Re: [ogpx] one virtual world, or many?

[Joshua Bell <[e-mail]>]

I apologize for the potentially unclear responses. These are decidedly
non-trivial questions being asked, and demand an attempt at rigorous
answers. Saying "yes" would imply policy decisions and extensive technical
work by independent parties. Saying "no" would incorrectly imply a
limitation on the scope of the intended work.

On Sun, Aug 30, 2009 at 6:19 PM, Charles Krinke <[e-mail]> wrote:

> I'm confused with all that is going on and need to ask my two main
> questions again.
>
> "Are we considering interop with virtual worlds running OpenSim which do
> not run agent domains but rather a similar, but not quite identical notion
> we call UGAIM (User, Grid, Asset, Inventory, Messaging) servers?"
>

It's unclear what you mean by "interop with" in the above sentence. What are
the end-points involved? What is the data and/or user experience being
shared?

And, of course, a protocol specification doesn't dictate implementation in
any way. Second Life today doesn't implement the parts of OGP that have been
drafted yet, and of course there is a lot more to be drafted! I couldn't say
what OpenSim will do in the future!

However, I can take a guess at your intent, but let me state my guess:

"Will the output of the VWRAP group be protocol definitions that would prove
useful for implementing movement of agents between OpenSim instances?"

To that question, I'd answer: "Hopefully, yes."


> "Are we considering a full handoff? That is, a user on an OpenSim grid may
> teleport to a SecondLife grid and retain its connection to the OpenSim
> UGAIM, in an analogous fashion that a user has been demonstrated to teleport
> from the Betagrid to OSGrid using OGP last year?"
>

First off, any such interoperation is subject to the whims of policy.

I'm unfamiliar with the details of OpenSim UGAIM, but I would posit that the
VWRAP group would define protocols that an OpenSim instance would need to
implement to support such interop with a VWRAP-implementing non-OpenSim
system. I believe UGAIM would be an "implementation detail" hidden behind
the protocol facades. This may imply that the properties of UGAIM (again,
I'm ignorant of any details) will be beneficial to the specification of the
protocols.

The worst-case analogy that comes to mind is taking a pre-web server
application and slapping a web interface (HTTP+HTML) on it. Depending on the
design of the server application, that may be trivial and clean, or may be a
hideous hack. In this case, we have the advantage that the server
application engineers are active (it's not a defunct legacy system!) and can
inform the creation of the protocol to potentially ease the migration.

Joshua