Re: [mmox] Permissions (Gareth Nelson)

Gareth, although you've withdrawn your proposal, it addresses interesting
issues so I'd like to make a brief comment on your last point:

On Tue, Feb 24, 2009 at 6:40 AM, Gareth Nelson <gareth@litesim.com> wrote:

> Essentially, you'd need to look at each sub-component and then
> evaluate what the restrictions are, applying the restrictions on
> sub-components to the collected whole (since they're part of the
> whole).

While such examination is just about feasible at interop time in SL's
non-hierarchical scheme for objects, it is totally impractical for the
hierarchical objects that would be the norm in a world without those highly
SL-dependent restrictions on object composition.

An object will in general be composed of an arbitrary number of other
objects arranged in an arbitrary tree, just like objects are composed of
smaller components in the real world.  This is the basis of engineering and
the key to riding on the shoulders of giants.  "Arbitrary number" could be
hundreds for virtual objects with current-day levels of complexity, or
thousands or more as our means of object production improve.

Consequently, such piecemeal examination of permissions cannot be done
realistically at interop time, but must be done at object construction time
instead, propagating permissions up the composition tree as required.  All
we then need to do at interop time is to look at the root of the object tree
and no deeper.

Hierarchical objects are rather central to the success of the virtual object
concept in VWs, and they are a linch-pin for the expansion of virtual
economies with more complex technology too, so they're coming.  More on this
from the SL perspective is available here:
https://wiki.secondlife.com/wiki/Prim_and_Object_Hierarchy .

Morgaine.

On Tue, Feb 24, 2009 at 6:40 AM, Gareth Nelson <gareth@litesim.com> wrote:

> Essentially, you'd need to look at each sub-component and then
> evaluate what the restrictions are, applying the restrictions on
> sub-components to the collected whole (since they're part of the
> whole).
>
> i've actually withdrawn my proposal.
>
> On Mon, Feb 23, 2009 at 8:25 PM, William George
> <wjgeorge@dceo.rutgers.edu> wrote:
> > Nice first step, but it's not clear how this would handle composite
> objects:
> > objects that have two different creators and/or permissions settings. The
> SL
> > convention of using the root prim's permission for the entire object
> leads
> > to restrictions concerning the reuse of "no mod" sub assemblies in a
> larger
> > assemblies.
> >
> >
> > _______________________________________________
> > mmox mailing list
> > mmox@ietf.org
> > https://www.ietf.org/mailman/listinfo/mmox
> >
> >
> _______________________________________________
> mmox mailing list
> mmox@ietf.org
> https://www.ietf.org/mailman/listinfo/mmox
>