Is there a way for adding and removing nodes from existing cluster (not just the node objects in master but actual instances too)? How do the authors see it happening or how is it done currently?

If there is no existing solution, I'm thinking of implementing this functionality either by

1. Two additional scripts like cluster/kube-spawn-node.sh and cluster/kube-destroy-node.sh. The logic would be similar to how kube-up creates nodes. I used spawn & destroy to avoid confusion with create/delete, hopefully signifying that actual machines are being created.
2. One one script like kube-resize-cluster which would take number of nodes as an argument and would spawn or destroy nodes based on that.

Do you feel there is a need for these kind of scripts? What approach would be better?

Regarding terminology, I've seen node and minion both being used. Do they mean the same or is node meant to replace minion?