Is the linear-attenuation light test acting over meters or centimeters?

I believe StandardDataSets\collada\library_lights\point\point_linear_attenuation_light\ is doing the linear attenuation calculation in terms of meters. (The light's inherited unit is centimeters. Like the rest of the CTS.)

The COLLADA manual only says it's done in terms of "distance" but does not define a unit. Normally distances are in the scale of the `` but is lighting meant to be an exception? It makes a difference if so. Managing scale factors for lights is difficult.

Nailing down a hard unit for lighting might be a good idea. But is this the idea? Or is it an accident?

Start with StandardDataSets/collada/library_lights/point/point_linear_attenuation_light/ and compare its expected behavior with the COLLADA manual's definition of distance. Determine whether the CTS assumes meters or centimeters, then document the agreed unit and update the relevant test expectations or documentation so the result is explicit.

Written by the indexing model from the issue text.