Difference between revisions of "Talk:Linkability Rules"
(New section: New algorithm overview graphic?) |
|||
| Line 17: | Line 17: | ||
* Show that the spheres interact, and can be linked | * Show that the spheres interact, and can be linked | ||
...any good? [[User:Celierra Darling|Celierra Darling]] 08:24, 3 April 2008 (PDT) | ...any good? [[User:Celierra Darling|Celierra Darling]] 08:24, 3 April 2008 (PDT) | ||
I actually made a graphic before I read this and stuck it in there (the brightly colored prim shapes), but it doesn't cover the details of the math, just the order in which single prims are checked, then the failures are sub-checked, etc. I was trying to explain that part to someone, and the graphic just came about as a result. Perhaps a short animation would better explain the more detailed picture? [[User:Tiger Crossing|Tiger Crossing]] 17:21, 4 April 2008 (PDT) | |||
Revision as of 16:21, 4 April 2008
Although the initial failure mode of unlinkable sets is correct in principle. It doesn't take the fact into account the checking algorithm is 'global coordinate/position based', while linked sets can be rotated over general angles. A new linkability check over a valid linkable set, which is rotated (causing a different order in the prim selection of the linkability check) can cause this set to become invalid.
Linkability checks over already linked sets should be performed using the coordinate system as established by the rotation of the linked set, instead of a global coordinate system, to prevent a rotation of a valid linked set breaking the linked set. Damanios Thetan
I changed the SMALLEST_MAX to be LINK_BONUS, and bumped it from 1.0 to 2.0, as per changes in the code. User:Andrew_Linden
New algorithm overview graphic?
Kelly Linden and I added some attempts at better ways to explain the algorithm in a more-intuitive way. I think there probably needs to be something graphical, though. Assuming I transformed the equations correctly in the 'visualization' section, I'm thinking something like
- Two identical complicated-looking prims with bounding boxes with nice numbers, like 1.5x2x6 (diameter: 6.5)
- Show the bounding boxes as shown by the 'stretch' tool
- Replace the complicated-looking prims by boxes
- Add a translucent bounding sphere around each prim
- Add translucent spheres which are triple each bounding sphere
- Add translucent spheres which are one more meter around
- Show that the spheres interact, and can be linked
...any good? Celierra Darling 08:24, 3 April 2008 (PDT)
I actually made a graphic before I read this and stuck it in there (the brightly colored prim shapes), but it doesn't cover the details of the math, just the order in which single prims are checked, then the failures are sub-checked, etc. I was trying to explain that part to someone, and the graphic just came about as a result. Perhaps a short animation would better explain the more detailed picture? Tiger Crossing 17:21, 4 April 2008 (PDT)