llDetectedPrimitiveParams()

Hey there, everybody! I have a suggestion that hopefully all you scripters out there will support. Although we have features like llDetectedPos(); and llDetectedRot();, our ability to know about our surroundings through scripts is fairly limited. If we had the ability to obtain more information through a function like llGetPrimitiveParams() for other objects, countless new fun and helpful scripts could be produced. The possibilities seem unlimited.

Perhaps the method could be passed the key of an object within the same sim, or it could work like the current llDetected() methods. Being able to know an object's prim type, lighting, color, size, and more, would be a wonderful new feature for LSL.

Make your way to www.secondlife.com/vote, search for prop #1709, and cast your votes!

Please post your thoughts on the topic here! I'd love to have this function implemented in a patch in the future.

OMGosh! I can now use LSL to copy someone else's creations!
Woohoo!

And aside from that there's a helluva lot of information involved in a primitive. iirc one primitive uncompressed uses about 4k of memory, when scripts can only deal with 16k total, and a sensor can return up to 16 primitives....

All we need really is the ability to perform something like llGetLinkedPrimitiveParams and llSetLinkedPrimitiveParams, that allow you to apply changes to any prim in a linked set, all in one call to the function without many scripts. Though many scripts is quite neat so even that's not too useful.

I'm having a hard time thinking up uses for this beyond prim mirroring, which is already extremely controversial. Whilst that is pretty handy for builders it's open to abuse and I don't think many people would like it at all.

I'd love to be able to at least get a detected object's "outline" in a script. The bounding box isn't good enough... getting a list of the prims and the individual bounding boxes for each prim would probably be enough. Right now autonomous robots simply can't "see" the world well enough to navigate except by doing things like spraying bursts of sensor bullets... which is a LOT more abusively expensive.

Guess you'd nee an interior bounding box for hollow objects, too, or else it'd be kind of embarassing for a robot trying to go through a doorway.

But the original proposal is clearly giving scripts Too Much Information.