as open source efforts and continued acceptance of virtual reality technology should draw more customers towards all providers. I suspect such a business strategy could be quite fruitful for them, provided they can manage these changes to their business model without losing vision or falling behind.

    On the subject of interoperability with other services, there may be so little in common between services such as SLâ„¢, WOW, and others, that none of the viewers could be adapted to work in the other worlds and provide a usable experience. In this case, interoperability may be confined to simple communication such as voice or text chat and probably would not present a threat to “walled garden” services until a common viewer platform could be developed.

1 Trackbacks For This Post

  1. The discourse about Interoperability: 1 year wrap-up - Connecting Worlds Says:

    [...] here’s a first list of selected interviews, posts, audio and presentations (newer to older): – Philip Rosedale: Opensource, Interoperable Virtual Worlds – The Interoperability panel at the recent LA Virtual Worlds Conference and Expo. Full audio here – [...]