Having consistent texture density is important to give a sense of realism but sometimes models carry insanely large textures. Goldilocks can, for a given viewpoint, identify textures that are too large for your usage of them. You will gain nothing but file size by keeping the textures that Goldilocks identifies. It certainly won't help your raytracing etc.
The original Goldilocks is here viewtopic.php?f=323&t=22658
Goldilocks v2.0 adds Geometry analysis. Its not just texture resolution that needs to be consistent, its also a useful sanity check to have consistent "Content Density" in your model. ie your model is built with a consistent amount of detail. Before you say it, yes I know sometimes you have a reason for having more detail in one area than another.
However, I keep finding models running slowly that turns out an espresso cup with a zillion faces in the corner of the room. So "Content Density" is a measure of "geometry per unit of space".
Goldilocks v2.0 adds another menu item that shows all your Groups/Component Instances ranked with respect to the whole models "Content Density". Basically, if you run it on a model and see a Group/Component waay out in front - so the amount of geometry per unit of space for that Component is far higher than the rest of the model - you might want to have a quick look to make sure thats what you really want.
Here what a report for geometry looks like. You can see a Group called "plate" has a much higher Content Density than everything else. Thats because that 1 plate has 38000 faces..
Adam



