a bell labs paper on optimal task scheduling from 1971! i hate it https://www.sciencedirect.com/science/article/pii/0012365X73901131
imagine:
- you are given a list of other people's work to do.
- you are told whose work is higher priority than others to perform. this precedence relationship cannot be questioned.
- each work item has its own constant uniform chance of catastrophic failure.
- catastrophic failure is to be interpreted solely as a failure to complete the work item, incurring a fixed "failure cost".
- success of a work item also incurs a constant "success cost".
this reads more like really bleak sci-fi than an actual published work. i am still waiting for anyone to even consider code–data locality and shared resource contention as a scheduling concern, let alone an optimization goal!
