I'm currently trying to figure the best way to use OnTime with Scrum and here's what I've come up with so far (it's important to keep in mind that none of this has progressed beyond the theoretical yet at this point, and I'm neither a ScrumMaster nor an OnTime expert.):

• Product Backlog = list of Features
• Feature item type = Requirement
• The Product Owner and Customer primarily view  the Project at this level
  
• Tasks are performed during Sprints to implement the Requirements included in the Sprint Backlog
• Developers primarily work at this level
• Tasks are created by the developers as they work on implementing Requirements ( not created until the Sprint planning session)
• Developers log all time worked and maintain estimated remaining work against Tasks (this allows incomplete Requirements to be moved around between Sprint folders w/o losing the Sprint work was actually performed in).
• Tasks are listed as Related Items under the appropriate Requirement
  

 Project Folder Hierarchy:

• Root folder = Project (same as usual)
• A new folder is created for each Release below the Root folder.
  
• A new folder is created for each Sprint inside the appropriate Release folder
• All Product Backlog items are initially kept in a at the project's root level
• Requirements are moved to a specific Release or Sprint folder during Release or Sprint planning sessions
  
• Add a custom field called Component to Requirements, Defects & Tasks to keep track of the screen, page, etc that the Item is related to. This allows you to group all Items related to a particular feature together without having to know your exact component folder structure upfront
  

Again this hasn't been tested in the Wild yet, but  In thinking it might work. I'd love to get anyone's feedback or suggestions.