steph Arnott Not all tortured prims are bad. The ones that greatly increase the polygon count, like twist, are bad. Hollow is not bad, and simply letting objects be a bit hollow isn't any worse.

steph Arnott The primary reason the LI goes up is the physics - set it to Convex hull and LI drops dramatically

You clearly do not understand what 'convex hull ' does, it is an outdate kludge and you can not hollow more than 95%

steph Arnott Steph, you can easily check this by creating a sphere, hollowing it by various amounts, and see if the Li changes at all. It does not. You state a sphere cannot be hollowed more than 95 percent. Creators on Opensim would disagree with you, as they've been doing so for years. On Second Life of course 95 is currently the limit... which is why I posted this feature request. Spheres (and other prims) can go to 99 with no change in Li at all.

MIght also mention that tortured prims really don't affect much. They did at one time. In the early days of SL tortured prims were murder on lag. LL fixed that years ago, and now tortured prims have very little impact.

To double check this, I created a sphere and hollowed it. The Display reading (viewer impact) went from 459 to 534, but remained at 534 regardless of how much it was hollowed. No other reading changed. I then twisted it both directions and cut it in half. The Display reading stayed at 534 (torturing had no effect on Viewer-side graphics). The physics readings were all over the board and were highly inconsistent, but had zero effect on the server readings. Changing the object between prim and convex hull had no effect on anything but physics, which brought the object back down to .2. But do physics even matter for non-physical objects?

As with all things on SL, these things depend largely on the builder, the tricks they know, and how they build the object.

When I ran the same test on Opensim the physics readings remained constant, changing from .1 to .2 and sticking there. I don't know why the SL readings are so extremely inconsistent, but that's SL and indicates there is something wrong with the physics measurement reading, not the physics themselves.

So tortured prims have very little impact, certainly no worse than a mesh shape. And an object with 99 hollow will have no different effect on SL than a solid object. A solid sphere measured .1 physics, as did a fully hollowed sphere, on both SL and Opensim.

99 is impossible, simple as that and you can shout at the stars and it still will be impossible

steph Arnott Wow. You're still shouting impossible... when Opensim has already proved it's quite possible... and has been doing so for years. I was a professional coder for over 35 years in RL. When someone tells me something is "impossible" with computers, my answer is they don't know what they're talking about. If this was "impossible" Spidey would have marked this as closed rather than tracked.

Andromeda Quonset I agree, although I find that 99 is really, really thin. (Can check this on Opensim.) Still, no objection to 99.9 if it can be achieved. : )

Open sim is NOT the main grid and is irrelivent.

Pazako Karu While that's a different feature suggestion (and might want to post that to go on the work list), I surely agree. It would be nice to make a hollow rectangle or oval with even thickness all around. I'm glad you mentioned this here, because perhaps both features could be implemented in the same project.

Vincent Nacon Opensim has "pre-visited" it for years. Their hollow works up to 98.999 (technically 99). No problem with physics or Li. There is a chance this may have caused SL physics issues 21+ years ago, but I doubt that would still be the case.

Eren Padar That's because Opensim doesn't have Havok physic engine. It wasn't from 21+ years ago, it came up during the Havok 4 update in 2011. We used to have 99 before.

Vincent Nacon Okay so are you stating that the Havok engine can't hollow a sphere to 99? I would understand if it's a limitation such as that... but it's hard to believe that Havok is that limited.