Re: [ogpx] VWRAP Draft Charter: 2009 08 28 revision

[Morgaine <[e-mail]>]

I support Carlo's request that we define the terms that we use before we use
them.

As Carlo's says, undefined terms should not be used in the charter or in the
group name --- it introduces uncertainty and confusion.  I would add that *even
less* should undefined terms appear in our protocol specifications, which
are relatively precise documents of interest to implementors.  If we don't
know what a virtual world is, the term should be abandoned in favour of one
that we do understand.

Carlo offers one way of defining the topological entities in our problem
space, based on size and administration, which sounds quite viable.  A
meaning of virtual world might be agreed in this way.

As an alternative, I propose a simpler and much more flexible approach:
define virtual worlds as anything that implements the required endpoints of
the VWRAP protocol.  We do not need to know anything else about them.  This
lets the people who actually design the virtual worlds decide what
constitutes a virtual world.  The only thing that interests us is that these
"virtual worlds" implement the VWRAP endpoints.

This is highly workable in practice.  After all, Linden Lab knows that SL is
a virtual world, they don't need it defined.  Likewise, the owners of OSgrid
know that OSgrid is a virtual world too, they don't need it defined.  As a
result, a definition of "virtual world" by an IETF workgroup will have
little practical effect outside of the specifications, because the
implementors will design their worlds as it suits them.  If they are to
interoperate through VWRAP however, then both LL and OSgrid will ensure that
they implement the protocol correctly, no matter their internal differences.

Note that this is the approach taken in a large number of IETF protocols:
the implementation behind the endpoint is just an *implementation detail*,
and is of no consequence as long as it implements the protocol correctly.
Precedent suggests that this is the right approach.

I believe that this is simpler, more powerful, and much more flexible than
defining "virtual world" on the basis of topology.  All we need to know is
that VWRAP endpoints are available, and not what kind of entity actually
implements them.

Morgaine.








===================================

On Mon, Aug 31, 2009 at 12:08 AM, Carlo Wood <[e-mail]> wrote:

> I think that any effort to work on the final wording
> should be stopped until we finally agree on what "virtual world" means...
>
> I understand that you don't want to define what virtual world means,
> in which case you shouldn't use it at all in the charter and also
> not in the protocol name (VWrap).
>
> There are several levels of 'locations' in an SL-like whatever:
>
> First, the words that are already in use by the community:
>
> 1) simulator
> 2) estate
> 3) region
> 4) land
> 5) island
> 6) virtual world
>
> Then the locations that we need terms for to be able to even
> TALK about them:
>
> A) A smallest partition
> B) A collection of adjacent A's run by a single administration
> C) A collection of adjecent B's run by different administrations (which
> very likely use the same TOS etc)
> D) A collection of C's that are not adjacent but still fall under the same
> TOS etc.
> E) A collection of D's that have totally different administrations and
> possibly different TOS etc, but which interoperate.
> F) The whole of all E's that do not interoperate, but still use VWRAP.
> G) The rest that use the term "virtual world", but do not use VWRAP.
>
> MMOX and WoW is part of G. I consider that not relevant as it should be
> clear by now what is part of G and what not.
>
> Meadhbh, do you disagree with my list(s)?
> Anything you want to add, or change my rough definitions of things?
>
> Please don't reply with "we don't want to define this" :p
> We NEED to define this, or we can't TALK about this!
>
> --
> Carlo Wood <[e-mail]>
>
> PS I hope it should be clear by now: the problem we are having
>   is that you call 'E' "a virtual world", while Morgaine, Kari,
>   me and others want to call 'D' "virtual world". Moreover,
>   you do as if D or E do not exist, so that we will not
>   be able to talk about them separately later on :/
>
>   You might convince me to call 'E' "virtual world", but
>   then we need a well defined term for 'D' in order to
>   communicate about D, too.
>
>   In my previous post I proposed to use "Galaxy" for E.
>
> _______________________________________________
> ogpx mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/ogpx
>