The linearization and non-linearization of Z(Depth) Value in OpenGL

Now I'm implementing the point shadow map. However, The code of the learnopengl.com uses the linearization of the depth value to extend the precision of depth value. It means that most of the graphics application uses the non-linearized z-value in the shader to distinguish objects closer well.

So, I'd like to know the mathematical description of it, specifically the equations of linearization and non-linearization of z-value.

While finding some tips and articles on the internet, I found a good description of it on the learnopengl.com comments.

https://learnopengl.com/Advanced-OpenGL/Depth-testing

Among many comments, the comment which 'wang dingqiao' wrote is so good for the understanding of this topic.

Therefore, for the record of it, I just wanna copy the comment, because the description is perfect.