List of features to implement:

	* progress meter

	* simulation mode: no execution

	* randomize node assignment

	* task graph scheduling
		* pure simulation mode: do only scheduling

	* write a man page

	* time limit for node completing the job -> migrate

	* re-try failed nodes after a given time period

	* add timestamps, calculate total time, average time

	* stop jobs, edit machinelist / jobs, continue jobs
