In the Content and Mesh User Group some interesting things came up today.
If a mesh object is transparent should that change its Land Impact cost? Should it even render? If it doesn’t render, doesn’t it use fewer resources? It seems the Lindens are looking at this special case and considering how to implement such a change.
Runitai says the code to NOT render 100% transparent objects is in the pipeline. The release of the code has been delayed because it broke highlight transparent. (SH-2963)
Dropping Mesh Things
If you have a mesh attachment, you cannot drop it as you would a prim or sculpty attachment. Mesh objects rigged or not, detach back into inventory.
Nyx Linden says, ‘Unfortunately we don’t have a good way of enabling that [dropping mesh] without breaking things right now.’
Some have tried using a prim as the root of a link set to see if that would allow them to drop mesh attachments. It doesn’t. So, it seems this is not going to be possible for some time.
If you are wondering why people would want to drop mesh, think of throw-away things used in role play. Medhue Simoni uses a mesh brain as zombie food. Eck! The brain would be partially eaten and dropped. Eventually the brain would de-rez itself and disappear.
Davido Chrome brought up the idea of a water material that would allow one to make lakes at elevations other than sea level. So, think of a lake up in the mountains. Another possible use is as water in a spa or bathtub.
Somewhere that is a feature request for a complete and arbitrary water simulation system. This would allow swimming, surface ripples, and reflection. Not a simple thing to do. I suspect in the Linden priority system that assigns less priority to complex tasks a water system rates petty low. We won’t see this any time soon.
Runitai Linden points out that creating multilevel water plans using the existing system is non-trivial… way complex.
Seems Havok, SL’s physics engine, has support for objects with primitive physics shapes; box, cylinder, and spheres. These primitives are predefined mathematical representations in Havok that it is optimized to handle.
If the Mesh Importer could use the Collada definition for the Havok primitives, we might be able to improve vehicle performance. We might also be able to reduce some of the streaming cost of mesh.
It is unclear how much of a change this would require to the existing SL system. Some think much of what is needed is already built into SL.
It seems the Havok primitives is something the Lindens may consider implementing.
There are some problems with mesh objects that need to be physical. Think bowling pins. The above image gives an example of a mesh pin and a sculpty pin. The mesh pin is on the left. The sculpty is on the right. A JIRA is being filed and the Lindens will be looking at the problem.
For the past two weeks SL has averaged well over 1400 objects uploaded per day to the main grid. The all time high is 1540 for a single, which was the day after mesh was released. All the beta testers were going for it. Over the last month the rate has increased from 1,000/day to near 1,500/day. Charlar says the trend line is steady.
Mesh viewer stats:
- Over 70% of residents use a mesh capable viewer.
- Over 63% of sessions are with a mesh capable viewer.
- Over 67% of time logged in is with a mesh viewer.
Not too much exciting stuff. But, things are getting worked out and improving.