[sldev] Optimus Prim vs. Megatrim: "The Big Prim Problem"

[Callum Lerwick]

On Wed, 2007-10-17 at 10:18 -0700, Kelly Linden wrote:
> > You need to perform collision detection on megaprims. The fact that the
> > megaprim itself is not physical however should prevent the need to merge
> > "simulation islands" in the physics engine, which was previously cited
> > as a major performance concern.
> >   
> This last bit is not true.  Large prims, even if they are "static",
> still effect the simulation islands.

You're being unclear. Yes they "affect" them, that's the whole point. My
point is, say you have a long, nonphysical horizontal 2x4, with an SI on
each end:

   SI                                     SI
--------------------------------------------------

While both SI's are interacting with the same stick, since the stick
itself is non-physical, there should be no need to merge the two SIs.

If your physics engine is doing so, then your physics engine's early
culling sucks^H^H^H^H^H is less than optimal. :P

Note that the long 2x4 doesn't even have to be "megaprim" size.

Considering the AWG, the last thing we need to be doing is letting low
level implementation detail dictate arbitrary limitations upon the upper
layers. If you do so, you end up with crap like IMAP.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.secondlife.com/pipermail/sldev/attachments/20071017/926cc476/attachment.pgp