[sldev] Optimus Prim vs. Megatrim: "The Big Prim Problem"
Dzonatas
dzonatas at dzonux.net
Wed Oct 17 11:37:24 PDT 2007
Any form of dyslexia is a good argument that we really need to do away
with textual syntax for computer program languages and go with more
graphical or iconic symbols! D'oh!
Dzonatas wrote:
> Argent Stonecutter wrote:
>> Just thinking of the obvious ways this could break makes me want to
>> run screaming.
>
> After you run wild and scream away, just list them. We can address
> the concerns and break them out into the appropriate components.
>
>
>> And I'm saying it shouldn't be exposed outside the sim itself. There
>> is no reason this kind of optimization - breaking a single prim or
>> linkset (they're the same thing as far as the the physics engine is
>> concerned, as far as I can tell) into multiple objects for the
>> physics engine - should require even transmitting the fact that this
>> optimization has taken place to the viewer, or to other sims.
>
> Consider SHARE for an example, perhaps that functionality should not
> have been exposed at all outside the core mainframe and related
> developers. However, developers wanted to open it up to allow other
> components to interoperate with it, and it became what it is known today.
>
> The same feature basically applies here. It is a typical compiler
> issue. We expose it in order to improve it. That improvement may
> entail further automation.
>
> For example, look at sculpties. The sculpt map is exposed. That really
> should not be exposed in the same way you suggest "it shouldn't be
> exposed outside the sim itself". However, look at all the different
> tools created to build sculpties because it has been exposed. Sculpt
> maps are created by tools that automate the process.
>
> That is the same ability I suggest that we should do. I figure
> optimized prims will as perfect in the same way that sculpt maps are
> not perfect, but that imperfection should not stop from being able to
> expose the functionality. Complex prims can be created by an automated
> process in the same sense sculpt maps are now created. It may even
> give the ability do have complex and concave physical topologies that
> the sim cannot render on-the-fly. Such compilation (the protoform,
> transform, and other object optimizations) can be done.
>
> Objects that are not complex and not concave probably can be rendered
> on-the-fly at the sim level as you suggest. However, once rendered by
> one sim, we should allow another sim to gather that same data so it
> does not have to render redundantly, which can get very costly as it
> does now.
>
>
--
Power to Change the Void
More information about the SLDev
mailing list