Linking meshes changes land impact?

[Publik]

So I've got a small mix of meshes and prims (4:3). Two of the meshes are 13 and 14 PE, the other two are 1 (13+14+1 * 2 + 1 * 3 = 32). When I link them, it jumps from that 32 up to 65LE.... What?

[Drongle McMahon]

When you link ordinary prims to meshes, the whole linkset, including the ordinary prims, are evaluated by the new land impact system. This can increase (or sometime decrease) the contribution of the ordinary prims to the linkset land impact. The changes can be very large. As long as the ordinary prims are not linked to meshes (and have "Prim" physics shape type), they will be evaluted by the old familiar system.

[Kwakkelde Kwak]

---

Drongle McMahon wrote:

[...] This ca increase (or sometime decrease) the contribution of the ordinary prims to the linkset land impact. [...]

---

 

Already working on the table for this Drongle?

[Drongle McMahon]

"Already working on the table for this Drongle"

I can't think of a useful one. The download and physics costs of the linked prims are too variable and too dependent on torturing and size. I did have a note somewhere about how far you could go before the download cost reached 0.5 and 1.0 for some sizes. Maybe I will visit that again. :matte-motes-smile:

[Publik]

That doesn't quite make sense to me, but whatever

[Kwakkelde Kwak]

It's not that hard basically... if you link something simple like a standard box to a mesh it will cost less than one prim, if you link something with lots of faces like a torus to a mesh it will cost (a lot) more than one prim.

[Publik]

Oh, so it's treating the prim as a mesh when linked to a mesh? Still doesn't make any damn sense, but I can live with it. I'd rather keep my stuff as one happy object than break it up like that, but whatever

[Kwakkelde Kwak]

Yes typical how the Lindengods have done this, it's a great idea, primcost according to render/serverweight, but only if there's a mesh in the set. That's just asking for very very clumsy linksets...It does show you however how inefficient some prims really are, but I don't think it will do any good. People will link their boxes to meshes and the rest to eachother, only adding to the data traffic......

edit: or even worse people will unlink a set on a full parcel and they will find it in their inventory in pieces because the seperate parts cost more than the set...

[Publik]

Yeah, and determining new weighting shouldn't be that hard...Something like:

list linkset = //whateverinteger weight = 0;// loop through set{    if(linkset[index] == MESH)    {        weight += // weight of mesh    }else // PRIM and MESH    {        weight += 1;    }}

 Do sculpties calculate differently as well?

[Pamela Galli]

My understanding is that whatever is linked to mesh, is counted as mesh is.

[Drongle McMahon]

"Do sculpties calculate differently as well?"

Just like standard prims, if sculpties are not linked to mesh and have physics shape type "Prim", they are evaluated, for parcel capacity consumption, as they always were. Otherwise they are evaluated accounting using the new system, in which case they will (generally) be very expensive. It's not a good idea to lin mesh and sculpties, or to try to use physics shape type "None" on sculpties (which is a pity).

[Publik]

---

Drongle McMahon wrote:

"Do sculpties calculate differently as well?

"

Just like standard prims, if sculpties are not linked to mesh and have physics shape type "Prim", they are evaluated, for parcel capacity consumption, as they always were. Otherwise they are evaluated accounting using the new system, in which case they will (generally) be very expensive. It's not a good idea to lin mesh and sculpties, or to try to use physics shape type "None" on sculpties (which is a pity).

---

IE you get better impact just tossing the physics and using prims for collision boxes? So effing lame. I can't even fathom why they thought this made sense. It actually solves a little problem I had, since the collision boxes seem off anyways. I've got a building made of mesh (the one I OP'd about) and the collision floats you a 1/4-1/8m off the floor. Does a lot for my confidence in LL, not that it was too high to begin with.

[Kwakkelde Kwak]

---

Publik wrote:

Yeah, and determining new weighting shouldn't be that hard...Something like:

list linkset = //whateverinteger weight = 0;// loop through set{    if(linkset[index] == MESH)    {        weight += // weight of mesh    }else // PRIM and MESH    {        weight += 1;    }}

 Do sculpties calculate differently as well?

---

 

Yes they could have left the "old" way of counting intact, even with a mesh in the linkset. That shouldn't be hard. The thing is I agree it's fair to do it the way they do it now, with SL prims represented by their actual weight. I also think they can't change the rules on old objects, so linksets with only prims still count as 1 erm... prim per prim. The combination of the two doesn't work though, as I said, it's an open invitation to put more pressure on the grid by linking everything under 1 prim equivalent to a mesh and everything over it to eachother. Maybe it's an idea to leave existing objects untouched and apply the new rules to new linksets. However this means most SL prims will become useless. A small torus for example would cost something like 100 prims.

Maybe what they did is a first step in this direction, letting us see what old prims cost. The future will tell...

 

@Drongle who said it's not a good idea to link sculpties and meshes....

It's not a good idea to even USE sculpties:)