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

[Argent Stonecutter]

On 17-Oct-2007, at 13:15, Andre Roche wrote:
> I understand the basis of the suggestion, and for prims that span
> multiple regions, the largest of them, that would be a workable
> compromise.  For the other prims, the more used ones, that wouldn't
> solve anything.

The only absolutely unreplacable megaprims are larger-than-10-meter  
curved objects, like spheres and cylinders for walls and domes, and  
they are already required to be phantom to work. All other megaprims  
can be replaced by multiple prims.

> There is no "solution" here in my opinion, only compromises.  SL
> should be as free as possible from complication and compromise as is
> feasible.

Which is why I am suggesting the simplest possible compromise, in my  
response to Callum:

>> You can continue to treat them ALL as cubes, as the physics engine  
>> currently does, but then you get the inadvertent invisible  
>> obstruction problem.
>>
>> Perhaps the thing to do is to have them ignored UNLESS they're  
>> cubes with no twist, profile cut, taper, or sheer. Path cut should  
>> be OK because that is a simple transform that the system already  
>> models correctly.
>>
>> OK, then:
>>
>> Have an option to remove restrictions.
>>
>> If a prim > 10m [and <54m, per Kelly Linden's comment on the  
>> maximum linkset size] is a cube with simple geometry, it is  
>> treated as such.
>>
>> Otherwise, it's ignored by the physics engine.