Interesting thoughts. Currently they are configurable. Though, there's nothing to say that material rubber has different settings then material plywood.
There's also no defined procedure for changing the settings. One of the reasons being... it's something that's partially mathematic and partially qualitative (does this object react well to being hit with a train?). I believe that all current LSL scripts that are worth their weight in $L get the mass of the avatar first.. and multiply the intended forces by it to get the 'effective' force.. so you're right in that the default density of avatar makes no difference to scripts. The PID controller, however, is specifically tuned. I'm sure there have been some people who have changed some of the properties and saw weird things happen like avatar go through walls when flying at high speeds in certain directions or falling through objects as a result. Anyway, I think what ckrinke is advocating here is a group of physics settings based on material. Regards Teravus On 3/22/09, Dahlia Trimble <[email protected]> wrote: > I'm trying to imagine the consequences of having the wrong density for an > avatar. The most common physics application that I'm aware of for avatars > are flight assists, which may be less common in OpenSim as there is no > flight ceiling. Another application may be push type teleporting systems. > Applications which require accurate physics usually use the results of > llGetMass() as a term for any calculations which could lessen the need for > accurate default density. > > It is a problem that many of the values for physics are empirically derived, > and that there are so many variations across implementations, such as the > differences between 32 and 64 bit systems and between different 64 bit > systems. This suggests that any deviation from the established values may > need to be a configuration option (if they aren't already), and that some > experiments may need to be designed and carried out to determine optimum > values. Perhaps, if all of the important values would be user-configurable, > a procedure could be developed for calibrating the physics engine for any > given installation? > > > > On Sun, Mar 22, 2009 at 3:03 PM, Charles Krinke <[email protected]> wrote: > > > > > > > > > > We have m_density defined in two different places with two different > values. > > > > One is in: > ...\OpenSim\Region\Physics\OdePlugin\ODECharacter.cs - (72, > 22) : > > > > public float m_density = 60f; > > > > Another is in: > ...\OpenSim\Region\Physics\OdePlugin\ODEPrim.cs(152): > > > > > private float m_density = 10.000006836f; // Aluminum g/cm3; > > > > And in looking at the usage to calculate mass, it looks like the math is > right, but yet, the avatar seems to have a higher density then the prim. > > > > I know that these are empirically determined, but in trying to look > forward to a world where we actually might use the density of metal, rubber, > glass, rock, plywood to do some additional physics *stuff*, it seems that we > might want to see if we can harmonize our notion of density a little bit. > > > > Perhaps I am just missing the boat here. But, I would like to understand > how we might be able to move forward a little bit on density some time. > > > > Charles > > _______________________________________________ > > Opensim-dev mailing list > > [email protected] > > https://lists.berlios.de/mailman/listinfo/opensim-dev > > > > > > > _______________________________________________ > Opensim-dev mailing list > [email protected] > https://lists.berlios.de/mailman/listinfo/opensim-dev > > _______________________________________________ Opensim-dev mailing list [email protected] https://lists.berlios.de/mailman/listinfo/opensim-dev
