Re: [ogpx] OGPX Draft Charter, 2009 08 25 edition

[Infinity Linden <[e-mail]>]

On Thu, Aug 27, 2009 at 4:02 AM, Carlo Wood<[e-mail]> wrote:
> On Tue, Aug 25, 2009 at 02:18:44PM -0700, Infinity Linden wrote:
>>                               Within  a single virtual  world, avatars
>> exist in  at most one location  in a shared  virtual space.
> ...
>> Adjacent locations  in virtual worlds accessible by  this protocol may
>> be   explicitly   partitioned  into   "regions"   to  facilitate   the
>> computational  and communication load  balancing required  to simulate
>> the virtual  environment. Such virtual  worlds may consist  of regions
>> administered  by distinct organizations.
>
> I see you insist on using a definition for "virtual world" that does
> NOT mean "different administrations", but rather means "all reachable
> regions".

yes. as discussed earlier, "virtual world" is a non-normative
definition, and defined from the user's point of view. this definition
also allows us to say that "OSGrid is a single virtual world, and not
a collection of adjacent virtual worlds."

>
> However,
>
>> Though these  virtual worlds
>> may  be partitioned,  they  remain "un-sharded;"  all inhabitants  and
>> objects  in a  particular location  in  a virtual  world may  initiate
>> interaction with  all other inhabitants and objects  in that location;
>> and, service  endpoint addresses  refer to at  most one  location. The
>> state of  a virtual  world is independent  of the  client applications
>> that access it and may persist between user sessions.
>
> this paragraph reveals the actual (or additional?) view of what
> "a virtual world" is:
>
>  A virtual world are all regions that collaborate under a single avatar domain.
>
> Correct me if I'm wrong, but from this I conclude that your vision of
> a virtual world is a collection of regions that all use a single point
> (and administration) for authentication of a users avatar (and inventory).
>
> For example, one logs in with the agent domain run by LL and than can
> move between regions and IM every other avatar in "the virtual world"
> (using the handle of this agent domain provider). If one would prefer
> to login with a different agent domain provider, it would not be possible
> to IM every other avatar, or have other interactions  such as 'giving'
> inventory objects to other avatars, and thus one would not be able to
> be part of the same virtual world.
>

yes. kind of. there is no reason a region may limit the number of
agent domains it accepts to one. so as i said earlier, you could have
agent domain A that is trusted by region domains 1, 2 and 3 and agent
domain B that is trusted by region domains 3, 4 and 5. Avatars
representing users from agent domain A and agent domain B can meet in
region domain 3.

so, some might say that region domains 1 through 5 are the same
virtual world. others might say that domains 1,2 and 3 and 3, 4 and 5
define distinct virtual worlds.

ultimately it doesn't matter to the protocol because the term "virtual
world" is not used in the protocol definition. or rather, it's not
used in a normative fashion.

there is no requirement in the protocol that asset services
(inventory) be offered by the same administrative entity that offers
the user authentication service. this is to support cable beach (or
anyone else who wants to decouple user authentication from asset
management.)

the protocol does not REQUIRE Linden Lab (or any other region domain
provider) to choose only a single agent domain to trust. however, it
does not REQUIRE that they trust multiple agent domains. it assumes
that the region domain will establish a policy of which agent domains
it trusts (or even a permissive policy where it will accept protocol
interactions from all agent domains.

since this is a policy issue, we don't mention it in the charter.

that being said, Linden Lab has a number of policies regarding
acceptable behavior in it's virtual world, so i would imagine that by
policy, Linden would require a foreign agent domain to agree to
enforce those policies as a precondition to allowing avatars defined
by those agent domains to rez in regions it administers. but again,
it's policy, not protocol.

which is why we have the next paragraph.

>> Regions and  Services implemented according to  the specifications may
>> be deployed by separate  organizations with varying policies and trust
>> domains.  The OGPX  protocols will  provide the  mechanisms  for these
>> virtual world  services to interoperate, when permitted  by policy and
>> shared trust domains.
>
> This paragraph is OK.
>
>> The protocol  should describe interaction semantics  for these virtual
>> worlds, independent of  transport, leveraging existing standards where
>> practical. It  should define interoperability  expectations for server
>> to server  interactions as well as  client-server interactions. Though
>> the  protocol  is  independent  of transport,  early  interoperability
>> trials used HTTP(S) for non-real-time messages. The working group will
>> define specific  features that must be replicated  in other transports
>> and  will  define  the use  of  HTTP(S)  as  a transport  of  protocol
>> messages.
>
> OK
>
>> Foundational components of the protocol include the publication of:
>>
>>   * an abstract  type system, suitable for  describing the application
>>     protocol in an implementation neutral manner,
>>
>>   * a   security   model   describing  trust   relationships   between
>>     participating entities,
>>
>>   * guidelines   for   the   use   of  existing   authentication   and
>>     confidentiality mechanisms,
>>
>>   * an application-layer  protocol for establishing  the user's avatar
>>     in a virtual world,
>>
>>   * an application-layer  protocol for moving a  user's avatar between
>>     adjacent and remote locations in a virtual world,
>>
>>   * format descriptions  for objects and  avatars in a  virtual world,
>>     and
>>
>>   * an   application-layer  protocol   for  identifying   agents,  and
>>     requesting information about them.
>>
>> The protocol  defined by this  group will carry information  about the
>> virtual  environment,  its contents  and  its  inhabitants.  It is  an
>> application layer protocol,  independent of transport, based partially
>> on these previously published internet drafts:
>>
>>   * http://tools.ietf.org/html/draft-hamrick-ogp-intro
>>   * http://tools.ietf.org/html/draft-hamrick-llsd
>>   * http://tools.ietf.org/html/draft-hamrick-ogp-auth
>>   * http://tools.ietf.org/html/draft-hamrick-ogp-launch
>>   * http://tools.ietf.org/html/draft-lentczner-ogp-base
>>   * http://tools.ietf.org/html/draft-levine-ogp-clientcap
>>   * http://tools.ietf.org/html/draft-levine-ogp-layering
>>
>> Goals and Milestones:
>>
>>   * October  2009   "Introduction  and  Goals"  to  the   IESG  as  an
>>     Informational RFC
>>
>>   * October 2009 "Abstract Type System for the Transmission of Dynamic
>>     Structured Data" to the IESG as Proposed Standard
>>
>>   * October 2010 "Foundational Concepts and Transport Expectations" to
>>     the IESG as Proposed Standard
>>
>>   * February 2010 "Guidelines for  Host Authentication" to the IESG as
>>     an Informational RFC
>>
>>   * February  2010 "Service  Establishment"  to the  IESG as  Proposed
>>     Standard
>>
>>   * February 2010  "Client Application Launch Message" to  the IESG as
>>     an Informational RFC
>>
>>   * February 2010  "Simulation Presence Establishment" to  the IESG as
>>     Proposed Standard
>>
>>   * June  2010  "Primitive Object  Format"  to  the  IESG as  Proposed
>>     Standard
>>
>>   * June 2010 "Digital Asset Access" to the IESG as Proposed Standard
>>
>>   * June 2010 "Entity Identifiers" to the IESG as Proposed standard
>> _______________________________________________
>> ogpx mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/ogpx
>
> --
> Carlo Wood <[e-mail]>
>