Sorry if this is a stupid question (and sorry for my english, its not my native tongue). My approach to learning houdini is “try stuff until you hit a brick wall”. Well, I hit it.

I'm trying to do the following: A number of Points (generated from an sphere via scatter) move because of a spring node with some turbulence. A custom operator with some python code type runs trough all the points and checks for the nearest neighbor. A new poly is then created to connect every point to its nearest neighbour (if this neighbour is inside a certain distance). This works.
But now I'm trying to stop all the points that have been connected from moving. To accomplish that I did the following: Created an attribute “attrMove” with default value of 1, in the python script I set this attribute to 0 if the point has been connected, then I create a group (filtered by the expression @attrMove so only the points that should still move are in the group). This group is used in the spring node, so it moves only unconnected points.

The problem: As soon as the point stops, it jumps back to it's initial position. But of course I need it to stop where it was a moment ago, not where it started.
I'm afraid I am too much of a beginner to solve this on my own, help is greatly appreciated.

since you want to use information from previous frames to influence next, you need to use simulation environment
you can do this using Solver SOP
and inside just copy P from previous frame if attrMove is 0

Looks good, but when I try to move the points slightly after they have been “frozen” (with a node below the solver) I experience strange behaviour: The Points don't move from their frozen origin but some other location.

I was not very precise with what I'm trying to accomplish: The Points should move randomly before they are “captured”, but afterwards they move also, just slower and in a smaller area.