discussing tasks & time/#911

Apr 13, 2006 / pixtur
Jan 5, 2009 / phsouzacruz

Attached files

This is a draft:

Times of Tasksπ

  • tasks have following parameters related to time:
    • date_started
    • date_closed
    • estimated
    • planned_start
    • planned_end
    • cur_start
    • cur_duration

  • additionally...
    • tasks can by planned as for_milestone
    • a task can be depending on other task(s) (precessor_task)
      • because we need more than one precessor should be done with a join table taskdepencies
      • additional parameter after_precessor_duration which clarifies the time of the task it depends on. E.g. "Task2 starts at (10 % (task1.planned_end - task1.planned_start)) + task1.planned_start
      • default for after_precessor_duration is 100% (start after the precessor is over
  • planned_start, planned_end and planned_duration are interconnected
  • if all three are non-zero, the planned duration is still returned. Rolling out task orders depends only on the planned_end, though.
  • planned_duration should be given in seconds
  • additionally we can state a planned_duration_max which allows a basic form of risiko management
  • Mental note:
    • I am unsure about the rollout-duration if planned_duration and planned_duration_max are different. We could use the avarage, but pratically an average closer to planned_duration would make more sense, since estimated time gets less precise for long forcasts...



Rolling out tasks...π

  • when automatically rolling out tasks we should cache the computed temporaty start-time cur_start
  • this time has to be recalculated if...
    • a dependend time changes cur_duration or cur_start
    • the task-order changes...
  • not sure what happens with task that have a planned start
  • conflicts happen if:
    • planned_start && depends_on are given -> mark as delayed, if cur_start > planned_start ?

No Comments