Re: [mmox] User Accounts / "Object" Transfer.

["Hurliman, John" <[e-mail]>]

>-----Original Message-----
>From: [e-mail] [mailto:[e-mail]] On Behalf Of
>Meadhbh Hamrick
>Sent: Thursday, February 19, 2009 12:06 PM
>To: Charles Krinke
>Cc: [e-mail]
>Subject: Re: [mmox] User Accounts / "Object" Transfer.
>
>the draft charter is the best guide for what we're hoping to achieve.
>
>you can find a copy at
>http://trac.tools.ietf.org/bof/trac/wiki/MmoxCharter
>
>the things to remember here are:
>
>a. we're not trying to solve the problem of interoperability between
>every existing virtual world. at least not in the near term.
>- and... by extension... we're not trying to specify a general-purpose
>virtual world application stack that would enable a single client
>application to interoperate with any virtual world.
>
>
>b. we're not trying to solve the general problem of "what is a virtual
>world." i would refer you to the VWR and MPEG-V groups for that.
>
>c. we're trying to build a flexible, extensible application layer wire
>protocol used for virtual world interoperability. but the protocol we
>define here may need to be augmented by an "interoperability profile"
>- in other words, we're defining mechanism, not policy. for example,
>second life and opensim now have a "one simulator per 256m square
>region." we may not want to enshrine this in the protocol (i mean...
>what does this mean in the context of a virtual world like EVE Online?)
>instead, we should put this in an interoperability profile. the profile
>would reference the ogp base documents, but add implementation details
>required for proper presentation of the virtual world, but not required
>for the definition of the wire protocol.
>
>d. OGP is not what Second Life currently uses. OGP was originally
>conceived as being the AWG's "next generation" virtual worlds protocol.
>(the AWG being a technical architecture working group with Linden, IBM
>and OpenSIm contributors) for what it's worth... we demonstrated
>interoperability last summer between systems managed by linden lab and
>various partner systems using open sim and PyOGP.
>

At the first AWG meeting in September 2007, it was made clear that Linden Lab had already defined the architecture of OGP, and the purpose of the AWG was to refine and implement this protocol. I'm not sure that demonstrating compatibility between three codebases that all speak the Second Life protocol could be categorized as interoperability.

>e. SL OGP (the OGP drafts that were written last summer) is not a
>complete body of work. there's still a lot of stuff to add. part of
>linden's motivation for participating in an IETF process is to continue
>to develop this protocol in a forum where we get maximal involvement
>from the internet community.
>
>f. HyperGrid is pretty darn'd cool, IMHO. and there's nothing in it that
>would eliminate it from consideration by this group. I believe John
>Hurliman is trying to define it in draft form to be considered at the
>upcoming face to face in san francisco. even if we don't have a HG draft
>for that meeting, i am confident it will attract considerable attention
>on the mailing list, and honestly, we're months (if not years) away from
>turning any internet drafts into RFCs.
>- if you've been following the HG discussion, you may have seen talk
>about the difference in trust models between OGP and HG. but... if we
>can create an abstract trust model that allows implementers to pick
>"OGP" or "HG" trust, i think we could easily define the wire protocol
>for both traditional OGP and HG users.
>

I don't know where you got the idea that I'm creating a HyperGrid proposal, but I don't think anyone is pretending that merging two administrative domains with some clever backend coding is interoperability. See my post at http://software.intel.com/en-us/blogs/2009/02/18/approaching-virtual-world-interoperability/ for the longer explanation of why OGP and HyperGrid are the same solution with different default settings, and why neither actually address interoperability.

>
>so... to recap...
>
>* we're about application layer wire protocol
>* we're about interoperability between current or near future
>implementations
>* we're NOT blessing Second Life's existing protocol as an official
>standard

That's not the question anyone is asking. The important question is: "Is the scope of MMOX limited to OGP, or mandating the discussion of OGP? Are virtual world interoperability proposals that are not compatible with OGP outside of the scope of MMOX?" We can either stop talking about MMOX as a working group for OGP, or put it in writing that that's the case and people who are not interested in OGP can save themselves the trouble.

>* it's entirely possible application developers using MMOX protocols
>would need to define an "interoperability profile," which defines how
>MMOX applies to their application.
>* in other words, "we define mechanism, not policy."
>
> ...