boiling the ocean and the profit motive

so VWRAP seems to be heading for the historical trash heap. there's been a little bit of conversation on the VWRAP list about rechartering. morgaine dinova has a wiki where she's working on a replacement intro draft, but after 3 1/2 months she's produced 681 words (the original VWRAP draft was around 9776 words.) at this rate, based solely on word count, we'll have to wait another 50 months. but i'm not hopeful the powers that be at the IETF will leave the working group open for another 4 years waiting on a single draft.

christa lopes posted a hatchet job on metaverse ink denouncing VWRAP as a linden-only club, worthy only of scorn. the article is largely inaccurate, being the memoir of events by a person who actively chose to not be involved. but as max ehrmann reminded us, it's good policy to listen to everyone.

but before VWRAP passes from people's minds, i wanted to offer a few notes about it's motivation; about why there there was VWRAP. diva and morgaine will tell you it's objective was to enable large scale interoperability between different virtual worlds, and that it failed miserably. nothing could be farther from the truth.

while it's true that VWRAP failed generally, it was never intended to be "all things to all people" in terms of virtual world interop. VWRAP was intended to provide a set of interoperability standards between hosts that implement services that created a "Second Life-Like" experience. you can actually read the group's charter, the OGPX mailing list and the original intro doc. it never says "we're going to provide interop between virtual worlds." it talks a lot about services and how they interact. so while you could use VWRAP to provide interop between worlds, that would be a side effect, not a stated goal.

before the VWRAP working group was chartered, we chartered a pair of "BoF Sessions." these are informal "Birds Of a Feather" meetings where people cuss and discuss the relative merits of forming a working group to work on some particular standardization regime. we started back in early 2009 with the MMOX BoF session.

david levine and i co-chaired the MMOX BoF and proposed a relatively narrow scope: standards for a second life-like world. we invited a LOT of people to participate from a LOT of different online game and virtual worlds companies. many simply said, "we don't think standards in this space are interesting" and didn't come. some said "we think standards are interesting, but don't want to limit the discussion to second life-like worlds."

because people like jon watte (then of forterra) and heiner wolf (of webkins) were generally interested, we put them on the agenda to talk about things they think should be standardized. after a fair amount of discussion on the MMOX list and at the MMOX BoF, we couldn't find consensus. but there's nothing wrong with that -- the world is big enough for multiple, even conflicting, opinions.

lisa dusseault proposed we that the "second life-like" camp go back and propose a second BoF that focused on "second life-like" virtual worlds while MMOX was maintained as a more general VW interop group. because i was more interested in "second life-like" worlds, i went on to propose the OGPX BoF.

the next summer, barry leiba, jon peterson and i co-chaired the OGPX BoF. this meeting was focused solely on OGP and "second life-like" worlds. representatives from IBM, linden and adobe were present while reps from intel participated remotely. the OpenSim core contributers were notable by their absence at this meeting.

following the OGPX BoF, a number of us hashed out the group's focus on the mailing list. in september of 2009 we reached consensus and agreed on a charter. we agreed that the working group would focus on the definition of inter-operable services that collaborate to render a virtual world. we also changed the name to VWRAP 'cause people thought the name "Open Grid Protocol" implied we were working on protocols to control general computing grids or cloud computing environments. we also agreed the space available for the group's charter wasn't large enough to contain all aspects of the group's work, so we also agreed the VWRAP intro would extend and refine the charter.

so people who tell you VWRAP was about "big I" Interoperability are fibbing. interestingly, the people who like to say these things, did not attend any of the meetings.

so why did we take the particular path we did? what makes service-level interop interesting?

i can only speak for myself with any authority, but here's what i gleaned from talking with David Levine and Suzy Deffeyes at IBM and Mic Bowman at Intel. let me apologize up front for any inaccuracies. david, suzy, mic -- let me know if i'm way off base here.

many of us work for technology companies. our work allows us to change people's lives for the better by proper application of technology. the world being an imperfect place, we sometimes have to resort to crass commercialism to feed ourselves. for the most part, we get to make cool products for discerning consumers and the money just happens. but at the end of the day, we must justify our actions to our investors or our directors.

one way we felt we could provide a legitimate service to the community while doing things that might be profitable to our investors was to start building a virtual world economy. it's pretty easy to see where people make money here. intel likes to sell CPUs that run virtual worlds. linden sells virtual land to people who want to inhabit virtual worlds. IBM makes coin by creating cutting edge virtual experiences for their global services customers. what you should note here is that no one company dominates this domain. there's no "apple" of virtual worlds; no one company makes CPUs, hosts online services and provides professional services.

this is likely due to virtual worlds being a somewhat immature and niche market. by niche i mean the market is much smaller than it could be and by immature, i mean we're at the beginning of the mass adoption of virtual experiences (which could be MMORPGs, Virtual Worlds or Augmented Reality experiences.) linden has long professed a believe that open source, open protocols and open interfaces will ultimately make them more profitable and make the industry more mature.

a number of people have pointed out similarities between the structure of the "virtual worlds market" now and that of online service providers in the late 80's / early 90's. like early AOL, Compuserve, The Source, Prodigy, BISON and others, Second Life is largely a walled garden. and everyone remembers what happened to the walled gardens: the ISPs ate their collective lunches.

when i pitched linden execs on the idea of a dedicated team inside linden to work on "open" protocols and interfaces, the refrain was "in five years we'll want to be known as the 'google' of virtual worlds and not the 'Compuserve' of virtual worlds." i have long been smoking from Eric von Hippel's pipe and believe that in the long run, "open" systems will always find ways to survive in business environments where walled gardens fail.

i am still a strong believer in the "make the pie higher" model where linden sacrifices a bit of virtual land revenue in exchange for the right to sell to other people's avatars. linden could, if they wanted to, throw open the gates and allow second life avatars to teleport into hypergrid regions (or vice versa.) linden is burdened somewhat by technologies, processes and a social contract that predates this virtual glastnost, but the problems are not insurmountable.

linden's participation in VWRAP was a gamble based on the belief: opening up one small part of the virtual world at a time would allow second life and its peers to grow together. taking a slow but steady path towards complete openness, concentrating on a small bit of functionality each quarter allows everyone to benefit.

david levine referred to this as "boiling the ocean one thimble at a time." start small with a constrained scope and just keep iterating, making incremental improvements.

some critics like prokofy neva (aka catherine fitzpatrick) argue that "open anything"is one step above godless communism. if i understand her arguments properly, she contends the ultimate goal of the open source and open protocol movement in virtual worlds is to institute global virtual collectivism where the "free culture" bolsheviks enforce global ownership of all virtual goods. (of course, i could be missing some of the finer points.) and she's probably partially right. there are plenty of free culture advocates that would prefer to see the concept of intellectual property consigned to the ash-heap of history.

but we don't have to make that mistake. open source, open protocols and open interfaces can still serve the global imperialist hegemony. or not. the fact of the matter is "open" is orthogonal to "free culture." one of the critical components of VWRAP was the trust model. it was discussed at length at IETF meetings and on the VWRAP mailing list, but sadly, was never documented to my satisfaction.

the trust model allowed individual service owners to decide whom they trusted and for what kinds of activities. for instance, if you ran a commercial asset service and you were worried that assets hosted by your service were in grave danger if they were ever instantiated in simulators run by the pirate party in sweden, then the trust model gave VWRAP implementations the tools to deny access to the pirate bay's servers.

that being said, open source and creative commons licenses serve an insanely useful purpose. to build protocols and systems that does not allow content owners the ability to chose the licenses their products were released with is just plain stupid.

so when people come back in a couple years and try to build another virtual worlds standardization group, just remember the following things about VWRAP:

• we weren't trying to boil the ocean. we weren't trying to ensure interoperability between worlds. we just wanted to define a few services and string them together.
  
  
• we wanted to support different IPR regimes ranging from rabid free-cultureists to DRM fans. we felt it wasn't our place as protocol designers to decide what license people used. we just wanted to deliver the bits and allow the endpoints to decide whether they trusted the person on the other side of a connection request.

thanks for listening.