Re: [mmox] The Story So Far...

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

>-----Original Message-----
>From: Meadhbh Hamrick (Infinity) [mailto:[e-mail]]
>Sent: Tuesday, February 17, 2009 3:42 PM
>To: Hurliman, John
>Cc: [e-mail]
>Subject: Re: [mmox] The Story So Far...
>
>
>On Feb 17, 2009, at 3:09 PM, Hurliman, John wrote:
>
>>> While there is plenty to do and discuss in order to develop OGP
>>> into a
>>> fully workable protocol, we think it represents an initial shared
>>> vision
>>> of an interoperable protocol and hope it can evolve and grow in the
>>> context of this group.
>>>
>>> - Mark Lentczner
>>>
>>> [1] http://wiki.secondlife.com/wiki/Open_Grid_Protocol
>>
>> It's great to see the amount of work that is being brought to the
>> table by Linden Lab and the Linden Lab Architecture Working Group
>> (AWG) participants, and I hope as much of it as possible can be used
>> to build an MMOX standard. I've been working on an LLIDL
>> implementation among other things based on what Linden Lab has
>> submitted as draft material.
>
>awesome!
>
>> It's my understanding that Open Grid Protocol (OGP) is a Linden Lab
>> project; part of the Linden Lab AWG which had no active contributors
>> that were also active developers of libSecondLife/libOpenMetaverse
>> or OpenSim during the time the OGP protocols were designed. The
>> "agent domain" concept in OGP terminology brings in a host of
>> assumptions that are diametrically opposed to the goals of
>> libOpenMetaverse and OpenSim (supporting millions of independent
>> administrative domains and service providers).
>
>huh? there were a number of non-lindens who participated in the
>development of OGP. the AWG does not, nor did it ever disallow
>participation by libSL or libOMV contributors / committers. last
>summer we demonstrated interoperability between Linden Lab's servers
>and an OpenSim instance. we hope that members of the libSL and libOMV
>teams will participate in MMOX.
>

I said "AWG had no active contributors that were also active developers of libSecondLife/libOpenMetaverse or OpenSim during the time the OGP protocols were designed". The participation by individuals that were not Linden Lab employees does not mean that OGP is a joint effort between Linden Lab and the libOpenMetaverse or OpenSim projects, or that the OGP as it is drafted has support from the libomv/OpenSim communities.

>the concept of an agent domain was defined explicitly to support many
>different independent administrative domains. that is... though the
>interoperability test last summer used the Linden agent domain, the
>specification does not mandate a single agent domain. in fact, i
>believe Christian Scholz (a non-linden) implemented his own Agent
>Domain based on code from the PyOGP project. There is also frequent
>mention in AWG meetings of an upcoming C# implementation of an Agent
>Domain, again by a non Linden-Lab employee.

The agent domain concept takes several independent services (identity, presence, inventory) and lumps them together into a single trust domain. This inhibits the development of a widely distributed ecosystem like we find on the web today, and is really only convenient for large virtual world service providers that already have a vertical stack of services. It also drags in the assumption that there must be a backend communication layer between services such as inventory and simulation regions.

John