# Evaluations across different geometries

I have two different point clouds. One creates a convex hull, for which I can access normals, face mid points, vertices, edges, etc.

The second point cloud (in blue) is scattered around the scene. For each point of this point cloud, I want to perform the following computation: $$\sum\limits_{\vec{n}} \left[\vec{n}\cdot (\vec{x_b} - \vec{x_g})<0\right]$$

The sum should be performed for each position of a blue point $$\vec{x_b}$$ over all normals and mid points $$\vec{x_g}$$ of the convex hull. The sum is a sum of boolean values and in the end, I want to know for each point, whether the sum is zero or larger. I was able to solve the task for one single point, where I had the position of the point as an input vector node and piped it into both geometry lines. But I have no idea, how I can do it for a collection of points, whose positions I do not know a priori.

Store your results in a custom attribute. I used an integer attribute that is increment when your comparison $$\vec{n}\cdot (\vec{x_b} - \vec{x_g})<0$$ is true.