Re: [mmox] OGP scalability concerns

[Morgaine <[e-mail]>]

As I see it, a good part of the reason for the tangle surrounding the word
"interop" seems to be that people are focusing on the *mechanisms* that they
believe it requires, and these mechanisms are often architecture-dependent
so there isn't much immediate agreement.

In contrast, if we look at *user expectations* then the situation becomes
much simpler and clearer.

Users already know what a single virtual world looks like, and they also
know what it means to move or teleport between different regions of that
single world.  So, when they hear that other worlds are out there and that
we are offering *interop* between old worlds and new, they know exactly what
to expect:  *the same as before, but extended to regions in the new worlds
as well*.

For our purposes in this protocol group, the above user expectation is
actually very easy to turn into a definition that has useful structural
properties, which I've stated before.  Full VW interop is, in essence:


*To create a mashup for the user, in which the user's avatar and objects
from one world are perceived to be in a new world, and are also perceived by
others in that new world.
*


Despite this interop goal being expressed in pure user terms, it's
remarkably structural too:  it tells us immediately that for successful
interop, we need to gather together data from many different sources (either
worlds or services employed by those worlds), otherwise that mashup will not
occur and user expectations will be dashed.

I think it's a good sign that everyone who has commented on this user view
of interop before has agreed with it --- it probably means that we actually
have the same goal in mind, but just seem to be approaching it from
different angles.

Lisa also provided a persona-oriented use
case<http://www.ietf.org/mail-archive/web/mmox/current/msg01275.html>that
underlined the social dimension of interop, and added one important
element to the picture:  *continuity* in social elements, such as names,
friends, places, and easy travel or access, as well as visual (or at least
perceptual) continuity.  The more seamless we can make the user experience
of interop, the higher will be our degree of success.

I think we actually know what *interop* is, from a user perspective.  If we
start from there, it might help us to converge on a common technical view as
well.

Morgaine.