I would like to set up our renderfarm as a user and assign tasks to it. But if I say the task duration is 18 hours then it assumes it's a human being which needs to sleep and eat.
Similarly if we have a part time freelancer who can only work 4 hour days then assigning him 20 hours of work should come out to be 5 days.
I read the feature request "% Time" https://shotgunsoftware.zendesk.com/forums/31278/entries/39330 and this would be applicable as well. Bonus points if it can be handled automatically through an expression. So for instance if we assign 5 tasks which all take 24 hours to render the farm will need to split its time between those 5 tasks 20% each. So its max time per day is 24 hours but its max time for those days per task is actually only ~5 hours per day.
Seeing those conflicts on the schedule would let us plan for an impending bottleneck by purchasing more rendernodes or moving up render tasks to a time which doesn't conflict.
This would interact well with the dependent jobs feature request as well. Our render times are often fixed and knowing what a render depends on helps us communicate with the client that after "3PM on friday there can be no more changes for a Tuesday delivery".
Also if we set a schedule for a render we can tie it into our render management software through python. If a job is 50% done but it's used up 80% of its schedule then emails need to start flying to the people involved to figure out what is going on.